A Roadmap to Future Space Connectivity: Satellite and Interplanetary Networks 3031307615, 9783031307614

Table of contents :
Preface
Acknowledgements
Contents
Contributors
Acronyms
Part I Satellite Communication Technology
1 Millimeter Waves and High-Throughput Satellites: The New Frontier Toward Terabit Connectivity in the Sky
1.1 Enhanced System Flexibility and Reconfigurability
1.1.1 Software Defined Networking
1.1.2 SDN-Enabling Payloads
1.1.2.1 Flexible and Reconfigurable Payloads
1.1.2.2 Inter-Satellite Links
1.2 Spectrum and Dynamic Spectrum Access
1.2.1 Current and Planned Allocations
1.2.2 Dynamic Spectrum Access
1.3 System-Level Capacity Improvement
1.3.1 Multi User-MIMO
1.3.1.1 Beam-Centric MIMO
1.3.1.2 User-Centric MIMO
1.3.1.3 MIMO Algorithms and Normalisations
1.3.1.4 Trends and Challenges
1.3.2 Beam-Hopping
1.4 Air Interfaces
1.4.1 Novel Waveforms
1.4.2 Adaptive Modulation and Coding
1.4.3 Air Interfaces for HTS in High Mobility Scenarios
1.4.4 Impact of HW Impairments
1.5 Feeder Link Evolution to Support
1.5.1 Smart Gateway Deployment and On-Ground Architecture
1.5.2 Feeder Link MIMO
1.6 On-Going/Planned Mission/Services and Mega-Constellations
1.7 Final Remarks
References
2 The Role of Satellite in 5G and Beyond
2.1 Non-Terrestrial Networks Standardization
2.1.1 3GPP Release 17: The First 5G NTN-Based Standard
2.1.2 NTN Radio Access Network
2.1.2.1 Direct User Access
2.1.2.2 Relay-Based User Access
2.1.2.3 Multi-Connectivity
2.1.2.4 Radio Protocol Issues and Adaptations
2.1.3 5G-Advanced
2.2 Services
2.3 System Architectures
2.3.1 User Segment
2.3.2 Space Segment
2.3.3 Ground Segment
2.4 Applications
2.5 Research and Development Activities
References
3 Futuristic Satellite Scenarios in 6G
3.1 Vision of 6G and Non-Terrestrial Networks
3.2 Architectural Perspectives of 3D Network in 6G
3.2.1 3D Network Platforms and Frequency Bands
3.2.2 Proposed Implementation
3.3 Localisation and RF Sensing in Satellite Networks
3.3.1 Satellite Network-Based Localisation
3.3.2 Satellite Network-Based RF Sensing
3.4 Design Parameters
3.4.1 UAV-Based Implementation of Relay Protocol in 3D Networks
3.4.1.1 Transmission Delay
3.4.1.2 Session Time
3.4.1.3 Signal-to-Noise Ratio
3.4.1.4 Throughput
3.4.1.5 System Parameters
3.4.2 UAV-Based Implementation of Radio Unit in 3D Networks
3.4.2.1 Fronthaul Bandwidth
3.4.2.2 Theoretical Throughput
3.4.2.3 Connection Density
3.4.2.4 Number of Functions
3.4.2.5 Fronthaul Energy Consumption
References
4 Quantum Satellite Communications
4.1 Introduction
4.2 Introduction to Quantum Communications
4.2.1 Superposition
4.2.2 Multiple Qubits and Entanglement
4.2.3 Bloch Sphere
4.2.4 Quantum Computing Gates
4.2.4.1 Single Qubit Gates
4.2.5 2 Qubit and Multi-Qubit Gates
4.2.6 Bell State Measurement
4.2.7 Entanglement Swapping
4.2.8 Measurement
4.2.9 Nondemolition Measurement
4.3 State of the Art in the Quantum Satellite Communications
4.3.1 Quantum Key Distribution
4.3.2 Entanglement Distribution
4.3.3 Entanglement Based QKD
4.4 Why Satellite-Based Quantum Communication
4.5 Recent Trends in Quantum Satellite Communications
4.5.1 Satellite-to-Ground Communication
4.5.2 Entanglement with Multiple Satellite Links
4.5.3 Continuous Variable Entanglement Distribution
4.5.4 Quantum Information Transfer Through Free Space
4.5.4.1 Single Photon Transfer along 7000km in Space
4.5.4.2 Optical Signals Through 38,600km into The Atmosphere
4.5.4.3 QKD Through Inter-Satellite Free-Space Links
4.5.5 Breakthrough in Quantum Satellite Communications Around 2017
4.5.5.1 Satellite-to-Ground QKD and Ground-to-Satellite Quantum State Transfer
4.5.6 Developments Towards a Global Network
4.5.6.1 Satellite Based Quantum Intercontinental Network
4.5.6.2 Global Navigation Satellite System Using Quantum Communication
4.5.7 Secured Quantum Entanglement-Based Cryptography Demonstration Using Satellites
4.5.8 Moving Boundaries from Lab to Real World
4.5.8.1 A 4600km Comprehensive Quantum Communication Network from Space to The Ground
4.5.9 An Overview of QKD experiments on-Orbit Over the Years
4.6 Final Remarks
4.7 Current Limitations in Quantum Satellite Communications
4.8 Future Works in Quantum Satellite Communications
References
Part II Systems and Infrastructures
5 Ground and Space Hardware for Interplanetary Communication Networks
5.1 Introduction
5.2 Ground Infrastructure for Interplanetary Communication Network
5.2.1 Deep Space Antenna Architecture
5.2.2 Uplink and Downlink Chains
5.2.3 DSN Capabilities
5.2.4 ESTRACK Capabilities
5.3 On-Board Equipment for Interplanetary Communication Networks
5.3.1 S/C TT&C Transponders
5.3.2 S/C Solid State vs Travelling Wave Tube Amplifiers
5.3.3 High- Medium and Low Gain Antennas
5.4 Direct Earth-to-Deep Space RF links vs Multi-Hop Links
5.4.1 Typical Performances and Limitations of a Direct Link
5.4.2 Data Relay Architectures
5.4.2.1 Data Relay Architectures for Earth Orbiting Users
5.4.2.2 Next Generation Data Relay Architectures for Earth Orbiting Users
5.4.2.3 Data Relay Architectures for Deep Space Users
5.4.3 Multi-Hop Links Advantages and Architectures
5.5 Guidelines for System Resources Allocation
5.5.1 Hardware Resources for a Traditional Single-Hop RF Link
5.5.2 Hardware Resources for a Relay Satellite
5.5.2.1 HW Resources for a Q/V EHF Band Link
5.5.2.2 HW Resources for an Optical Link
5.5.3 Technology Roadmap
5.6 Conclusions
References
6 End-to-End Space System: Engineering Considerations
6.1 Introduction
6.2 Statistical Nature of Onboard Data Generation
6.3 Statistical Nature of Communication Link
6.4 Link Optimization, Reliability, and Margin Policy
6.4.1 Review of Link Analysis Techniques
6.4.1.1 Link Budgeting Approach
6.4.1.2 Statistical Link Analysis
6.4.1.3 Minimum Margin for Link Design and Optimization
6.4.1.4 Illustrating Examples
6.4.1.5 Analysis Insight and Concept of Minimum Margin
6.4.2 ARQ Links for Reliable Communications and Its Statistical Characterization
6.4.2.1 Summary of Prior Results in ARQ Link Analysis
6.4.2.2 Statistical ARQ Link Analysis with Unlimited Number of Re-Transmissions
6.4.2.3 Statistical ARQ Link Analysis of Truncated ARQ with K Re-Transmissions
6.5 Concluding Remarks
References
7 Intelligent Space Communication Networks
7.1 Introduction
7.2 AI Improvements in Satellite Networks
7.2.1 Communication Resource Allocation
7.2.2 Security
7.2.3 Orbital Edge Computing (OEC)
7.2.4 Remote Sensing
7.2.5 Space-Air-Ground Integrated Network
7.2.6 Satellite Operations
References
8 Technologies and Infrastructures for a Sustainable Space
8.1 Space Sustainability: The Problem
8.2 Space Debris Mitigation/Removal
8.3 Sustainable-by-Design Approach: Enabling Technology
8.3.1 Concept of Sustainability by Design
8.3.2 BW/FW Compatibility: Federated Satellite Systems
8.3.3 BW/FW Compatibility: Joint Communication and Sensing
8.3.4 Ally Technologies
8.3.4.1 Softwarization
8.3.4.2 Autonomy and AI Tools
8.3.5 Very High-Speed Inter-Satellite/Inter-Layer Links
8.4 Conclusions
References
Part III Interplanetary Networking
9 Softwarization in Satellite and Interplanetary Networks
9.1 Introduction
9.2 Computational and Communication Technologies for Massive Space Exploration
9.2.1 Cloud, Edge, and Fog Computing
9.2.2 Network Coverage, Network Softwarization, and Network Automation
9.2.2.1 Virtualization and Softwarization
9.2.2.2 Backbone Network Technologies for Space Applications
9.2.2.3 Network Coverage Technologies on Remote Environment
9.2.2.4 Network Automation
9.2.2.5 Service Oriented Architecture for Intelligent Network Design
9.2.3 Internet of Things
9.2.4 Artificial Intelligence for Space Applications
9.3 Network Softwarization and In-Network Intelligence for Teleoperation, Telerobotic and Telepresence in Massive Space Exploration
9.3.1 Teleoperation
9.3.2 Telerobotic
9.3.3 Telepresence
9.3.4 Augmented Telerobotic
9.4 Teleoperation Using Edge Computing
References
10 Extraterrestrial Radio Access Network: The Road to Broadband Connectivity on Mars
10.1 Introduction
10.2 The Current (and Near Future) Picture of Martian Surface Connectivity
10.3 Martian Communications Supported by Sky Connections
10.4 LTE Connections Operating on the Mars Soil: Would They Be Practicable?
10.5 Advanced Solutions Based on 3D NTN and C-RAN
10.6 Conclusion
References
Part IV New Space Applications
11 Integration between Communication, Navigation and for Space Applications: Case Study on Lunar Satellite Navigation System with Focus on ODTS Techniques
11.1 Introduction: From Earth to Space Applications
11.2 Integrated Communication, Navigation and Sensing Systems for Space: Moon Case Study
11.2.1 Lunar Communication System Architecture
11.2.2 Lunar Satellite Navigation System
11.3 Orbit Determination Techniques for Lunar Satellite Navigation
11.4 Tracking from Earth or from Earth Orbit
11.5 Satellite Laser Ranging
11.6 Very Long Baseline Interferometry
11.7 High Sensitivity Spaceborn Receiver
11.8 Multiple Spacecraft Per Antenna Approach in TT&C Design
11.9 Tracking from Moon or from the Moon Orbit
11.9.1 Orbit Determination Candidate Configurations and Baseline Concept Introduction
11.10 Lunar Navigation ODTS System Architecture Baseline
11.11 Conclusions
References
12 The Internet-of-Things, the Internet of Remote Things, and the Path Towards the Internet of Space Things
12.1 Introduction
12.2 3GPP Non-Terrestrial Networks for the IoT
12.2.1 Innovations of the Latest 3GPP Releases
12.2.1.1 Release 16
12.2.1.2 Release 17
12.2.1.3 Release 18
12.2.2 Innovation Projects
12.2.2.1 5G-EMERGE (2022)
12.3 LPWAN for IoT
12.4 Open Challenges Towards the Internet of Space Things
12.4.1 Independent LEO Satellites for IoRT/IoST
12.4.1.1 LEO Satellites of IoT with and Without ISL
12.4.1.2 Commercial LEO Constellations for IoT
12.4.2 Physical Layer
12.4.3 Medium Access
12.4.4 Network and Higher Layers
12.4.5 Edge Computing
12.5 Conclusions
References
Epilogue
Index

Citation preview

Signals and Communication Technology

Claudio Sacchi · Fabrizio Granelli · Riccardo Bassoli · Frank H. P. Fitzek · Marina Ruggieri   Editors

A Roadmap to Future Space Connectivity

Satellite and Interplanetary Networks

Signals and Communication Technology Series Editors Emre Celebi, Department of Computer Science, University of Central Arkansas, Conway, AR, USA Jingdong Chen, Northwestern Polytechnical University, Xi’an, China E. S. Gopi, Department of Electronics and Communication Engineering, National Institute of Technology, Tiruchirappalli, Tamil Nadu, India Amy Neustein, Linguistic Technology Systems, Fort Lee, NJ, USA Antonio Liotta, University of Bolzano, Bolzano, Italy Mario Di Mauro, University of Salerno, Salerno, Italy

This series is devoted to fundamentals and applications of modern methods of signal processing and cutting-edge communication technologies. The main topics are information and signal theory, acoustical signal processing, image processing and multimedia systems, mobile and wireless communications, and computer and communication networks. Volumes in the series address researchers in academia and industrial R&D departments. The series is application-oriented. The level of presentation of each individual volume, however, depends on the subject and can range from practical to scientific. Indexing: All books in “Signals and Communication Technology” are indexed by Scopus and zbMATH For general information about this book series, comments or suggestions, please contact Mary James at [email protected] or Ramesh Nath Premnath at [email protected].

Claudio Sacchi • Fabrizio Granelli • Riccardo Bassoli • Frank H. P. Fitzek • Marina Ruggieri Editors

A Roadmap to Future Space Connectivity Satellite and Interplanetary Networks

Editors Claudio Sacchi Department of Information Engineering and Computer Science (DISI) University of Trento Trento, Italy

Fabrizio Granelli Department of Information Engineering and Computer Science (DISI) University of Trento Trento, Italy

Riccardo Bassoli Deutsche Telekom Chair of Communication Networks Technische Universität Dresden Dresden, Germany

Frank H. P. Fitzek Deutsche Telekom Chair of Communication Networks Technische Universität Dresden Dresden, Germany

Centre for Tactile Internet with Human-in-the-Loop (CeTI) Dresden, Germany

Centre for Tactile Internet with Human-in-the-Loop (CeTI) Dresden, Germany

Marina Ruggieri Center for Teleinfrastructures (CTIF) University of Roma “Tor Vergata” Roma, Italy

ISSN 1860-4862 ISSN 1860-4870 (electronic) Signals and Communication Technology ISBN 978-3-031-30761-4 ISBN 978-3-031-30762-1 (eBook) https://doi.org/10.1007/978-3-031-30762-1 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Space and terrestrial systems are more tightly related then anyone can suspect. Apart from the obvious cooperation and synergy that must exist between the space and terrestrial components of an integrated infrastructure, another subtended relationship characterizes terrestrial and space activities: the virtuous osmosis of architectures and technologies adopted on Earth to space and vice versa. It is perhaps more instinctive to expect a from-Earth-to-space osmosis of technologies and architectures and the spin-in as the natural way to go. The capability, instead, of space technologies and related architectures to pour innovations on Earth is amazing and the spin-out from space to Earth is very rewarding and surprising since at least the first Apollo mission! A consequence of the bidirectional contamination between Earth and space is that actions and choices of human beings in the space realm have an impact on Earth. It is a great opportunity, a fascinating challenge but also a very strong responsibility. Selecting the suitable technology for a space system is then much more than finalizing a new mission or a service in a cost and operation effective manner. To be more precise, let’s say that the measure of effectiveness risks to be inaccurate in terms of predictable and unpredictable effects in the medium and long term. This book is focused on a roadmap for future space connectivity that moves from awareness and responsibility in the use of the space domain. The above approach makes space indeed a main actor not only in the progress of knowledge and in the recognition of the recently added right of Humanity of “connecting the unconnected” but also in the capability of assuring a satisfactory future to all of us. Despite the neutrality of technology, that is neither good nor bad by itself but that can become either one according to the use we make of it, the roadmap for future space connectivity is studded with choices on both technologies and architectures that can turn out to be right or wrong for the ambitious goal of a suitable way ahead for mankind. Let’s be somehow disruptive here in measuring the suitability of technologies and architectures with their success of failure in passing a Glue Tech test of compliance. In fact, a Glue Technology (GT) is a powerful means of integrating v

vi

Preface

various components while effectively maintaining their autonomy. By applying the GT paradigm to future connectivity infrastructures, the integration between terrestrial, aerial and space components could pave the way for very ambitious goals, including “connecting the connected” in a truly sustainable manner. The GT paradigm is so promising that a group of experts worldwide created a technical panel of the IEEE Aerospace and Electronic Systems Society (AESS), named “Glue Technologies for Space Systems”, whose focus is the conception, design and application of Glue Technologies in space missions, infrastructures and services under the sustainability umbrella. The editors and most of the chapter authors of this book belong to the above AESS panel. The holistic nature of sustainability makes it not effective to focus on Earth without caring effectively for the surrounding domain (space). Therefore, space sustainability is part of the picture and its implementation contributes to reach the more obvious goal of a green Earth. In the book, space sustainability is both an underlying topic for different frameworks and a specific topic in a dedicated chapter. A powerful equation relates the capability of implementing sustainable space systems to the identification—and consequent use—of only Glue Technologies and related architectures. In fact, the horizontal ranges of application realms that a GT guarantees along with its related software-driven architectures offer to the system intrinsic pillars for its sustainable design, implementation and operations along with a good potential for its future recycling or, even better, upcycling. It is not dreamful to envisage in the future a system certification based on the use of only pure Glue Technologies and to expect a standardization activity related to the GT paradigm. We hope the reader will be captured by the GT concept and its broad range of implications and potentials and that he/she will be stimulated to contribute to the future of space connectivity in a sustainable manner. The book is organized in four parts dedicated to satellite communications technology (Part I), systems and infrastructures (Part II), interplanetary networking (Part III) and new space applications (Part IV). Disruptive technologies, configurations, implications, design guidelines and verticals will guide the reader in the articulated domains of the GT paradigm. Enjoy the journey .. . . Trento, Italy Trento, Italy Dresden, Germany Dresden, Germany Roma, Italy February 2023

Claudio Sacchi Fabrizio Granelli Riccardo Bassoli Frank H. P. Fitzek Marina Ruggieri

Acknowledgements

The editors of the book would like to thank the IEEE Aerospace and Electronic Systems Society (AESS) for allowing the vision about “Glue Technologies for Space Systems” become an AESS Technical Panel and for having recognized that vision and panel activities through the “Best Panel of the Year” award for two years in a row. The trust and recognition, in turns, have consolidated the idea of this book and profoundly inspired during its finalization. R. Bassoli and F. H. P. Fitzek would like to thank Deutsche Telekom for supporting over the last year and for their motivation to work on the topic of 6G and three-dimensional networking. We also thank CeTI team for supporting R. Bassoli and F. H. P. Fitzek. CeTI is funded by the German Research Foundation (DFG, Deutsche Forschungsgemeinschaft) as part of Germany’s Excellence Strategy— EXC 2050/1—Project ID 390696704—Cluster of Excellence “Centre for Tactile Internet with Human-in-the-Loop” (CeTI) of the Technische Universität Dresden. R. Bassoli and F. H. P. Fitzek also acknowledge the financial support by the Federal Ministry of Education and Research of Germany in the programme of “Souverän. Digital. Vernetzt.” Joint project 6G-life, project identification number: 16KISK001K.

vii

Contents

Part I Satellite Communication Technology 1

Millimeter Waves and High-Throughput Satellites: The New Frontier Toward Terabit Connectivity in the Sky. . . . . . . . . . . . . . . . . . . . . . . Ernestina Cianca, Marina Ruggieri, Alessandro Guidotti, Tommaso Rossi, and Giorgia Parca

3

2

The Role of Satellite in 5G and Beyond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mario Marchese, Fabio Patrone, and Alessandro Guidotti

41

3

Futuristic Satellite Scenarios in 6G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Justine Cris Borromeo, Koteswararao Kondepu, Mauro De Sanctis, Luca Valcarenghi, Riccardo Bassoli, and Frank H. P. Fitzek

67

4

Quantum Satellite Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sonai Biswas, Riccardo Bassoli, Janis Nötzel, Christian Deppe, Holger Boche, and Frank H. P. Fitzek

85

Part II Systems and Infrastructures 5

Ground and Space Hardware for Interplanetary Communication Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Paolo Tortora, Dario Modenini, Marco Zannoni, Edoardo Gramigna, Eliseo Strollo, Andrea Togni, Enrico Paolini, Lorenzo Valentini, Oreste Cocciolillo, and Lorenzo Simone

6

End-to-End Space System: Engineering Considerations. . . . . . . . . . . . . . . 139 Kar-Ming Cheung

7

Intelligent Space Communication Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Mario Marchese, Simone Morosi, and Fabio Patrone

8

Technologies and Infrastructures for a Sustainable Space . . . . . . . . . . . . . 185 Ernestina Cianca, Simone Morosi, and Marina Ruggieri

ix

x

Contents

Part III Interplanetary Networking 9

Softwarization in Satellite and Interplanetary Networks . . . . . . . . . . . . . . 203 Sisay Tadesse Arzo, Riccardo Bassoli, Michael Devetsikiotis, Fabrizio Granelli, and Frank H. P. Fitzek

10

Extraterrestrial Radio Access Network: The Road to Broadband Connectivity on Mars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Stefano Bonafini, Claudio Sacchi, Fabrizio Granelli, Koteswararao Kondepu, Riccardo Bassoli, and Frank H. P. Fitzek

Part IV New Space Applications 11

Integration between Communication, Navigation and for Space Applications: Case Study on Lunar Satellite Navigation System with Focus on ODTS Techniques . . . . . . . . . . . . . . . . . . . 243 Cosimo Stallo, Henno Bookmap, Daniele Cretoni, Martina Cappa, Laura De Leo, Mattia Carosi, and Carmine Di Lauro

12

The Internet-of-Things, the Internet of Remote Things, and the Path Towards the Internet of Space Things . . . . . . . . . . . . . . . . . . . . . . . . . 271 Fabrizio Granelli, Claudio Sacchi, Marco Centenaro, and Cristina Costa

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

Contributors

Sisay Tadesse Arzo University of New Mexico, Albuquerque, NM, USA Riccardo Bassoli Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany Centre for Tactile Internet with Human-in-the-Loop (CeTI), Dresden, Germany Sonai Biswas Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany Holger Boche Department of Electrical and Computer Engineering, Technische Universität München, München, Germany Stefano Bonafini Department of Information Engineering and Computer Science (DISI), University of Trento, Trento, Italy Henno Bookmap Telespazio, Darmstadt, Germany Justine Cris Borromeo Scuola Universitaria Superiore di Sant’Anna, Pisa, Italy Martina Cappa Ranstad Italia, Roma, Italy Mattia Carosi Thales Alenia Space, Roma, Italy Marco Centenaro System Solutions and Innovation Department, Bolzano Vicentino, Italy Kar-Ming Cheung Jet Propulsion Lab, California Institute of Technology, Pasadena, CA, USA Ernestina Cianca University of Rome “Tor Vergata”, Rome, Italy Oreste Cocciolillo Radio-Communication and TTC Products, Thales Alenia Space Italia, Rome, Italy Cristina Costa Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), Parma, Italy

xi

xii

Contributors

Daniele Cretoni Thales Alenia Space, Rome, Italy Christian Deppe Department of Electrical and Computer Engineering, Technische Universität München, München, Germany Mauro De Sanctis University of Rome “Tor Vergata”, Rome, Italy Carmine Di Lauro Thales Alenia Space, Roma, Italy Laura De Leo Ranstad, Roma, Italy Michael Devetsikiotis University of New Mexico, Albuquerque, NM, USA Frank H. P. Fitzek Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany Centre for Tactile Internet with Human-in-the-Loop (CeTI), Dresden, Germany Edoardo Gramigna Department of Industrial Engineering, Alma Mater Studiorum - Università, di Bologna, Forlì, Italy Fabrizio Granelli Department of Information Engineering and Computer Science (DISI), University of Trento, Trento, Italy Alessandro Guidotti Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), University of Bologna, Bologna, Italy Koteswararao Kondepu Department of Computer Science and Engineering Indian Institute of Technology Dharwad, WALMI Campus, Dharwad, India Mario Marchese Dipartimento di ingegneria navale, elettrica, elettronica e delle telecomunicazioni (DITEN), University of Genova, Genova, Italy Dario Modenini Department of Industrial Engineering, Interdepartmental Center for Industrial Research in Aerospace, Alma Mater Studiorum - Università di Bologna, Forlì, Italy Simone Morosi University of Florence, Florence, Italy Janis Nötzel Department of Electrical and Computer Engineering, Technische Universität München, München, Germany Enrico Paolini Department of Electrical, Electronic and Information Engineering “Guglielmo Marconi”, Interdepartmental Center for Industrial Research in Aerospace, Cesena, Italy Giorgia Parca Agenzia Spaziale Italiana (ASI), Via del Politecnico, Rome, Italy Fabio Patrone Dipartimento di ingegneria navale, elettrica, elettronica e delle telecomunicazioni (DITEN), University of Genova, Genova, Italy Tommaso Rossi University of Rome “Tor Vergata”, Rome, Italy Marina Ruggieri Center for Teleinfrastructures (CTIF), University of Roma “Tor Vergata”, Roma, Italy

Contributors

xiii

Claudio Sacchi Department of Information Engineering and Computer Science (DISI), University of Trento, Trento, Italy Lorenzo Simone Radio-Communication and TTC Products, Thales Alenia Space Italia, Rome, Italy Cosimo Stallo Thales Alenia Space, Roma, Italy Eliseo Strollo Department of Industrial Engineering, Alma Mater Studiorum Università di Bologna, Forlì, Italy Andrea Togni Department of Industrial Engineering, Alma Mater Studiorum Università di Bologna, Forlì, Italy Paolo Tortora Department of Industrial Engineering, Interdepartmental Center for Industrial Research in Aerospace, Alma Mater Studiorum - Università di Bologna, Forlì, Italy Luca Valcarenghi Scuola Universitaria Superiore di Sant’Anna, Pisa, Italy Lorenzo Valentini Department of Electrical, Electronic and Information Engineering “Guglielmo Marconi”, Interdepartmental Center for Industrial Research in Aerospace, Cesena, Italy Marco Zannoni Department of Industrial Engineering, Interdepartmental Center for Industrial Research in Aerospace, Alma Mater Studiorum - Università di Bologna, Forlì, Italy

Acronyms

3GPP ACM ACT ADC AEHF AFC AHM AI ALOS AoD API ARMA ARTES ASI BH BHC BHTC BHTP BSM BSS BW-Comp BWG CA CDM CDMA CEPT CF-MIMO CHAMP CIR CN0

3rd Generation Partnership Project Adaptive Coding and Modulation Adaptive Coding Transmission Analog-to-Digital Converter Advanced Extreme High Frequency Analog Fountain Code Active Hydrogen Maser Artificial Intelligence Advanced Land Observing Satellite Age of Data Application Programming Interface Auto-Regressive Moving Average Advanced Research in Telecommunications Systems Italian Space Agency Beam Hopping Beam Hopping Cycle Beam Hopping Transmission Channel Beam Hopping Time Plan Bell State Measurement Broadcasting Satellite Service Backward Compatibility Beam Waveguide Carrier Aggregation Code Division Multiplexing Code Division Multiple Access European Conference of Postal and Telecommunications Administrations Cell-Free Multiple Input Multiple Output Challenging Minisatellite Payload Carrier-to-Interference Carrier-to- Noise Ratio xv

xvi

CNN COMPASS CP CPRI CRAN CRT CSI CSTD CT CV DAC DAS DC DC DCU DD DL DRL DROM DS DSA DSA DSN DSRSS DSS DSS DSSS DST DT DTE DTP EDRSS EDSN EGS EHF EIRP ELFO eNodeB EOL EOORT ESA ESIM ESTEC ESTRACK

Acronyms

Convolutional Neural Network Combined Observational Methods for Positional Awareness in the Solar System Control Plane Common Public Radio Interface Cloud Radio Access Network Contention Resolution Timer Channel State Information Concurrent Spatial and Time Division Core network & Terminals Continuous Variable Digital-to-Analog Converter Distributed Antenna System Dual-Connectivity Direct Current Digital Channelizer Unit Delay-Doppler Deep Learning Deep Reinforcement Learning Data Relay for Moon Deep Space Dynamic Spectrum Access Deep Space Antenna Deep Space Network Deep Space Relay Satellite System Deep Space Station Distributed Satellite System Direct Sequence Spread Spectrum Deep Space Transponder Dwell Time Direct to Earth Digital Transparent Processor European Data Relay Satellite System ESA Deep Space Network Earth Ground Segment Extremely High Frequency Effective Isotropic Radiated Power Elliptical Lunar Frozen Orbits Extended Node B End Of Life Earth-Orbiting Optical Communication Relay Transceiver European Space Agency Earth Station In Motion European Space Research and Technology Centre ESA Tracking Station Network

Acronyms

FCN FDD FER FFR FHSS FR FS FSO FSS FSS FW-Comp GaAs GaN GEO gNB-CU gNB-DU GNSS GOCE GRACE GSFC GSO GW H/W HAPS HARQ HEMT HEO HGA HiRISE HPA HTS HydRON IAB ICI ICT IF ILRS INS IOAG IoRT IoT IP ISFFT ISI ISL

xvii

Fully Convolutional Network Frequency Division Duplexing Frame Error Rate Full Frequency Reuse Frequency Hopping Spread Spectrum Frequency Range Fixed Service Free-Space Optics Fixed Satellite Service Federated Satellite System Forward Compatibility Gallium Arsenide Gallium Nitride Geostationary Earth Orbit gNB Centralised Unit gNB Distributed Unit Global Navigation Satellite System Gravity Field and Steady-State Ocean Circulation Explorer Gravity Recovery and Climate Experiment Goddard Space Flight Center Geosynchronous Satellite Orbit Gateway Hardware High Altitude Platform Hybrid Automatic Repeat reQuest High Electron Mobility Transistors High Elliptical Orbit High Gain Antenna High Resolution Imaging Science Experiment High-Power Amplifier High Throughput Satellite High Throughput Optical Network Integrated Access and Backhaul Inter-Carrier Interference Information and Communication Technologies Intermediate Frequency International Laser Ranging Service Inertial Navigation Interagency Operations Advisory Group Internet of Remote Things Internet of Things Internet Protocol Inverse Simplistic Fast Fourier Transform Intersymbol Interference Inter-Satellite Link

xviii

ISS ITU JCS JDRS JPL kNN KPI LCAWG LCNS LCRD LEO LFO LGA LIDAR LLCD LLO LLR LMO LNA LOLA LOS LPWAN LRD LRNS LRO LSS LSTM LT LTE LUCAS LuGRE LUS MAC MarCO MC MEO MER MESFET MF MGA MIMC MIMO MIT ML MMIC

Acronyms

International Space Station International Telecommunications Union Joint Communication and Sensing Systems Japanese Optical Data Relay System Jet Propulsion Laboratory k-Nearest Neighbours Key Performance Indicator Lunar Communications Architecture Working Group Lunar Communication and Navigation Service Laser Communications Relay Demonstration Low Earth Orbit Lunar Frozen Orbit Low Gain Antenna Light Detection and Ranging Lunar Laser Communication Demonstration Low Lunar Orbit Lunar Laser Ranging Low Mars Orbit Low Noise Amplifier Lunar Orbiter Laser Altimeter Line Of Sight Low-Power Wide-Area Network Long-Range-Dependence Lunar Radio Navigation System Lunar Reconnaissance Orbiter Lunar Space Segment Long Short-Term Memory Luby Transform Long Term Evolution Laser Utilizing CommunicAtion System Lunar GNSS Receiver Experiment Lunar User Segment Medium Access Control layer MarsCubeOne Multi-Connectivity Medium Earth Orbit Mars Exploration Rovers MEtal-Semiconductor Field-Effect Transistor Matched Filter Medium Gain Antenna Monolithic Microwave Integrated Circuit Multiple Input Multiple Output Massachusetts Institute of Technology Machine Learning Monolithic Microwave Integrated Circuit

Acronyms

MMS MMSE mmWave ModCod MPA MPC MRO MSL MSPA MSS MSS MT NB-IoT NCC NEO NFV NGC NGSO NLOS NN NR NTN O-RAN OAI OBBF OBP OBPU ODTS OEC OFDM OGBF OOBE OTFS PA PAC PAE PAPR PDCP PER PHY PIMT PLMN PMD PNT PPP

xix

Magnetospheric Multiscale Minimum Mean Square Error millimiter waves Modulation and Coding Medium Power Amplifiers Minimum Power Constraint Mars Reconnaissance Orbiter Mars Science Laboratory Multiple Spacecraft Per Antenna Moon Surface Segment Mobile Satellite Service Mobile Termination Narrowband IoT Network Control Center Near-Earth Objects Network Function Virtualization Next Generation Core network Non-Geosynchronous Satellite Orbit Non Line Of Sight Neural Network New Radio Non-Terrestrial Network Open Radio Access Network Open Air Interface On-Board Beamforming On-Board Processor On-Board Processing Unit Orbit Determination and Timing Synchronisation Orbital Edge Computing Orthogonal Frequency Division Multiplexing On-Ground Beamforming Out-Of-Band Emissions Orthogonal Time Frequency Space Pilot Aided Per Antenna Constraint Power-Added Efficiency Peak-to-Average Power Ratio Packet Data Convergence Protocol layer Packet Error Rate Physical layer Propagation Impairment Mitigation Techniques Public Land Mobile Network Post-Mission Disposal Position Navigation and Timing Precise Point Positioning

xx

PU PVT QBER QKD QND QoS RA RAN RF RIS RL RLC RLM RMS RN RNN RPAS RRM RSE S/C SAE SAGIN SAR SatCom SCaN ScyLight SD SDAP SdE SDM SDN SDP SDR SEP SFFT SFGC SFU SGD SI SINR SISE SISO SNR SoL SOOP

Acronyms

Primary User Position Velocity and Time Quantum Bit Error Rate Quantum Key Distribution Quantum Nondemolition Quality of Service Random Access Radio Access Network Radio Frequency Reflecting Intelligent Surfaces Reinforcement Learning Radio Link Control layer Return Link Message Root Mean Square Relay Node Recursive Neural Network Remotely Piloted Aircraft Systems Radio Resource Management Radio Science Experiment Spacecraft Society of Automotive Engineers Space-Aerial-Ground Integrated Network Search and Rescue Satellite Communications Space Communications and Navigation SeCure and Laser communication technology Software Defined Service Data Application Protocol Sustainability design Efficiency Space Debris Mitigation Software Defined Network Software Defined Payload Software Defined Radio Sun-Earth-Probe Simplistic Fast Fourier Transform Space Frequency Coordination Group Solar Flux Unit Smart Gateway Diversity Study Item Signal-to-Interference-plus-Noise Ratio Signal In Space Signal Error Single Input Single Output Signal-to-Noise Ratio Safety of Life Signals Of OPportunity

Acronyms

SPC SRI SS SSA SSPA SSR SST STIN STK SU SubyD SVM SWE TA TC TDRS TDRSS TF TM TN TSG TT&C TTFF TWT TWTA TX TZD UAS UAV UE UHF UL-PC UP UTC VHTS VLBI VLMO VNF VSAT WI WLAN ZF

xxi

Sum Power Constraint Satellite Radio Interface Subsystem Space Situational Awareness Solid State Power Amplifier Space Sustainability Rating Space Surveillance and Tracking Satellite-Terrestrial Integrated Network Software Systems Tool Kit Secondary User Sustainability-by-Design Support Vector Machine Space Weather Timing Advance Telecommand Tracking and Data Relay Satellites Tracking and Data Relay Satellites System Time-Frequency Telemetry Terrestrial Network Technical Specification Group Tracking, Telemetry & Command Time To First Fix Travelling Wave Tube Travelling-Wave Tube Amplifier Transmission Troposphere Zenith Delay Unmanned Aerial System Unmanned Aerial Vechile User Equipment Ultra High Frequency Uplink Power Control User Plane Coordinated Universal Time Very High Throughput Satellite Very Long Baseline Interferometry Very Low Mars Orbit Virtual Network Function Very Small Aperture Terminal Work Item Wireless Local Area Network Zero Forcing

Part I

Satellite Communication Technology

Chapter 1

Millimeter Waves and High-Throughput Satellites: The New Frontier Toward Terabit Connectivity in the Sky Ernestina Cianca, Marina Ruggieri, Alessandro Guidotti, Tommaso Rossi, and Giorgia Parca

1.1 Enhanced System Flexibility and Reconfigurability Being the first chapter of a visionary book is both an honour and a responsibility. The road for future space connectivity is paved with many challenges and opportunities that space players might or not decide to embrace. Future space can be developed by selecting innovative technologies and system architectures, whose benefit is not simply to improve previous technologies and infrastructures but—most important— to allow a new approach to space that is inclusive, responsible, and sustainable. In this framework, this chapter deals with both a technology and an architecture that are key for future space infrastructures, envisioned in Fig. 1.1. In fact, the use of millimeter waves links and HTS based systems are not the mandatory evolution of connectivity in space, but they rather represent a disruptive way ahead that profoundly changes configurations, services, operators, users, and market related to the space realm [1–3]. Why does space need a disruptive, instead of a predictable—and likely boring— evolution? The answer perhaps is hidden behind Humanity’s future that seems paved

E. Cianca () · T. Rossi University of Rome “Tor Vergata”, Rome, Italy e-mail: [email protected]; [email protected]; [email protected] M. Ruggieri Center for Teleinfrastructures (CTIF), University of Roma “Tor Vergata”, Roma, Italy e-mail: [email protected] A. Guidotti Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), University of Bologna, Bologna, Italy e-mail: [email protected] G. Parca Agenzia Spaziale Italiana (ASI), Via del Politecnico, Rome, Italy © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_1

3

4

E. Cianca et al.

Fig. 1.1 High-level system architecture for future HTS systems

with very non-linear challenges: space is and must continue to be an active part of the picture and, thus, the capability to be disruptive in space is a way to align faster and better with Humanity’s issues. Connectivity has become a right of Humanity and a potential ally for a sustainable future. The “connecting the unconnected” goal is very ambitious [4]: connecting the .34% of the world population that is still offline needs both to assess the current efforts to reach the unconnected and to identify innovative solutions at architectural and service levels [5]. In this frame, a disruptive space component can greatly support the connectivity infrastructure to meet the goal. The “connecting the unconnected” paradigm calls for a magic blend between a need of the supply side for a wider coverage to be provided by the infrastructure as well as a need of the demand side for more relevant and cheaper services in order to succeed. Both needs can be fulfilled with a visionary approach to the architectural aspects of the network, a visionary approach to the use of technologies therein and a proper dissemination activity to make the unconnected aware of the deep need for connectivity. A suitable configuration for “connecting the unconnected” has to conjugate a sophisticated coverage with a flexible and reconfigurable architecture. It has also to care for space sustainability. The latter topic will be dealt with extensively in Chap. 8. Therefore, let’s focus here on coverage, flexibility and reconfigurability. In the above frame, both the widespread use of mm-wave links and HTS-based architectures could play a key role to make the magic blend for “connecting the unconnected” become a reality in a suitable timeframe. The HTS approach builds the coverage as a sophisticated combination of a large number of spot/narrow beams and frequency reuse in non-adjacent beams that brings a deep exploitation of the

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

5

bandwidth and then the high throughput capability. Network management, antenna configuration and satellite processing result as the main actors for the successful implementation of the HTS approach. The use of “brave” mmWave links in at least Q/V-band could then be “glue” for the development of a very effective architecture [1, 2, 6–8]. Below, we focus on a pillar of the HTS implementation, a networking based on the SDN paradigm and some related enablers at payload level.

1.1.1 Software Defined Networking One of the pillars that concur to a flexible and reconfigurable connectivity infrastructure is the increase of software pervasiveness into the network functions and management. The SDN paradigm is a powerful answer to the software pervasiveness through the decoupling between hardware and services by means of a softwarebased control and encapsulation. A simple conceptual view on SDN is sketched in Fig. 1.2, highlighting, in particular, the switching devices on the infrastructure, the SDN control software in the Control Plane (CP), and the Application Programming Interface (API) in the application plane. A software-defined approach is beneficial not only to the network but also to other key-elements of the configuration, like data storage. An SDN-based infrastructure is naturally keen to a fast innovation and to cloud orchestration and, consequently, to a high degree of reconfigurability and re-programmability. On the terrestrial side, SDN and NFV, along with cloud and fog computing, are driving the development of connectivity, while only recently the focus on SDN for space and space-terrestrial integrated networks is gaining the deserved

Fig. 1.2 Software defined networking: a conceptual view

6

E. Cianca et al.

importance [9–13]. The alignment between the efforts on the software definition on both terrestrial and space components is crucial. In fact, the ambitious “connecting the connected” goal can be reached in an effective time frame only if both the terrestrial and the space components have the same degree of flexibility, reprogrammability and reconfigurability in the global integrated infrastructure. It is worth mentioning that flexibility, re-programmability, and reconfigurability are also key for a sustainable environment, both on Earth and in space. This crucial topic is dealt with in Chap. 8. The effort to conceive, design, and deploy a global terrestrial-space connectivity infrastructure based on the SDN paradigm finds good allies in architectural and technology choices that give to the space network on one hand the “perfect” integration with the terrestrial component, on the other hand a suitable independence degree from the terrestrial portion. The latter is not a contradiction with respect to the integration goal, but instead it is a good strategic planning for resilience and flexibility of the global network over an extended timeframe. Among the space SDN allies it is worth mentioning the use of mmWave links, because they concur to both the capacity and the equipment size of the infrastructure. Pros and cons of the choice will be highlighted by the matter dealt with in the rest of the Chapter. A second SDN ally, that is both architectural and technological, is the provision to the satellites of the Inter-Satellite-Link (ISL) capability. ISL reduces the dependence of the space component on ground sections and it provides to each satellite of any constellation a more convincing role of intelligent node in a truly softwareprone network configuration. Both EHF (V or W-band) and optical ISLs can be adopted and their role is becoming crucial in the evolution of Global Navigation Satellite Systems (GNSS) as well as remote sensing and connectivity infrastructures, like mega-constellations [14–17]. As any choice, ISL brings pros and cons. It certainly increases the credibility of a future global infrastructure where terrestrial and space play a peer-to-peer role in a flexible, reconfigurable, re-programmable and sustainable manner.

1.1.2 SDN-Enabling Payloads The feasibility of the presented vision on SDN in satellite systems calls for extremely flexible payloads. In the following paragraph, the impact at payload level of such flexibility is discussed in terms of: (1) ISLs; and (2) reconfigurable payloads.

1.1.2.1

Flexible and Reconfigurable Payloads

The next generation of HTS/VHTS satellite systems will be supported by advanced on-board technologies to improve flexibility and reconfigurability. In this framework, the payload are evolving from monolithic bent-pipe architecture to flexible software-defined systems. Two are the main development trends for this new

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

7

generation of payloads, the first one is DTP (Digital Transparent Processor), the second one is the full SDP (Software Defined Payload). DTP initial concepts and architectures date back to the beginning of this century [18] and have been extensively analyzed and developed during the last decades [19]. These architectures are mainly applied to large Geostationary Earth Orbit (GEO) satellite systems offering broadband connectivity; P/Ls based on DTP include Analog-to-Digital Converter (ADC) and Digital-to-Analog Converter (DAC) at the Radio Frequency (RF) front-end and use digital processing to provide transparent connectivity with the objective to implement programmable and flexible channel routing with adjustable bandwidth allocation. The main element of DTP is the Digital Channelizer Units (DCU); the DCU implements digital filtering and switching with programmable mapping of input and output ports and channels between different satellite beams. DTP architecture offers full spatial routing, high granularity channelization, spectrum equalisation and gain control. The main objective is to support multiple beam antenna connectivity of feeder-links and user-links, offering high reconfiguration flexibility. In general, DTP performs processing of the signal without demodulation and decoding but new generations include also On-Board Processing Units (OBPU) sections able to perform demodulation/modulation and decoding/coding functions as well as IP-based onboard switching, thus improving the capabilities of on-board flexible payloads. Some examples of GEO satellite payloads/platforms embracing the software-defined paradigm are: • Eutelsat KONNECT VHTS, launched in 2022 and based on Spacebus NEO platform developed by Thales Alenia Space; • Eutelsat Quantum, launched in 2021 and developed through a PPP between ESA, Eutelsat and Airbus Defence and Space; • Inmarsat GX7/8/9 and Intelsat 42/43 satellites, under development and based on Airbus OneSat software-defined satellite architecture; • Thales Alenia Space platform named Space Inspire (INstant SPace In-orbit REconfiguration). Full SDP architectures are currently under development and test for small satellites (including cubesats); this architectural concept extends SDR capabilities, introducing flexibility provided by modularization of satellite components with standard interfaces, virtualization of the full infrastructure, development of applications from a large set of general micro-services. SDP functionalities can be adapted over-the-air, being it possible to use the same equipment to provide communications, navigation and sensing services. It has to be outlined that a strict standardization activity of SDP is needed in the following years; this activity has to be jointly conducted and integrated with the one of terrestrial networks. Examples of Non-Geosynchronous Satellite Orbit (NGSO) SDP approaches are: • Lockheed Martin’s SmartSat software-defined satellite platform; • Iridium NEXT system; • SES’s O3b mPOWER MEO satellites based on Boeing 702X platform;

8

E. Cianca et al.

It has to be underlined that one of the main drawback of software-defined architectures is the intrinsic increase in control complexity, thus requiring Artificial Intelligence (AI) based capabilities; moreover, this type of architectures bring vulnerabilities with respect to cyber-security. On the other hand, these drawbacks can also be considered as opportunities to improve overall design and implementation efficiency. On the antennas of such flexible payloads, the use of mmWaves allows to work with reduced on-board array antenna dimensions or with narrower beams. On the other hand, the power generation is less efficient with respect to lower frequencies. For what concerns active array antenna power generation, Travelling-Wave Tube Amplifiers (TWTA) still have a high efficiency performance with respect to Solid State Power Amplifier (SSPA), roughly .60% versus .35%, but they could not be the best solution for Low Earth Orbit (LEO) platform array antennas. Moreover, wide band-gap semiconductors technology based on gallium nitride (GaN) can be effectively used to develop very efficient, compact, linear, reliable and robust to radiation SSPAs and Low Noise Amplifiers (LNA). As a matter of fact, current technological developments provide the possibility to design SSPA with saturation power that can reach 30 W at Ka band and 20 W at Q/V-band. On the other hand, their efficiency still needs to be increased in order to be compliant with the constrained power budget of a LEO spacecraft and thermal management issues have to be faced.

1.1.2.2

Inter-Satellite Links

In a pervasively softwarized environment, higher volumes of signalling and control data will have to be exchanged by satellites. LEO-LEO ISLs can be divided in: intra-plane (same orbital plane), inter-plane (different orbital planes), cross-seam (in orbital planes moving in nearly opposite directions). Establishing cross-seam ISLs is rather challenging due to the very high Doppler shifts that are involved, especially when mm-waves are used (order of MHz). Intra-plane links are rather stable. Interplanes ISLs are characterized by higher dynamicity due to the slight asymmetries that are introduced to minimize the propellant when avoiding physical collisions between satellites at crossing points. Therefore, these ISLs are more challenging and must be established “on-the-fly”. To establish two intra-plane and two inter-plane ISLs (minimal configuration to get an advantage), we need to place on the satellite 4 transceivers and 4 antennas with related pointing system. Trade-offs to be considered are: dimension of antennas, achievable data rates vs satellite area (bigger antennas means higher data rates and/or lower transmitted power but also a smaller area to place the solar panels). A suitable solution for the antennas could be planar antennas. Limited onboard resources prohibit activating all ISL antennas simultaneously. Instead, an active antenna switch unit is connected to all ISL antennas. Once a session is established on the physical layer, attitude control must be used to keep the Line-OfSight (LOS) stable within the antenna field-of-view. Such high-data rate ISLs could

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

9

be implemented at optical frequencies or mm-waves. Optical ISLs are well suited as they offer small size, weight and power advantage with respect to traditional RF links. Moreover, they are not ITU regulated and hence, easier to operate. On the other hand, optical ISLs are susceptible to pointing errors, which could be rather challenging (and complex to implement over a small satellite) in case of interplane ISLs. Even if several planned networks suggest the use of laser ISLs that may achieve data rates of 100 Gbps, none of them so far features ISLs. RF ISLs present wider beams that reduce the sensitivity to the pointing errors and enables neighbor discovery procedures. However, to get also high data rates, narrow beams even with small antennas, the most interesting option for RF ISLs are at frequencies higher that Ka-band (Q/V/W bands). Recently a KA-band ISL has been tested with the following characteristics: distance of 4700 km, de-pointing of 2.◦ , within the antenna beamwidth of 5.◦ ; data-rate > 1.5 Mbps in nominal mode; power 30 W (output power from the amplifier 1.6 W).

1.2 Spectrum and Dynamic Spectrum Access 1.2.1 Current and Planned Allocations HTSs are driving the need to expand the system capacity as well as strong efforts on spectrum resources optimization, with the aim to respond to the accelerating global demand for reliable broadband communications. The delivery of such services is in fact enabled by HTS operating at Ka-band and above which are able to support sustainable business cases by leveraging on multi-beam technologies and intensive frequency reuse to increase the efficient use of spectrum and reach the Terabit connectivity. HTS embark the capabilities to use the same amount of allocated spectrum of traditional Fixed Satellite Service (FSS) satellites while expanding the available throughput, thus achieving a reduced cost of the Gigabits per second. Conventionally, as reported in [20], the satellite throughput is quantified on the basis of the re-use factor of the same portion of spectrum identified by N. Throughput is thus considered high when N spans from 5 to 10, and very high (VHTS) when .N > 10. Satellites for broadband communications high throughput capabilities up to the order of 100 Gbps have been launched in the last decade and mostly exploit the Ka-band; some examples are reported in Fig. 1.3 based on the launch date and providing an indication of the system capacity. The availability of spectrum resources in the EHF portion of the electromagnetic spectrum corresponding to the frequency range 30–300 GHz, as defined by the ITU [21] has the potential to be considered the new broadband frontier for satellite communications. Access to the radio spectrum at EHF is based on the ITU frequency allocations provisions [22], where defined categories of radio service, for satellite fixed, mobile and broadcasting services, are allocated to different portion of the spectrum and for different ITU regions, also reported in Fig. 1.4 for Space-to-Earth and Earth-to-Space communications, respectively; please note

10

E. Cianca et al.

Fig. 1.3 Evolution of the capacity of HTS/VHTS systems over time. Not exhaustive

Fig. 1.4 ITU frequency allocations in EHF to Space-to-Earth (above) and Earth-to-Space (below) FSS, MSS, and BSS

that we reported the allocations for FSS, Mobile Satellite Services (MSS), and the Broadcasting Satellite Services (BSS). In addition, it shall also be mentioned that ISLs are currently allowed to operate in the 32.3–33, 54.25–58.2, and 59–71 GHz spectrum. Given the limits on the use of the frequency spectrum, several bands are allocated for more than one radio service by adopting a shared approach based technical or operational compatibilities allow to provide services in the same (or adjacent) frequency bands without causing unacceptable interference to each other. System parameters such as antenna radiation patterns, transmission power etc. are com-

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

11

monly regulated by allowed margins in order to provide the proper compatibility and provided by decisions reflected in the International and National Tables of Frequency Allocations. The static sharing approach based on hard frequency or geographic separation is considered an effective solution, which, however, appear to be not sustainable in the short-term perspective due to the increasing satellite density. In fact, in the next few years the number of satellites is expected to grow rapidly as proved by the development plan of Starlink or oneWeb constellations which foresee the launch of tens of thousands of new satellites Interference management has to account for the scenario of GSO and NGSO operating in overlapping coverage areas in the same frequency bands, and this aspect becomes crucial in the view of the deployment of several systems with thousands of such satellites all sharing the same frequency bands. Considering the increasing satellite density, the static and fixed spectrum allocation model is not sustainable and will lead to a shortage of spectrum. Fixed spectrum assignment will eventually lead to limited performance as a consequence of interference as well as inefficient spectrum utilization. Therefore, next generation of spectrum management techniques are looking at overcoming conservative rules and big margins by introducing intelligence in the system and innovative dynamic spectrum sharing solutions based on spectrum sensing, location knowledge, ad-hoc databases, as instance, able to face and mitigate the interference occurrence in the high-density SatCom scenario.

1.2.2 Dynamic Spectrum Access Interference management in satellite systems is essential to guarantee the effective reuse and the sharing of spectrum by adopting a dynamic approach whenever different satellites, i.e., GSOs and NGSO, make use of overlapping coverage areas in the same frequency bands. The complexity increases in the scenario of multiple systems with thousands of satellites all sharing the same frequency bands and even more for providing connectivity to earth stations in mobility. Furthermore, the everincreasing diffusion of new terrestrial services is also translated in their presence within the satellite services bands thus incurring in the need for two different spectral coexistence paradigms, one between multiple satellites and the other between satellite and terrestrial services. The optimized exploitation of spectrum resources thus requires to overcome conservative and static frequency assignments and to move towards a dynamic spectrum sharing approach. The coexistence of FSS and terrestrial networks on the use of Ka band represents an exemplar case for which the European Conference of Postal and Telecommunications Administrations (CEPT) has adopted a decision [23] for providing guidance on downlink communication in the range 17.7–19.7 GHz used either by FSS and terrestrial networks. However, the new perspective to effectively mitigate the interference between multiple satellites and between satellite and terrestrial links is offered by the capability to dynamically adapt the spectrum allocation depending on the actual level of usage and detecting underused bands, by exploiting spatial, time, frequency separation with

12

E. Cianca et al.

dynamic allocation of spectrum resources. When dynamic spectrum management is concerned, a reference model considers authorized users as primary users (PUs), and users without license as secondary users (SUs). The latter can access the spectrum PUs if the primary spectrum is idle or share the primary spectrum provided that the services of the PUs can be protected at adequate conditions [24]. This spectrum access policy is known as dynamic spectrum access which provide a rule for the coexistence between PUs and SUs and two basic models can be identified, depending on the activity of PU, namely the opportunistic spectrum access and the concurrent spectrum access. In the first case, the inactivity of the PU on a certain frequency is recognized and the SU can access the resource for a defined timeframe. This can be done on deterministic base, by consulting a dedicated database, or by adopting spectrum sensing techniques for monitoring and detecting the PU inactivity. When the concurrent spectrum access occurs, the interference level between PU and SU cannot be avoided and has to be predicted and kept under a certain acceptable level. This case can also be treated by applying mitigation of interference with the help of advanced coding and transmission strategies. This set of opportunities for spectrum sharing, to be applied between two satellite systems or between satellite and terrestrial system, has led to the concept of cognitive SatCom [25], where the cognitive approach was already demonstrated to be effective for the optimized exploitation of spectrum resources in terrestrial systems. Cognitive SatCom sharing and interference mitigation techniques are mostly approached by research activities by exploiting centralized databases used to store and verify status and data on assignments for primary users and manage the availability of a channel for any secondary user for a certain timeframe, [26]. Cognitive radio techniques are applied to the satellite-terrestrial network operating in downlink and uplink in various frequency bands by cooperative spectrum sensing, where each earth station detects primary users’ transmitted signal and reports it to the satellite which makes the final decision on the state of the spectrum. In other approaches, power allocation algorithms are used to adapt the power transmitted by FSS satellite terminals, thus keeping the aggregated interference caused at the Fixed Sservice system below some acceptable threshold. It has to be noted that the majority of the proposed methods are developed for GSO satellites and FSS provided to static Earth stations. In this case only a static use case is concerned and the mobility aspects related to the coverage of NGSO and the mobile users should be taken into account in a more general and complete dynamic scenario. Several techniques are found in the literature addressing the coexistence among NGSO/GSO satellite systems and static or dynamic users based on adaptive power control, beam hopping, signal interferent to noise ratio thresholding, [27], however the major challenge is related to achieving the multipurpose capability to manage the spectrum utilization together with the interference mitigation.

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

13

1.3 System-Level Capacity Improvement Legacy HTS systems provide large capacity and spectrum efficiency over large geographical areas thanks to high frequency reuse by means of multiple on-ground spot beams. While the first HTSs were based on tens of spot beams, current designs of State-of-the-Art solutions envisage thousands of beams (e.g., ViaSat-3, launching in 2023, with 2000 beams per satellite providing terabit-capacity). While these systems promise to reach the terabit-capacity, future 6G services are foreseen to request further increased capacities, which can only be satisfied by means of VHTS for multi-terabit connectivity. In this framework, the past years have seen a massive effort on the exploitation of the available bandwidth, by either adding unused/under-used spectrum chunks through advanced and flexible spectrum usage paradigms (e.g., Cognitive Radios [28, 29]) or by decreasing the frequency reuse factor down to 1, i.e., full frequency reuse (FFR). In the latter case, which is addressed in this section, the co-channel interference from the adjacent beams is significantly increased, due to the side-lobes of the antenna radiation patterns. This calls for the adoption of advanced interference management techniques, as discussed below.

1.3.1 Multi User-MIMO During the last years, building on the success showed in terrestrial networks and thanks to the introduction of the Super-Frame structure in the DVB-S2X standard, the implementation and assessment of Multi User-Multiple Input Multiple Output () algorithms in SatCom has gained momentum, [30, 31]. In general, digital beamforming solutions can be broadly classified based on: (1) where the beamforming coefficients are applied to the signals, i.e., on-ground at the gateway (GW) or on-board the satellite; and (2) in which signal space the beamforming coefficients are computed, i.e., beam-space or user-space. In the following, .NF denotes the number of on-board radiating elements (elements of a Direct Radiating Array or feeds in an Array Fed Reflector antenna, for instance) and .NB the number of beams. We then assume .NU uniformly distributed (f eed) ∈ CNU ×NF the overall system users in the coverage area and we denote by .HT channel matrix, where: (1) the generic k-th row is the Channel State Information () vector of the k-th user; and (2) the generic .(k, n)-th element represents the channel between the n-th radiating element (feed) and the k-th on-ground user. In each time slot, the considered Radio Resource Management (RRM) algorithm selects a subset of users to be simultaneously served on the same frequency resources. Dennoting by .S the function representing the RRM scheduling, we thus obtain a .K × NF (f eed) channel matrix .H(f eed) ⊆ HT representing the channel during the considered time slot. Finally, we denote as .s = [s1 , . . . , sK ]T the K-dimensional column vector containing the unit-variance users’ symbols.

14

1.3.1.1

E. Cianca et al.

Beam-Centric MIMO

Beam-centric (or beam-space) MIMO solutions correspond to the first implementations of MIMO via satellite. In general, without MIMO, the signal received at the on-ground terminals with FFR can be written as: y = H(beam) s + z = H(f eed) Bs + z

.

(1.1)

where .z is a vector of independent and identically distributed (i.i.d.) complex circularly-symmetric Gaussian random variables modelling the system noise. The (beam) = H(f eed) B is an equivalent channel matrix in the beam .K × K matrix .H space: it is the projection of the .K × NF feed-space matrix .H(f eed) onto the Kdimensional space defined by the .NF × K beamforming matrix .B.1 Typically, legacy systems were designed to mitigate intra-system interference through multiple spot beams and frequency reuse or spatial separation; in particular, the beam shapes and the colour to be assigned to each beam are defined so as to minimise the interference.2 Aiming at increasing the system capacity through FFR, digital beamforming can be implemented in the beam-space defined by the equivalent channel matrix .H(beam) ; this is typically referred to as precoding. The baseline principle is that of computing a .K × K precoding matrix .W such that the product (beam) W is nearly diagonal, i.e., interference is nearly cancelled out. The received .H signals become: y = H(beam) Ws + z = H(beam) x(beam) + z

.

(1.2)

where we highlighted that the users’ symbols in .s have been precoded. In particular, the transmitted symbols are now given by .x(beam) = Ws and it is worthwhile highlighting that each precoded symbol is the linear combination of all of the K (beam) = w:,k s = users’ symbols, in fact: .xk i=1 wi,k si . More in general, it can be observed that: (1) the generic n-th row of the precoding matrix, .= wn,: , defines how the K signals are linearly combined at the n-th equivalent antenna (i.e., the nth beamformed channel); and (2) the generic k-th column of the precoding matrix, .w:,k , defines how the K equivalent antennas contribute to the signal directed towards the k-th user. The signal received at the k-th terminal is given by: yk =

.

(beam) hk,: w:,k sk





intended



+

K  j =1 j =k



(beam)

w:,j sj +zk





hk,:

(1.3)

interference

≤ NB in the beam space and, when .K = NB , one user per beam is served. Eq. (1.1), FFR is assumed. To take into account a frequency reuse factor greater than one, it is sufficient to select the rows of .H(beam) , .x, and .z corresponding to the desired colour. 1 .K

2 In

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

15

(beam)

where .hk,: represents the Channel State Information (CSI) vector (or channel signature) of the k-th user. The above formula leads to the following Signal-toInterference-plus-Noise Ratio (SINR):  2  (beam)  w:,k  hk,: .γk =  2    K  (beam) 2 σN +  j =1 hk,: w:,j   j =k 

(1.4)

where .σN2 is the noise variance. From the above SINR, the achievable spectral efficiency can be obtained by means of any standardised Modulation and Coding (ModCod) scheme or through the Shannon bound.

1.3.1.2

User-Centric MIMO

The performance of beam-centric MIMO in satellite systems has showed significant benefits in terms of achievable spectral efficiency and overall system capacity. However, as previously mentioned, precoding is applied to a system which has already been designed and optimised to limit inter-beam interference. The geographicallydefined beam coverage, .B, already defines the interfering patterns on which the precoding matrix, .W, operates, which might bring to non-optimal solutions, i.e., solutions obtained by optimising the problem on a limited search set. To tackle this issue, user-centric beamforming (also denoted as Cell-Free MIMO, , in line with the terrestrial communications) can be implemented. With CF-MIMO, the beamforming and precoding matrices are jointly optimised in the feed space and the vector of users’ symbols .s is directly projected onto the .NF -dimensional feed space: y = H(f eed) Ws + z = H(f eed) x(f eed) + z

.

(1.5)

where .W is the .NF × K complex beamforming matrix and .x(f eed) = Ws. In this case, from each on-board radiating element (feed) a linear combination of the K (f eed) = wn,: s = K users’ symbols is transmitted: .xn k=1 wn,k sk . Similarly to the beam-centric scenario, we can observe that: (1) the generic n-th row of the beamforming matrix, .= wn,: , defines how the K signals are linearly combined at the n-th radiating element; and (2) the generic k-th column of the beamforming matrix, .w:,k , defines how the .NF radiating elements contribute to the signal directed towards the k-th user. In line with Eq. (1.3), the signal received at the generic k-th terminal is given by:

16

E. Cianca et al.

Fig. 1.5 Detailed architecture for on-ground beamforming in the feed space

Fig. 1.6 Detailed architecture for on-board beamforming in the feed space

yk =

.

(f eed) hk,: w:,k sk





intended



+

K  j =1 j =k



(f eed)

w:,j sj +zk





hk,:

(1.6)

interference

and the SINR is still computed according to Eq. (1.4). Compared to beam-centric MIMO, with user-centric MIMO the precoding and beamforming matrices are jointly optimised based on the selected algorithm. Since, usually, .NF  NB , this means that the optimisation is performed with many more degrees of freedom, which leads to a much improved performance. As mentioned above, two architectures are possible: (1) On-Ground Beamforming (OGBF), in which the precoding matrix is computed and applied to the symbols on-ground, shown in Fig. 1.5; and (2) On-Board Beamforming, depicted in Fig. 1.6,

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

17

in which the precoding matrix is computed on-ground, but it is then applied to the users’ symbols on-board. Please note that in these figures we are assuming user-centric solutions, but the same architecture principles apply to beam-centric systems. Assuming that each user is allocated the entire beam bandwidth .Bbeam : • with OGBF, the K-dimensional users’ signals are projected onto the .NF dimensional feed space. Thus, the feeder link shall carry .NF signals, for a total required bandwidth equal to .NF · Bbeam ; • with OBBF, while the beamforming coefficients are still computed on-ground, their application is moved on-board. Thus, the feeder link shall carry K signals requiring a bandwidth .K · Bbeam , which si typically much lower than the request with OGBF. However, in this case also the beamforming coefficients shall be transmitted to the satellite, as well as an accurate clock reference so that the correct coefficients are applied to the corresponding users’ signals, with either in-band or out-of-band signalling. In both cases, the problem of the feeder link bandwidth requirement can be alleviated by using multiple GWs and advanced solutions as discussed in Sect. 1.5. In terms of requirements on the feeder link, it shall be noticed that, with beamcentric solutions, in both cases the total requested bandwidth is .K · Bbeam . With OBBF, also the precoding coefficients shall be sent together with an accurate clockreference. When regenerative payloads are available, the RRM algorithm and the computation of the beamforming matrix can be implemented on-board as well. This is applicable in the OBBF solution, since with OGBF the coefficients and scheduling must be available at the ground control segment and, thus, computed on-ground. For the sake of clarity, it is worthwhile recalling that On-Board or On-Ground Beamforming refer to where the beamforming coefficients are applied to the signals, not to where they are computed. Clearly, the solution with on-board RRM and beamforming computation has an increased system complexity and cost, but it has the advantage of reducing the overall latency and, consequently, the impact of CSI aging.

1.3.1.3

MIMO Algorithms and Normalisations

All of the main linear MIMO algorithms extensively assessed and discussed during the last years are based on the knowledge of the CSI vectors at the transmitter side. Thus, each user terminal shall estimate the channel coefficients from all of the beams or radiating elements, in the beam and feed spaces, respectively, and send it to the ground control segment in charge of computing the beamforming matrix. The most used CSI-based algorithms are: • Matched Filter (MF), in which the beamforming vectors for each user are given by the hermitian of their channel signatures:3 3 In

the following, we drop the beam/feed indication in the equations since the algorithms and the normalisations are applicable to both cases.

18

E. Cianca et al.

WMF = H

.

(1.7)

The hermitian operation can be interpreted as the digital steering of the signals (at beam or feed level) towards the actual users locations. • Zero Forcing (ZF), which is the least squares solution to the following system of linear equations: 2 .s = arg min Hs − x s

(1.8)

Notably, the solution to the above system is .H−1 . However, this spatial equalization completely neglects the impact of noise and, as such, it can lead to a performance degradation in terms of Signal-to-Noise Ratio (SNR). To circumvent this issue, and to also apply ZF to non-square not full-rank channel matrices, the following solution can be implemented:

† WZF = HH H HH

.

(1.9)

in which .† denotes the Moore-Penrose pseudo-inverse matrix. It shall be noticed that the inner Gram matrix .HH H has a dimension .NB × NB or .NF × NF in the beam and feed spaces; considering the trend in the number of beams/feeds for future VHTS systems, the dimension of this matrix might be extremely large and thus pose computational issues. In [32], an alternative solution is provided in which the inner Gram matrix has a dimension .K × K in both cases:

−1 WZF = HH HHH

.

(1.10)

In general, ZF minimises the co-channel interference; however, it does not maximise the overall system capacity and often the Gram matrix is ill-conditioned. • Minimum Mean Square Error (), which is a Regularized ZF (RZF) algorithm aimed at circumventing the above mentioned issues on the Gram matrix. In this case:

−1 WMMSE = HH H + diag (α) I HH

.

(1.11)

where .α is a vector of regularisation factors, for which the optimal value has been 2 /P , with .P being the total available on-board power, shown to be .K · sigmaN tx tx [33]. Also in this case, the inner Gram matrix might be challenging due to the number of beams or feeds. To circumvent this issue, the following equivalent, and computationally efficient, solution can be implemented: −1

WMMSE = HH HHH + diag (α) I

.

(1.12)

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

19

It is worth highlighting that specific power constraints are usually taken into account when computing the beamforming matrix; in fact, the total power emitted by the

satellite when beamforming is implemented is given by .tr WWH = W2F , which if left without control might lead to values larger than the total available onboard power .Ptx . . · F and .tr (·) are the Frobenius norm and the trace operators, respectively. Thus, proper normalisation approaches shall be implemented, aimed at regularising the combination of the transmitted power level from each on-board antenna so as to satisfy the power constraint requirements, [34, 35]. While several solutions have been proposed in the literature, the most general are the following: • Sum Power Constraint (SPC). An upper bound is imposed on the total on-board transmission power: .

W = W W

= tr WWH  W 2F

(1.13)

The total transmission power is upper bounded by .Ptx and the optimal vector relationships among the beamforming matrix columns are preserved; however, SPC does not provide any limit on the power per transmitted signal, which can pose some issues when taking into account the capability of the on-board HPAs. In particular, they might be working in the non-linear region, leading to intermodulation products and, thus, performance loss. • Per Antenna Constraint (PAC). In this case, the limitation is imposed on the power transmitted by each antenna: .

  1 1  = √1 diag W ,..., W  w1,:   wN,:  N

(1.14)

where .N = NB , NF for beam and feed space beamforming, respectively. In this case, while the maximum power per antenna is limited to account for the amplifiers capabilities, the vector relationships among the beamforming matrix columns are modified; this might lead to a loss in the performance due to an increased interference level. • Maximum Power Constraint (MPC). This normalisation aims at upper bounding the power per antenna, so as to not exceed .Ptx , while not modifying the result obtained by the beamformer: .

= W

W

(1.15)

N maxj  Wj,: 2

where it can be easily noticed that the maximum power transmitted by any on-board antenna is upper bounded by .Ptx , without modifying the vector relationships in the beamforming matrix. However, this approach might lead to sub-optimal performance since it tends to decrease the SNR due to the fact that not all of the available on-board power is being exploited.

20

1.3.1.4

E. Cianca et al.

Trends and Challenges

As outlined above, the implementation of MIMO techniques in Satellite Communications is extremely promising in terms of enhanced achievable capacity. However, there are still several technical challenges to be addressed, as detailed hereafter: • CSI aging and accuracy: in order to implement the MIMO algorithms defined in the previous paragraphs, the transmitter shall know the exact CSI from each user, i.e., the estimates at each user terminal of the channel coefficients from each of the on-board radiating elements. This can be achieved by means of Pilot Aided (PA) channel estimation algorithms as in DVB-S2X, where optional SuperFrame formats include Walsh-Hadamard pilots. However, the SuperFrame format is based on the concept of beam, thus allowing to estimate the channel at each user terminal from each beam. As such, it is not yet suited for cell-free (or usercentri) solutions that need the channel coefficients from each radiating element, i.e., before beamforming is applied. It is not straightforward to compute the feedspace coefficients from the beam-space ones, due to the non-linear operations involved, and, thus, this is a major challenge. In addition, MIMO algorithms are also impacted by the CSI aging and accuracy, i.e., by any misalignment between the actual channel matrix when the transmission occurs and the estimated one exploited to compute the beamforming matrix. This is impacted by both the accuracy of the channel estimation and, in particular,by the channel fluctuations and the users’ and satellites’ movement; notably, the latter is a major factor in NGSO systems. It shall also be mentioned that not all of the CSI coefficients might be estimated at the user terminal; in fact, the pilots from the adjacent beams might be received with a very low Channel-to-Interference Ratio (CIR) and, thus, the channel matrix shall be padded with nulls, leading to a significantly loss of performance. To tackle this challenge, location-based algorithms have been recently proposed, [36, 37]. In this case, the users provide their locations to the transmitter, which then estimates all of the channel coefficient terms that can be pre-computed based on the relative position of the user and the satellite, e.g., radiation pattern, slant range. This approach has been shown to yield a performance close to CSI-based algorithms, with the additional advantages of a reduced receiver complexity and lower overhead due to CSI reporting. The main challenges in this case are the impossibility to a priori know the stochastic terms in the channel coefficients (such as shadow fading or scintillations) and potential privacy issues due to the need for the accurate user position. As for the latter, it shall also be mentioned that looking at 3GPP Non-Terrestrial Networks (NTN), described in the next chapter, the users’ locations are not known at RAN level, i.e., at the gNB or GW, but rather at the core network. This is another challenge for location-based solutions. Another possibility to tackle the challenges related to CSI estimation is that of implementing Artificial Intelligence or Machine Learning algorithm for predictive channel modelling; this is still a quite unexplored area, which is likely to become a hot research topic in the near future.

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

21

• Scheduling: scheduling aspects for MIMO systems are particulary challenging. In fact, the actual SINR with which the receiver is served can be known only after the beamforming matrix has been computed, while this information shall be known in advance so as to properly select the users to be served in the same time slot. In general, these aspects have been tackled by means of clustering strategies, Genetic Algorithms, and pragmatic iterative approaches in which the users are iteratively added to a time slot subject to satisfying specific interference and traffic constraints, as proposed in [32] and the references therein. In the last couple of years, Artificial Intelligence and Machine Learning solutions have been gaining attention for MIMO scheduling and they can indeed provide a viable solution in the near future, also thanks to the enhanced on-board computational capabilities. • multiple GWs: as previously discussed, in order to cope with the large capacity requirements on the feeder link, multiple GWs are needed. This applies in particular to systems implementing MIMO. In general, it can be assumed that each GW manages a different subset of the total number of beams; in this case, the multiple on-ground GWs need to share the CSI information among them, or with a central unit in charge of computing the beamforming matrix, in order to build the overall channel matrix. To this aim, several approaches can be foreseen: (1) private CSI, in which there is no information sharing and each GW only knows the CSI of the users in its cluster of beams, which leads to an increased inter-beam interference due to the lack of cooperation; (2) shared CSI, in which each GW shares information on the CSI of the adjacent beams, which leads to a reduced inter-beam interference; and (3) full CSI knowledge, in which all of the CSI information is shared. Considering that, typically, the GWs in the ground segment are interconnected to manage atmospheric events, it can be assumed that the full CSI sharing is implemented. However, this still poses challenges in terms of signalling overhead, also on the ground segment, and additional latencies leading to a more relevant CSI aging. Aspects related to the implementation of multiple GWs in MIMO are discussed in [38–40] and the references therein. • Multicast precoding: in unicast precoding, each signal sent from a beam is dedicated to a specific user, i.e., the codewords sent on that beam are userspecific. However, when considering DVB-S2X, the SuperFrame format contains a large payload and, thus, it might be difficult to always fill it with a single user data. To this aim, multicast precoding was proposed (see [41] and the references therein), in which multiple users are multiplexed into the same codeword. In this framework, the main challenge is related to the construction of the channel matrix. In fact, multiple users are multiplexed into a single codeword that is then precoded according to an equivalent channel obtained as the average of the users’ channel coefficients vectors. Since the average channel coefficients exploited to compute the precoding vector depend on the selected users and the modulation and coding scheme (ModCod) used for the multiplexed users depends on the worst-case user, i.e., the user with lowest SINR, so as to guarantee that all users can decode their information, user selection, and grouping in the same codeword directly and deeply impacts the overall system performance in terms

22

E. Cianca et al.

of achievable throughput. In this framework, the main challenge is related to the optimal selection of the users to be inserted in the same multicast group, which can be modelled as a clustering problem, thus resorting to Machine Learning algorithms. Other advanced approaches can be envisaged as well, e.g., Artificial Intelligence and Deep Neural Networks, but, for the moment being, they are still unexplored. • Distributed NGSO CF-MIMO: NGSO systems, in particular LEO and VLEO, have been gaining an increased attention in both the academic and industrial fora. The implementation of MIMO techniques in this framework is a relatively recent topic, which has been addressed for instance in the context of European research project, [36, 37]. The larger velocity of the satellites poses challenging issues in terms of the CSI estimation and aging, management of the multiple on-ground GWs, and scheduling. In addition to these, also the implementation of MIMO from multiple NGSO satellites operating as a formation, or swarm, has been proposed. In fact, by exploiting Inter-Satellite Links, is it possible to implement distributed CF-MIMO algorithms based on the CSI from all of the radiating elements located on-board the different satellites, i.e., without the assumption of co-located radiating elements. These systems is promising, in particular thanks to the increased channel diversity leading to benefits in harsh propagation conditions and NLOS channels. The main challenge is related to the need for very accurate time and frequency synchronisation among the satellites operating as a Distributed Antenna System (DAS) and the increased signalling overhead.

1.3.2 Beam-Hopping Exploiting the enhanced on-board flexibility in terms of resource allocation and computational capabilities, Beam Hopping (BH) is a promising technique providing the ability to allocate the capacity where it is actually needed in the time, space, and frequency domains. In fact, the traffic is significantly non-uniform and it may vary significantly over the satellite lifetime. Traffic-agnostic solutions, in which the resource allocation does not take into account the users’ needs, a significant amount of the available capacity might be allocated to users not needing it, leading to potentially large unmeant capacity levels for users requiring larger throughput. BH is a technique in which the different beams are served by time-division multiplexing. As shown in Fig. 1.7, a BH satellite system consists of .Nb beams that are grouped into K clusters (.K = 3 in the figure). In each time slot: (1) only one cell per cluster is active; (2) the active cell uses the full cluster power and bandwidth; (3) each cluster is served by one transmission channel, i.e., the users belonging to beams in the same cluster are served by time-multiplexing. In general, as reported in [42], the geographical distribution of the cells in a cluster can be determined based on system optimization (e.g., the traffic demand or an interference management strategy) or payload constraints (BH implementation technology and payload architectures).

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

23

Fig. 1.7 General beam hopping concept: cells and clusters

With uniform cluster sizes, there are . Nb /K beams per cluster, while in nonuniform cases the number of beams per cluster can vary. There are two main strategies to implement BH: (1) pre-scheduled BH; and (2) traffic-driven BH. In the former case, one or multiple BH transmission channels (BHTC) are transmitted periodically to serve one or multiple clusters.4 With a predetermined scheduling, each cell in a cluster is revisited periodically according to the Beam Hopping Time Plan (BHTP) and the BHTC repeats its BHTP with a repetition equal to the BH cycle (BHC). A BHTC can contain a single carrier or multiple carriers, with the latter allowing the different carriers to carry variable symbol rates. In general, the BHTC can include different carrier frequencies, bandwidths, and number of carriers; in this framework, it shall be noticed that the frequency band assigned to different BHTCs can full overlap (leading to BH in full frequency reuse), partially overlap, or not overall at all (leading to Frequency Division Multiplexing). An example of BHC is provided in Fig. 1.8, where it shall be noticed that each BHTC can operate on multiple or single carriers and for a variable time, denoted as Dwell Time (DT). In order to avoid inter-cluster interference, guard times are required before and after the transmission occurs, so as to allow

4A

BHTC is the transmission channel in charge of serving one cluster.

24

E. Cianca et al.

Fig. 1.8 Example of beam hopping time plan

the satellite payload to modify its beamforming matrix; these DTs can include sequences to support fast beam re-acquisition at the user side. The DT can be completely arbitrary from one BHTC to the other; however, when they are integer multiples of the same value, a regular time grid is obtained, which leads to a more robust BH synchronisation since the terminal can search for the signal at fixed time intervals. In pre-scheduled BH, the allocated band and DT are pre-determined based on the knowledge of the traffic demand distribution. As such, in case the traffic demands vary, the BHTP needs to be modified accordingly; moreover, pre-determining the scheduling might lead, before the BHTP reconfiguration, to a large delay in serving the end users requiring larger traffic. To tackle these issues, traffic-driven BH can be implemented. Based on regenerative payloads and systems with on-board processors, the BHTP can be tailored to the current traffic demands. In this case, the BHTP is not repetitive and each packet is sent as soon as it arrives to the system, thus reducing the delay. There is thus a single queue for all users, taking advantage of statistical multiplexing. In this case, the illumination plan is actually random and this might lead to interference when adjacent cells are served simultaneously; these aspects shall be taken into account when designing the scheduling algorithm. It is worthwhile highlighting that previous versions of the DVB-S2X SuperFrame formats were all based on a fixed payload length. As such, they did not allow the efficient implementation of BH solutions, as the DTs would have been fixed as well. Thus, in the most recent version of the DVB-S2X standard, three additional SuperFrame formats were added based on variable-length payload sizes. These formats allow both pre-scheduled BH and traffic-driven BH [42]. There are several advantages thanks to the implementation of BH: (1) the number of on-board transponder RF high-power chains is reduced, leading to a reduced payload complexity and cost (also in terms of mass and accommodation); (2) one

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

25

HPA is serving multiple beams, which is well suited for large coverage areas; (3) it is possible to operate with a single carrier per HPA, thus lowering the required backoff; and (4) there is a limited additional complexity to implement it in payloads equipped with active antennas and beamforming. The potential applications and scenarios are manifold, including GEO, LEO, or MEO satellites and broadband bidirectional traffic (Business-to-Business and Business-to-Customer), maritime and air-borne in-flight communications, Voice over IP, and Internet of Things (IoT). In general, pre-scheduled BH is likely to be applied to HTS GEO systems, in which the coverage area is extremely large with more limited variations in the traffic demand per cell; traffic-driven BH solutions might be preferred when larger variations are expected, i.e., in NGSO systems with fast moving satellites and reduced coverage areas. It shall be also mentioned that the capacity provided by a single beam might not be sufficient for some hot-spot cells; in this case, more advanced solutions can be foreseen as: (1) the combination of MIMO and BH, as also proposed in [32]; (2) modifying the cell size based on the traffic request, i.e., identifying the number of beams in hot-spot areas and merging multiple beams in cold-spot areas, [34, 43].

1.4 Air Interfaces The design choices of the air interface for HTS working at mm-waves, must be performed considering the following impairments [44]: • high level of phase noise; • propagation impairments (especially related to tropospheric precipitations); • HW impairments [45]. The competing drivers of reaching Giga-Baud rates and building low-cost, lowcomplexity user terminals cause analog RF components, such as analog mixers, antialiasing filters and amplification in quadrature frequency-conversion architectures to exceed their tolerance limits and create direct-current offset and in-phase/quadrature (I/Q) imbalance that is strong and frequency-selective. The receiver design should include techniques able to mitigate such effects. As a matter of fact, the air interface of current HTS systems is based on DVB-S2 standard and its evolution in DVB-S2X. Therefore, traditional modulation techniques are used with a single carrier approach, even if with the key capability to select several transmission modes, i.e., different combinations of modulation and coding schemes. It is worth outlining that operators buy their modems from vendors that manufacture very standard and consolidated solutions. Most of the effort in the design of an air interface that is robust to the typical impairments of a mm-waves satellite link, such as the high phase noise, strong hardware and propagation impairments, has been moved to the optimization of Adaptive Coding and Modulation (ACM) techniques, smart gateway diversity, beam-hopping. On the other hand, several research works have tackled the issue of designing novel waveforms for mmWave satellite links [7, 44, 46]. The design guidelines proposed

26

E. Cianca et al.

in the above-mentioned research works will become more important when W-band or even the THz range will be considered for transmission of HTSs and gaining few dB of margin could become crucial. It must be also outlined that some of the most interesting application scenarios for satellite systems involve high mobility nodes as either the user terminal is highly mobile (e.g., airplanes, trains) or the satellite is a LEO satellite and hence, moves fast. At mmWave, high mobility scenarios are characterized by very high Doppler shifts thus making more challenging the design of the air interface. Finally, it is worthwhile mentioning the 3GPP NTN standard included in Release 17, which specifies the features enabling 4G and 5G systems to support a satellite component [47]. More than technical specifications, it also enables the integration of the satellite industry in the 3GPP ecosystem to ensure a global market. The role of NTN in 5G is extensively addressed in Chap. 2. Below, we discuss: (1) insights on the design of novel waveforms more suitable for mmWave satellite links; (2) the research activity in the framework of the ACM techniques optimization; and (3) the proposed solutions and future trends to face the challenges related to high mobility scenarios and the impact of the HW impairments.

1.4.1 Novel Waveforms Guidelines for the waveforms design of mmWave satellite communications, in particular for Q/V and W bands, have been presented in [46], where a detailed comparison of several waveforms in presence of nonlinear distortions and typical values of phase noise have been shown. Two types of waveforms have been compared: Constant Envelope multicarrier waveforms (CE-OFDM and CE-SCFDMA) [48] and single carrier impulse-based waveforms such as Time Hopping-UWB (THUWB), Direct Sequence UWB (DS-UWB), and Prolate Spheroidal Wave Functions (PSWF)-based Pulse Shape Modulation (PSM) [7]. PSWF waveforms (originally proposed in short-range indoor ultra-wideband communications) optimizes the tradeoff between the concentration of pulse energy in a finite time interval and in a limited bandwidth. Theoretically, binary PSWF-based PSM exhibits the best tradeoff between envelope compactness, spectral compactness and robustness against EHF link impairments. On the other hand, the related waveform generation is complicated, computationally intensive and not affordable by current State-of-theArt signal processing architectures. The generation and the detection of TH-UWB signals are very simple and cheap. Such a solution is robust to nonlinear distortion, but it is vulnerable to the effects of phase noise. At the opposite, DS-UWB modulation format is more affected by nonlinear distortions. The use of CE-OFDM and CE-SC-FDMA may be suggested when the phase noise level is moderate (e.g., in Q band).

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

27

1.4.2 Adaptive Modulation and Coding The strong propagation impairments, especially related to the tropospheric precipitation makes mandatory the use of channel impairments mitigation techniques such as ACM [49–51], data rate reduction, up-link power control, on-board dynamic power allocation, antenna pattern reconfiguration, or smart gateway diversity [52]. In this framework, the optimization of ACM techniques over “beyond Ka-band” satellite transmission channels is an active field of research. Many of these solutions are based on rain attenuation prediction to enable timely decision based on the short-term weather state and improve the effectiveness of such techniques; most of these algorithms are based on machine learning approaches. Such prediction techniques should also take into account the limited space resources in terms of onboard processing capability and power consumption. In [53], the auto-regressive moving average (ARMA) model has been used to design a forecasting algorithm. This model can predict the N-state Markov chain of rain attenuation on a Ka-band link. Moreover, they propose an adaptive coding transmission (ACT) scheme over the Ka-band channel based on the analog fountain code (AFC), which is a rateless code that combines the Luby Transform (LT) code with m-ary modulation [54]. The proposed ACT scheme can approach the Shannon limit over a wide SNR range.

1.4.3 Air Interfaces for HTS in High Mobility Scenarios High mobility communication scenarios operating at high carrier frequency suffer from sever Doppler spreads. Moreover, the excessive phase noise associated with high-frequency oscillators at mm-waves, results in a time-varying composite channel. Conventional Orthogonal Frequency Division Multiplexing (OFDM) transmission is impaired by severe inter-carrier-interference (ICI) in such scenarios. Adaptive coherent/non-coherent detection have been shown to be robust against high Doppler Spreads [55]. Nevertheless, an increasing amount of research has been dedicated to the design of new modulation waveform for high mobility scenarios. Recently, a novel modulation scheme has been proposed as a candidate for highmobility scenarios and also robust to the high oscillator phase noise that is crucial at mm-waves, namely Orthogonal Time Frequency Space (OTFS) modulation [56, 57]. Moreover, OTFS is characterized by low Peak-to-Average Power Ratio (PAPR), which is important in scenarios where the user terminals such as airborne and spaceborne vehicles, have limited on-board power supply. In [58], the performance of OTFS modulation on satellite-to-ground high mobility communications has been analyzed, both for the case of GEO satellites and LEO satellites. Results show the feasibility of applying OTFS in air-to-ground communications with high mobility. OTFS modulates the information in the delayDoppler (DD) domain rather than in the Time-Frequency (TF) domain such as OFDM. While OFDM transforms a a frequency-selective channel to multiple

28

E. Cianca et al.

Fig. 1.9 OTFS TX-RX scheme

parallel frequency-flat subchannels, OTFS transforms a time-varying channel into a 2D quasi-time-invariant channel in the DD domain. Figure 1.9 shows the scheme of the transmitter and receiver, which consist of a cascade of two 2-dimensional transforms: the 2D Inverse Simplistic Fast Fourier Transform (ISFFT) and the Heisemberg transform at the transmitter; the Wigner transform and the direct 2D Simplistic Fast Fourier Transform (SFFT) at the receiver. It is worth mentioning that waveforms that are produced in OTFS appear to be natural candidates for joint communication and sensing as the estimations used for the radar signals rely on delay, Doppler and angular features of the resolvable paths, which are the parameters that must be estimated at the receiver to demodulate the information.

1.4.4 Impact of HW Impairments Satellite channels typically introduce signal degradation due to the following: • linear distortions, introduced in the form of Intersymbol Interference (ISI), which might be caused by the propagation channel (frequency selectivity and scintillations) or by imperfection in the magnitude and group-delay responses of the IMUX and OMUX filters on board the satellite; • nonlinear distortions due to high power amplifiers (HPAs) (with memory or memoryless, AM/AM and AM/PM characteristics). Moreover, satellite links are usually affected by I/Q imbalance and phase noise distortions introduced by the signal up/down-conversion circuits. At mmWave the above distortions are stronger and saving even few dB of margin could be important to increase the spectral efficiency or to reduce the packet error rate (PER). In this context, it is expected that the equalization may play an important role. Equalizers can be introduced at the receiver side to mitigate linear distortions. However, in satellite communication systems, equalization is also seen as one of the techniques to reduce the sensitivity to nonlinear distortions and enable a better use of satellite power [59–61]. Much work can be found on either assessing the performance of rather standard equalizers over satellite links, taking into account RF and HW impairments or in proposing more advanced equalizers for either improving

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

29

the performance or reducing the computational complexity and the length of training data [60, 62–64]. In [62], two characterization models of analog RF impairments in the presence of frequency offset have been proposed along with novel digital compensation algorithms with immunity to frequency offset. Most of the mentioned works, consider satellite links working up to Ka-band and show performance assessment through simulations with more or less accurate models of the RF and channel impairments. The CMA is known to be robust to frequency offset and phase noise and that makes it suitable for use in EHF links, which are characterized by large phase noise. Moreover, thanks to its low implementation complexity, it does not increase the cost of the satellite terminal; on the contrary, equalization can be useful to counteract degradation caused by terminal front end, thus reducing its costs. The performance of such equalizer have been experimentally tested over a real DVB-S2 satellite link in Q/V band. Different transmission schemes, based on the DVB-S2 standard, have been tested: QPSK, 8PSK, and 16APSK modulation with different code rates. In all the considered experimental scenarios, the CMA equalization provides a SNR gain that ranges between 0.2 and 0.5 dB (for a target PER of .10−4 ), which is a relatively high gain considering that the maximum achievable gain ranges between 0.4 and 1 dB. In a Q/V-band broadband transmission scenario where the channel impairments are very strong and ACM techniques are used, such an improvement is relevant and might result in a large throughput increase for an HTS system. Moreover, such a gain is achieved with a negligible additional complexity at the receiver side.

1.5 Feeder Link Evolution to Support 1.5.1 Smart Gateway Deployment and On-Ground Architecture The short-term scenarios of HTS systems foresee the exploitation of EHF bands, in particular Q/V and W bands, but also free-space optics (FSO) in the feederlink to support broadband user access services. It has to be underlined that one of the requirement for HTS feeder link is an availability of 99.5–99.9% of the total time. Considering a single satellite link, such strong requirement at frequency bands higher than the Ka-band could be reached only using a very large link margin, leading to a waste of resources and an increase in the overall cost of the system. In this context, spatial diversity plays a key role, being the only technique that can guarantee such high link availability maintaining reasonable link dimensioning (i.e. antenna size, required power, etc.). In this framework, the most promising spatial diversity technique is the so-called Smart Gateway Diversity (SGD). The latter has been presented for the first time in [65] and is a feeder link spatial diversity scheme based on a pool of synchronized gateways (GWs), connected by a terrestrial fiber network, whose architecture provides the possibility to route

30

E. Cianca et al.

and distribute feeder link traffic between GWs nodes in order to counteract fading on one (or more) GWs. Traffic management is performed by a network control center that is also responsible to handle the traffic handover when deep fading occurs. SGD configuration design and management is a complex task that has to be performed optimizing traffic handover and switching control procedure, as well as diversity management logic. This type of optimization will be one of the main task of future HTS systems because the largest part of the cost of these kind of systems is related to ground segment development and management. This is the reason why in the last few years, the interest of satellite operators and scientific community in SGD is rapidly growing; in general, the goal of the research is to tailor SDG scheme for a particular HTS system so that to guarantee a target feeder-link service availability and system capacity, reducing the number of GWs. Many recent scientific works have been focused on the design of HTS ground segment exploiting both EHF and optical bands. del Portillo et al. [66] Reports the optimization of ground segment architecture for large LEO constellations using EHF feeder links (Ka, Q/V and W bands), with the objective to reduce the number of GWs and maximize the capacity; another interesting work is [67] where an optimization of a Q/V-band feeder link with the objective to reduce the number of GWs and the cost of terrestrial links for traffic routing is presented. Feeder links based on Free-Space Optics (FSO) have been analyzed in [68–71], with the objective to: find the best optical gateway positioning (with respect to tropospheric propagation), reduce the number of GWs and optimize their combined use. The problem of satellite gateway placement for network reliability maximization has been analyzed using standard optimization algorithms in [72, 73]. Many research works have been also focused on the development of multiple gateways outage modeling with the objective to create tools that can be used by satellite operators to optimize SGD system design [52, 74]. Two different strategies for the SGD architecture optimization, given a set of constraints (geographical position of all available gateways, characteristics of satellite/ground terminals, target system availability, target throughput, etc.), based on the concept of GWs clustering is presented in [75]. There are two main implementation of SGD schemes, the so called N + P [74] and N + 0 [52, 76]. The first one is based on the use of N active GWs and P redundant GWs (that are in warm standby); these are activated when one (or more) feeder-link of an active GW experiences a strong fading level and is not able to guarantee the required traffic capacity. The second scheme is based on N active GWs working in capacity load-sharing; this means that they have a spare capacity that can be used when one or more GWs are in outage. For this scheme all the GWs are always active and have been designed to have the required additional spare capacity to satisfy system requirements. In order to support a SGD scheme, it is required to increase complexity in the satellite payload because it should guarantee the possibility to illuminate the same user beams from multiple GWs, using an on-board reconfigurable switching matrix. Hence, for traditional analog transparent payloads, the SG schemes do not foresee the possibility of a full traffic rerouting between all the GWs because this would require a high on-board routing complexity while, in general, the rerouting is allowed between a sub-group of GWs. This leads to the

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

31

need for a strong optimization of SG diversity scheme parameters. On the other hand, for next generation of HTS systems there are digital payload developments that provide a more flexible beam-to-beam connectivity, relaxing the limitation of analog ones but increasing power consumption and complexity of the payload.

1.5.2 Feeder Link MIMO Aiming at supporting HTS and VHTS services, also driven by the recent integration of a non-terrestrial component into the 5G (and 6G) infrastructure, it is already understood that a further leap in the achievable capacity is needed to cope with the massive demands from future enhanced Mobile Broadband (eMBB) services. These payloads are expected to generate more hundreds of spot beams, each allocated even hundreds of Mhz (e.g., 3GPP NTN assume 400 MHz in Ka-band): it is thus clear that feeder links might become the system bottleneck. Recently, to tackle this challenge, some research activities focused on the implementation of MIMO solutions on the feeder link combined with the exploitation of higher carrier frequency, i.e., Q/V (40–50 GHz) and W (70–95 GHz) bands, [77– 80]. In this case, combining the high directivity of the feeder link antennas and the high frequency, the links are predominantly LOS. As such, it is possible to achieve the maximum MIMO capacity by meeting specific phase relationships between the signals impinging at the receiver antenna. Two design approaches have been explored so far, both for GEO systems: (1) a payload equipped with two multibeam reflector antennas, separated by a few meters, and multiple gateways per site; and (2) a payload equipped with a single multi-beam reflector antenna and a single gateway per site. In general, in both approaches the SGD principles depicted in the previous section can be applied. The main findings are discussed below: • Multi-GW per site: each GW site spans several tens of km and includes two GWs, which are managed by a central processing unit. Phase and time synchronisation are guaranteed by the use of RF-over-fiber transport, [81]. It was designed by NASA for deep space applications with antennas separated by more than tens of km; as such, it can be assumed to provide the required, tight, synchornisation performance on-ground. The satellite is then equipped with two reflector antennas, separated by a few meters. Thus, the receiving antenna elements are not colocated, in order to better exploit the spatial diversity principle. The interference analysis performed in Q/V-band shows that the proposed MIMO architecture yields a gain of at least 3 dB in the CIR over half of the links compared to legacy systems. A detailed overview of the possible payload architectures is available in [78]. In general, for a fixed amount of data to be provided on the user link, it was observed that a reduced number of feeder links is required with MIMO compared to legacy Single Input Single Output (SISO) systems. In addition, the combination of feeder and user link MIMO algorithms yields significant gains, even up to .110% in the rate per beam compared to 4-colours schemes.

32

E. Cianca et al.

• Single-GW per site: a single Network Control Center (NCC) is connected to the GWs and ensures the overall time and phase synchronisation. This can be achieved again by means of RF-over-fiber transport, [81], as discussed above. The NCC is in charge of managing the distributed linear precoding algorithm and re-routing the traffic through the GWs that are currently active. To this aim, it shall provide the following functions: (1) storage of the CSI vector for each active GW; (2) computation of the precoding matrix based on the stored CSI vectors and set of active GWs; (3) transmission of either the precoded signals or of the nonprecoded signals together with the precoding coefficients; (4) transmission to the satellite of the list of currently active GWs, i.e., the beamforming coefficients for the receiver on the feeder link; and (5) overall system synchronization. The performance was assessed in [80] in Ku, Ka, Q/V, and W bands with ZF and MMSE precoding, considering SPC, MPC, and PAC normalizations. Moreover, the impact of imperfect CSI at the ground control segment and of different service availabilities was detailed. In general, MMSE-PAC solutions are the most promising with capacity gains that can be as large as .200% compared to a nonprecoded system with frequency reuse 3. An interesting outcome of this study is that, in W-band, the significantly directive radiation patterns on-board the satellite provide a significant interference rejection and, thus, the exploitation of EHF bands with FFR can already be implemented. In fact, asymptotically the precoded and non-precoded FFR schemes have the same performance. The advantage compared to multi-colour frequency reuse schemes is significant in all scenarios. Based on the above studies, MIMO is a promising solutions for the feeder links as well, allowing to: (1) reduce the number of required GWs; (2) better tackling the impact of the atmospheric events; and (3) increasing the capacity on the feeder link. The challenges in this area include the availability of the CSI vectors, and their reliability, in particular when MIMO is implemented in a distributed approach through multiple GWs. Moreover, the development of suitable transmission and channel coding formats for multi-gigabit satellite applications is still under investigation; in fact, the theoretical performance improvement yielded by new waveforms is often subject to trade-offs with the hardware feasibility of the modem chain, [7]. Finally, it is worth mentioning that the channel models, in particular related to the atmospheric impairments, are still approximated; more accurate models will be needed to further assess the feasibility of MIMO in the EHF bands.

1.6 On-Going/Planned Mission/Services and Mega-Constellations The study of the actual behavior of the mmWave travelling through the troposphere play a fundamental role in the perspective of exploiting EHF for satellite commu-

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

33

Fig. 1.10 The “Aldo Paraboni” payload antenna farm. Courtesy of AIRBUS Italia SPA, https://www.esa. int/Applications/ Telecommunications_ Integrated_Applications/ Alphasat/

nications. The main missions and activities on the transmission of EHF signals by operational satellite links are reported hereafter [44]. Alphasat Communications Experiment: The “Aldo Paraboni” Payload Based on the experience acquired by the Italian Space Agency (ASI) in the analysis and verification of Ka-band and Q/V band propagation in the framework of the Italsat F1 mission, in 2006, ASI proposed to the European Space Agency (ESA) to host an experimental payload in Q/V band on board the Alphasat GEO satellite, successfully launched in July 2013. The installed Technology Demonstration Payload number 5 (TDP#5) was named in honor of professor Aldo Paraboni (1940– 2013) of Politecnico di Milano (Italy), pioneer of research about the use of high frequencies in satellite communications [82]. The Q/V band TDP#5 antenna farm is shown in Fig. 1.10; two self-standing payloads are present, namely: Communication experiment payload, aimed at designing, optimizing and testing the effectiveness of adaptive transmission schemes, i.e., Propagation Impairment Mitigation Techniques (PIMTs), over the Q/V band satellite channel. Experimented PIMTs are ACM, uplink power control (UL-PC) and space diversity. Modulated signals are transmitted to the Earth stations located in Spino D’Adda (Italy), Tito Scalo (Italy) and Graz (Austria). The Principal Investigator of communication experiment is University of Rome Tor Vergata. Main target of the propagation experiment payload is to transmit an unmodulated carrier over a European coverage to derive first and second-order statistics of tropospheric attenuation, including rain fading, cloud fading and scintillations. Other interesting measurements are related to the acquisition of sky-noise temperature and the study of the parameters measured during the concurrent communication experiment. The Principal Investigator of propagation experiments is Politecnico di Milano. More details about the payloads’ features and the measurement approach followed by the communication and propagation experiments can be found in [83–85].

34

E. Cianca et al.

AEHF Satellite Constellation The Advanced Extreme High Frequency (AEHF) satellite program is a follow-up of US Milstar military satellite system designed in the 1990s [86]. AEHF has been launched in 2013, consisting of four satellites in Geostationary Earth Orbit and embarking 44 GHz EHF uplink transmitter, while the downlink is transmitting in the SHF-band (20 GHz frequency). AEHF satellites use a large number of narrow spot beams directed towards the Earth to relay communications to and from users. Crosslinks between the satellites allow them to relay communications directly rather than via a ground station. The satellites are designed to provide jam-resistant communications with a low probability of interception. They incorporate frequencyhopping radio technology, as well as phased array antennas that can adapt their radiation patterns in order to block out potential sources of jamming [86]. The maximum user data rate supported by AEHF is 8.2 Mb/s, allowing transmission of real-time videos, battlefield maps, targeting data, etc., while the total capacity provided is 430 Mb/s [87]. User terminals include mobile terminals, ship and submarine terminals, and airborne terminals used by all of the Services and AEHF international partners which are Canada, the Netherlands and the United Kingdom. ASI-DAVID Data Collection Experiment (DCE) The Data and Video Interactive Distribution (DAVID) is a satellite mission of ASI [88] that completed Phase B in 2003. The C/D phase proposal has been delivered to ASI. The Data Collection Experiment (DCE) of DAVID aims at proving the feasibility of high-speed data collection (100 Mb/s) from a content provider station via 84.5 GHz W-band link. DAVID satellite is designed as a Sun-synchronous lowearth orbit satellite with orbit altitude of 570 km, ensuring the re-visitation of the same site at the same hour every day. Details of the physical layer design of DAVIDDCE can be found in [52]. DAVID mission is currently on hold with the perspectives to be the basis for follow up activities. W-Band Analysis and Verification (WAVE) WAVE is an ASI project that can be regarded as the follow-up of DAVID. WAVE has been conceived to make possible the deployment of operative LEO/GEO satellites in the W-band for commercial services in 15–20 years [89, 90]. The project, whose A2 Phase has been successfully concluded in 2008, includes two demonstrative missions and two pre-operative missions. The demonstrative missions are: AeroWAVE, based on the use of High-Altitude Platforms, that will give preliminary results for the channel characterization and IKNOW, based on a small LEO, that will provide first-order statistics of satellite channel [89]. The two preoperative missions—one over LEO and the other one over GEO—will exploit the results achieved by demonstration missions in order to perform preliminary tests for operative use of W-band. Like DAVID, also WAVE is in a frozen status, waiting for Alphasat achievements.

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

35

W-Cube The development of the W-Cube was undertaken in the project ESA ARTES— Advanced Technology CubeSat-based W-band channel measurements. The project was run by Joanneum Research as project coordinator. The mission uses a 75 GHz carrier transmitting from space to a ground station receiver in Graz, Austria, to study, how atmospheric weather and space weather phenomena affect radio signals in the EHF frequency range. During the first two successful completions of project phases, the nanosatellite and a corresponding ground station were designed and built, which are collating data for using the Q and W band from LEO in the third phase. Fraunhofer IAF developed the high-frequency front-ends for the satellite and the ground station, including Medium Power Amplifiers (MPAs), LNAs, frequency multipliers and mixers. W-Cube was launched on 30th of June 2021 from Cape Canaveral, Florida as part of a 88 satellite launch on the Transporter-2 mission using the Falcon 9 rocket. Non-Geostationary-Satellite-Constellations at EHF Several companies such as SpaceX, OneWeb, Telesat, O3b Networks and Boeing applied to the FCC their plans to deploy constellations of V-band satellites in nongeosynchronous orbits and provide commercial communications services. SpaceX proposes a VLEO constellation consisting of two sub-constellations of space stations and associated ground control facilities, gateway earth stations and end user Earth stations, [91]. The first LEO constellation comprises the 4425 satellites operating in the Ku and Ka bands while the second component of the SpaceX V-band system is the VLEO constellation composed by 7518 satellites, each of which will occupy unique orbital planes in VLEO. The SpaceX System will use phased array beam-forming and digital processing technologies onboard each satellite payload supporting the efficient and flexible use of spectrum resources. The broadband services will be available for residential, commercial, institutional, governmental and professional users worldwide. SpaceX has designed its V-band system to meet the dual requirements of the world’s broadband demand—namely, connectivity for rural, remote and hard-to-reach end-users, as well as efficient, highcapacity connectivity at all locations. The expansion plan of OneWeb constellation by 2000 satellites included a “subconstellation” of 720 LEO V-band satellites at 1200 km, and another constellation in Medium Earth Orbit (MEO) of 1280 satellites for commercial communications services. OneWeb would implement the dynamic traffic assignment between the LEO and MEO V-band constellations based on service requirements and the data traffic within coverage areas. This OneWeb plan on MEO development in 2017 follows that of ViaSat for 24 MEO satellites to augment the HTS ViaSat-3.

36

E. Cianca et al.

1.7 Final Remarks In this first chapter, we explored the technologies and architectures, based on (V)HTS and mmWave, that are expected to play a fundamental role in the design and implementation of the next generation space systems. In particular, we provided a multifaceted overview of SDN-based payloads, future air interfaces, advanced system-level capacity enhancements, spectrum allocations and DSA, and the required evolutions on the feeder links. Moreover, we also reported some of the main on-going and planned missions operating in EHF. All of the discussed solutions revolve around the concepts of flexibility, reconfigurability, and sustainability aimed at making the “connecting the unconnected” a reality in the near future. In the next chapters of this visionary book, many of the above introduced concepts will be further detailed and expanded.

References 1. M. Ruggieri, S. De Fina, M. Pratesi, A. Salome’, E. Saggese, C. Bonifazi, The W-band data collection experiment of the DAVID mission. IEEE Trans. AES 38(4), 1377–1387 (2002) 2. T. Rossi, M. De Sanctis, E. Cianca, C. Fragale, M. Ruggieri, H. Fenech, Future space-based communications infrastructures based on High Throughput Satellites and Software Defined Networking, in IEEE International Symposium on Systems Engineering (ISSE)(2015), pp. 332– 337. https://doi.org/10.1109/SysEng.2015.7302778 3. Y. Guan, F. Geng, J. H. Saleh, Review of high throughput satellites: market disruptions, affordability-throughput map, and the cost per bit/second decision tree, in IEEE AES Magazine (2019), pp- 64–80. https://doi.org/10.1109/MAES.2019.2916506 4. I. Philbeck, Connecting the unconnected: working together to achieve connect 2020 agenda targets, in Broadband Commission and the World Economic Forum at Davos Annual Meeting 2017, ITU (2017). Background paper to the special session 5. https://www.itu.int/en/ITU-D/Statistics/Pages/stat/default.aspx 6. ESA, First W-band transmission from space (2021). https://artes.esa.int/news/first-wbandtransmission-space 7. C. Sacchi, T. Rossi, M. Ruggieri, F. Granelli, Efficient waveform design for high-bit-rate W-band satellite transmissions. IEEE Trans. Aerosp. Electron. Syst. 47(2), 974–995 (2011). https://doi.org/10.1109/TAES.2011.5751238 8. J. Nessel, R. Acosta, F.A. Miranda, Preliminary experiments for the assessment of V/Wband links for space-earth communications, in Proceedings AP-S International Symposium, IEEE International Symposium on Antennas and Propagation & USNC/URSI National Radio Science Meeting (2013). https://doi.org/10.1109/APS.2013.6711467 9. M.B. Jiménez, D. Fernandez, J.E. Rivadeneira, L. Bellido, A. Cardenas, A survey of the main security issues and solutions for the SDN architecture, in IEEE Access (2021), pp. 122016– 122038. https://doi.org/10.1109/ACCESS.2021.3109564 10. T. Rossi, C. Fragale, M. De Sanctis, E. Cianca, M. Ruggieri, H. Fenech, Software defined networking and high throughput satellite: the best matching for space-based communications infrastructures, in 21st Ka and Broadband Communications Conference, Bologna (Italy), On line Proceedings (2015). ISSN -2573–6124 11. M. Ruggieri, T. Rossi, New fascinating challenges for space systems: softwarization, ai-based robotization and sustainability: which role for cubeSats? in Advances in the Astronautical Sciences, vol. 173 (2020), pp. 609–615. ISBN 978-0-87703-671-5, paper AAS 20-268, ISSN

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

37

10816003, 00653438. Proceedings for the 5th IAA Conference on University Satellite Missions and CubeSat Workshop, 28–31 January 2020, Rome, Italy 12. Y. Bi, G. Han. S. Xu, X. Wang, C. Lin, Z. Yu, P. Sun, Software defined space-terrestrial integrated networks: architecture, challenges, and solutions. IEEE Netw. 22–28 (2019). https:// doi.org/10.1109/MNET.2018.1800193 13. A. Papa, T. de Cola, P. Vizarreta, M. He, C. Mas-Machuca, W. Kellerer, Design and evaluation of reconfigurable SDN LEO constellations. IEEE Trans. Netw. Serv. Manag. 17(3), 1432–1445 (2020). https://doi.org/10.1109/TNSM.2020.2993400 14. H. Maune, K. H. Hubner, G. Gold, Considerations for V- and W-band inter-satellite links, in IEEE 22nd Annual Wireless and Microwave Technology Conference (WAMICON) (2022). https://doi.org/10.1109/WAMICON53991.2022.9786212 15. A.U. Chaudhry, H. Yanikomeroglu, Temporary laser inter-satellite links in free-space optical satellite networks. IEEE Open J. Commun. Soc. 3, 1413–1445 (2022). https://doi.org/10.1109/ OJCOMS.2022.3198391 16. https://www.copernicus.eu/en 17. K. Han, B. Xu, F. Shao, W. Gong, Q. Ren, J. Chang, An adaptive topology optimization strategy for inter-satellite links in GNSS. IEEE Trans. Aerosp. Electron. Syst. 10(9), 1–12 (2021). https://doi.org/10.1109/TAES.2022.3181555 18. A. Le Pera, F. Forni, M. Grossi, M. Lucente, V. Palma, T. Rossi, M. Ruggieri, Digital transparent processor for satellite telecommunication services, in IEEE Aerospace Conference (2007), pp. 1–9. https://doi.org/10.1109/AERO.2007.352932 19. V. Sulli, D. Giancristofaro, F. Santucci, M. Faccio, G. Marini, Design of digital satellite processors: from communications link performance to hardware complexity. IEEE J. Sel. Areas Commun. 36(2), 338–350 (2018). https://doi.org/10.1109/JSAC.2018.2804198 20. ITU-R Report S.2461-0. spectrum needs for the fixed-satellite services in the 51.4-52.4 GHz band (2019) 21. ITU-R Recommendation V.431-8. Nomenclature of the frequency and wavelength bands used in telecommunications (2015) 22. ITU Radio Regulations, Articles, Edition of 2021 23. ECC ERC, Decision of 19 October 2000 on the Shared Use of the Band 17.7–19.7 GHz by the Fixed Service and Earth Stations of the Fixed-Satellite Service (Space-to-Earth) (2000) 24. Y.-C. Liang, Dynamic Spectrum Management - From Cognitive Radio to Blockchain and Artificial Intelligence, 1st edn. (Springer Singapore, 2020) 25. S.K. Sharma, S. Chatzinotas, B. Ottersten, Cognitive radio techniques for satellite communication systems, in 2013 IEEE 78th Vehicular Technology Conference (VTC Fall) (2013), pp. 1–5 26. M. H´’oyhty´’a et al., Database-assisted spectrum sharing in satellite communications: a survey. IEEE Access 5, 25322–25341 (2017) 27. S.K. Sharma, S. Chatzinotas, B. Ottersten, In-line interference mitigation techniques for spectral coexistence of GEO and NGEO satellites. Int. J. Sat. Commun. Netw. 31(1), (2016) 28. V. Icolari, A. Guidotti, D. Tarchi, A. Vanelli-Coralli, An interference estimation technique for Satellite cognitive radio systems, in 2015 IEEE International Conference on Communications (ICC) (2015), pp. 892–897. https://doi.org/10.1109/ICC.2015.7248435 29. S. Chatzinotas, B. Evans, A. Guidotti, V. Icolari, E. Lagunas, S. Maleki, S.K. Sharma, D. Tarchi, P. Thompson, A. Vanelli-Coralli, Cognitive approaches to enhance spectrum availability for satellite systems. Int. J. Sat. Commun. Netw. 35(5), 407–442 (2016) 30. G. Zheng, S. Chatzinotas, B. Ottersten, Generic optimization of linear precoding in multibeam satellite systems. IEEE Trans. Wirel. Commun. 11(6), 2308–2320 (2012). https://doi.org/10. 1109/TWC.2012.040412.111629 31. P. Arapoglou, K. Liolis, M. Bertinelli, A. Panagopoulos, P. Cottis, R. De Gaudenzi, MIMO over satellite: a review. IEEE Commun. Surv. Tutorials 13(1), 27–51 (2011). https://doi.org/10. 1109/SURV.2011.033110.00072

38

E. Cianca et al.

32. P. Angeletti, R. De Gaudenzi, A pragmatic approach to massive MIMO for broadband communication satellites. IEEE Access 8, 132212–132236 (2020). https://doi.org/10.1109/ ACCESS.2020.3009850 33. R. Muharar, J. Evans, Downlink Beamforming with transmit-side channel correlation: a large system analysis, in 2011 IEEE International Conference on Communications (ICC) (2011), pp. 1–5 . https://doi.org/10.1109/icc.2011.5962672 34. A. Guidotti, A. Vanelli-Coralli, Design trade-off analysis of precoding multi-beam satellite communication systems, in 2021 IEEE Aerospace Conference (2021), pp. 1–12. https://doi. org/10.1109/AERO50100.2021.9438169 35. A. Guidotti, C. Amatetti, F. Arnal, B. Chamaillard, A. Vanelli-Coralli, Location-assisted precoding in 5G LEO systems: architectures and performances, in 2022 Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit) (2022), pp. 154–159. https://doi.org/10.1109/EuCNC/6GSummit54941.2022.9815611 36. EC Horizon 2020 SPACE-29-TEC-2020 Project DYNASAT, Dynamic Spectrum Sharing and Bandwidth-Efficient Techniques for High-Throughput MIMO Satellite Systems. Deliverable D3.2 “Bandwidth Efficient Techniques Design” (2022) 37. EC Horizon 2020 SPACE-29-TEC-2020 Project DYNASAT, Dynamic Spectrum Sharing and Bandwidth-Efficient Techniques for High-Throughput MIMO Satellite Systems. Deliverable D3.5 “Bandwidth Efficient Techniques Evaluation” (2022) 38. V. Joroughi et al., Designing joint precoding and beamforming in a multiple gateway multibeam satellite system, in IEEE WCNC 2018 (2018) 39. C. Mosquera et al., Distributed precoding systems in multi-gateway multibeam satellites: regularization and coarse beamforming. IEEE Trans. Commun. 17(10), (2018) 40. G. Zheng, S. Chatzinotas, B. Ottersten, Multi-gateway cooperation in multibeam satellite systems, in IEEE PIMRC 2012 (2012) 41. A. Guidotti, A. Vanelli-Coralli, Clustering strategies for multicast precoding in multibeam satellite systems. Int. J. Sat. Commun. Netw. 38, 85–104 (2019) 42. DVB Document A171-2, Implementation guidelines for the second generation system for broadcasting, interactive services, news gathering and other broadband satellite applications, Part 2: S2 Extensions (DVB-S2X) (2020) 43. J.H. Puneeth et al., Satellite Beam Densification for High-Demand Areas, in ASMS/SPSC 2022 (2022) 44. E. Cianca, T. Rossi, A. Yahalom, Y. Pinhasi, J. Farserotu, C. Sacchi, EHF for satellite communications: the new broadband frontier. Proc. IEEE 99(11), 1858–1881 (2011). https:// doi.org/10.1109/JPROC.2011.2158765 45. B.F. Beidas, Radio-frequency impairments compensation in ultra high-throughput satellite systems. IEEE Trans. Commun. 67(9), 6025–6038 (2019). https://doi.org/10.1109/TCOMM. 2019.2926031 46. M. De Sanctis, E. Cianca, T. Rossi, C. Sacchi, L. Mucchi, R. Prasad, Waveform design solutions for EHF broadband satellite communications. IEEE Commun. Mag. 53(3), 18–23 (2015). https://doi.org/10.1109/MCOM.2015.7060477 47. 3GPP, Solutions for NR to support non-terrestrial networks (NTN). TR 38.821 v16.0.0 (2020) 48. R. Mulinde, T.F. Rahman, C. Sacchi, Constant-envelope SC-FDMA for nonlinear satellite channels, in Proceedings of IEEE Global Communications Conference (GLOBECOM) (2013), pp. 2939–2944. https://doi.org/10.1109/GLOCOM.2013.6831521 49. M. De Sanctis, T. Rossi, L. Rizzo, M. Ruggieri, G. Codispoti, Optimization of ACM algorithms over Q/V-band satellite channels with the Alphasat Aldo Paraboni P/L, in Proceedings of IEEE Aerospace Conference (2015), pp. 1–9 50. R. He, D. Yang, H. Wang, J. Kuang, Adaptive hierarchical coding and modulation scheme over satellite channels. IET Commun. 13(17), 2834–2839 (2019) 51. S. Mukherjee, M. De Sanctis, T. Rossi, E. Cianca, M. Ruggieri, R. Prasad, Mode switching algorithms for DVB-S2 links in W band, in IEEE Global Telecommunications Conference GLOBECOM (2010), pp. 1–5. https://doi.org/10.1109/GLOCOM.2010.5683138

1 Millimeter Waves and High-Throughput Satellites: The New Frontier. . .

39

52. T. Rossi, M. De Sanctis, F. Maggio, Evaluation of outage probability for satellite systems exploiting smart gateway configurations. IEEE Commun. Lett. 21(7), 1541–1544. 53. S. Gu, J. Jiao, Z. Huang, S. Wu, Q. Zhang, ARMA-based adaptive coding transmission over millimeter-wave channel for integrated satellite-terrestrial networks. IEEE Access 6, 21635– 21645 (2018). https://doi.org/10.1109/ACCESS.2018.2825256 54. M. Shirvanimoghaddam, Y. Li, B. Vucetic, Near-capacity adaptive analog fountain codes for wireless channels. IEEE Commun. Lett. 17(12), 2241–2244 (2013). https://doi.org/10.1109/ LCOMM.2013.101813.131972 55. C. Xu et al., Adaptive coherent/non-coherent single/multiple-antenna aided channel coded ground-to-air aeronautical communication. IEEE Trans. Commun. 67(2), 1099–1116 (2019). https://doi.org/10.1109/TCOMM.2018.2877471 56. G.D. Surabhi, M.K. Ramachandran, A. Chockalingam, OTFS modulation with phase noise in mmWave communications, in IEEE 89th vehicular technology conference (VTC2019-Spring) (2019), pp. 1–5. https://doi.org/10.1109/VTCSpring.2019.8746382 57. Z. Wei, W. Yuan, S. Li, J. Yuan, G. Bharatula, R. Hadani, L.H. Hanzo, Orthogonal timefrequency space modulation: a promising next-generation waveform. IEEE Wirel. Commun. 28, 136–144 (2021) 58. T. Li, R. He, B. Ai, M. Yang, Z. Zhong, H, Zhang, OTFS modulation performance in a satelliteto-ground channel at sub-6-GHz and millimeter-wave bands with high mobility. Front. Inf. Technol. Electron. Eng. 22, 517–526 (2021). https://doi.org/10.1631/FITEE.2000468 59. B.F. Beidas, Intermodulation distortion in multicarrier satellite systems: analysis and turbo Volterra equalization. IEEE Trans. Commun. 59(6), 1580–1590 (2011) 60. E. Kofidis, V. Dalakas, Y. Kopsinis, S. Theodoridis, A novel efficient cluster-based MLSE equalizer for satellite communication channels with M-QAM signaling. EURASIP J. Appl. Signal Process. pp. 1–16 (2006) 61. S. Dimitrov, Non-linear distortion cancellation and symbol-based equalization in satellite forward links. IEEE Trans. Wirel. Commun. 16(7), 4489–4502 (2017) 62. B.F. Beidas, R.I. Seshadri, N. Becker, Multicarrier successive predistortion for nonlinear satellite systems. IEEE Trans. Commun. 63(4), 1373–1382 (2015) 63. S. Cioni et al., Transmission parameters optimization and receiver architectures for DVB-S2X systems. Int. J. Satell. Commun. Netw. 34(3), 337–350 (2016) 64. A. Ugolini, A. Modenini, G. Colavolpe, G. Picchi, V. Mignone, A. Morello, Advanced techniques for spectrally efficient DVB-S2X systems, in Proceedings 7th Advanced Satellite Multimedia Systems Conference, 13th Signal Processing Space Communications Workshop (2014), pp. 158–164 65. H. Skinnemoen, Gateway diversity in Ka-band systems, in 4th Ka-Band Utilization Conference (1998) 66. I. del Portillo, B. Cameron, E. Crawley, Ground segment architectures for large LEO constellations with feeder links in EHF-bands, in 2018 IEEE Aerospace Conference (2018) 67. B.G. Evans, P. Thompson, C. Baudoin, M. Dervin, Gateway diversity for a future high throughput satellite system, in 32nd AIAA International Communications Satellite Systems Conference, AIAA SPACE Forum (2014) 68. F. Lacoste, A. Guérin, A. Laurens, G. Azema, C. Periard, D. Grimal, FSO ground network optimization and analysis considering the influence of clouds, in Proceedings of the 5th European Conference on Antennas and Propagation (EUCAP) (2011), pp. 2746–2750 69. S. Poulenard, M. Crosnier, A. Rissons, Ground segment design for broadband geostationary satellite with optical feeder link. IEEE/OSA J. Opt. Commun. Netw. 7(4), 325–336 (2015) 70. C. Fuchs, F. Moll, Ground station network optimization for space-to-ground optical communication links. IEEE/OSA J. Opt. Commun. Netw. 7(12), 1148–1159 (2015) 71. I. del Portillo, M. Sanchez, B. Cameron, E. Crawley, Architecting the ground segment of an optical space communication network, in 2016 IEEE Aerospace Conference (2016), pp. 1–13 72. Y. Cao, L. Zhao, Y. Shi, J. Liu, Gateway placement for reliability optimization in 5Gsatellite hybrid networks, in 2018 International Conference on Computing, Networking and Communications (ICNC) (2018), pp. 372–376

40

E. Cianca et al.

73. R. Dhaou et al., Gateway selection optimization in hybrid MANET-satellite network, in Wireless and Satellite Systems (Springer International Publishing, New York, 2015), pp. 331– 344 74. A. Gharanjik, B. Shankar, P.D. Arapoglou, B. Ottersten, Multiple gateway transmit diversity in Q/V band feeder links. IEEE Trans. Commun. 63(3), 916–926 (2015) 75. T. Rossi, M. De Sanctis, F. Maggio, M. Ruggieri, C. Hibberd, C. Togni, Smart gateway diversity optimization for EHF satellite networks. IEEE Trans. Aerosp. Electron. Syst. 56(1), 130–141 (2019) 76. A. Kyrgiazos, B.G. Evans, P. Thompson, On the gateway diversity for high throughput broadband satellite systems. IEEE Trans. Wirel. Commun. 13(10), 5411–5426 (2014) 77. T. Delamotte, R.T. Schwarz, K. Storek, A. Knopp, MIMO feeder links for high throughput satellites, in WSA 2018; 22nd International ITG Workshop on Smart Antennas (2018), pp. 1–8 78. T. Delamotte, K. Storek, A. Knopp, MIMO processing for satellites in the 5G era, in 2019 IEEE 2nd 5G World Forum (5GWF) (2019), pp. 629–635. https://doi.org/10.1109/5GWF.2019. 8911646 79. R.T. Schwarz, T. Delamotte, K. Storek, A. Knopp, MIMO applications for multibeam satellites. IEEE Trans. Broadcast. 65(4), 664–681 (2019). https://doi.org/10.1109/TBC.2019.2898150 80. A. Guidotti, C. Sacchi, A. Vanelli-Coralli, Feeder link precoding for future broadcasting services: architecture and performance. IEEE Trans. Aerosp. Electron. Syst. https://doi.org/ 10.1109/TAES.2022.3144243 81. K.Y. Lau, G.F. Lutes, R.L. Tjoelker, Ultra-stable RF-over-fiber transport in NASA antennas, phased arrays and radars. J. Lightwave Technol. 32(20), 3440–3451 (2014). https://doi.org/10. 1109/JLT.2014.2312930 82. G. Codispoti, G. Parca, M. Ruggieri, T. Rossi, M. De Sanctis, C. Riva, L. Luini, The role of the Italian space agency in investigating high frequencies for satellite communications: the alphasat experiment. Int. J. Sat. Commun. Netw. 37(5), 387–396 (2018) 83. T. Rossi, M. De Sanctis, F. Maggio, M. Ruggieri, G. Codispoti, G. Parca, Q/V-band satellite communication experiments on channel estimation with Alphasat Aldo Paraboni P/L, in 2015 IEEE Aerospace Conference (2015), pp. 1–11 84. T. Rossi et al., Satellite communication and propagation experiments through the alphasat Q/V band Aldo Paraboni technology demonstration payload. IEEE Trans. Aerosp. Electron. Syst. 31(3), 18–27 (2016) 85. R. Polonio, C. Riva, ITALSAT propagation experiment at 18.7, 39.6 and 49.5 GHz at Spino D’Adda: three years of CPA statistics. IEEE Trans. Antennas Propag. 46(5), 631–635 (1998) 86. J.A. Haulund, R. Khullar, Advanced extremely high frequency mission planning: will its legacy dictate its future? in 2012 IEEE First AESS European Conference on Satellite Telecommunications (ESTEL) (2012), pp. 1–5 87. M. Sarchet, AEHF Update, Presentation to AFCEA Los Angeles Chapter, Los Angeles (CA) [Online]. Accessed 22 Mar. 2012. Available: https://afcea-la.org/ 88. M. Ruggieri, S. De Fina, M. Pratesi, E. Saggese, C. Bonifazi, The W-band data collection experiment of the DAVID mission. IEEE Trans. Aerosp. Electron. Syst. 38(4), 1377–1387 (2002) 89. M. Lucente et al., Experimental missions in W-Band: a small LEO satellite approach. IEEE Syst. J. 2(1), 90–103 (2008) 90. A. Jebril et al., The WAVE mission payload, in 2005 IEEE Aerospace Conference (2005), pp. 903–912 91. FCC FACT SHEET, SpaceX V-band authorization memorandum opinion, order and authorization–IBFS. File No. SAT-LOA-20170301-00027 [Online]. Available: https://docs.fcc. gov/public/attachments/DOC-354775A1.pdf

Chapter 2

The Role of Satellite in 5G and Beyond Mario Marchese, Fabio Patrone, and Alessandro Guidotti

2.1 Non-Terrestrial Networks Standardization Recognising the added value that a non-terrestrial component would bring to the overall 5G communication infrastructure, a few years ago (within Rel. 15) 3GPP started to design and assess Non-Terrestrial Network (NTN) systems. Such integration is expected to, [1–6]: (1) complement 5G services in under-/un-served areas; (2) improve the 5G service reliability and continuity for Internet of Things (IoT) devices or Mission Critical services; and (3) enable the 5G network scalability by means of efficient multicast/broadcast resources for data delivery. In this Section, we provide an overview of the 3GPP standardisation activities related to NTN, extensively discuss the Radio Access Network (RAN) architecture options and the related challenges, and provide a glimpse at the path that is being defined for Beyond 5G (B5G), i.e., 5G-Advanced (5G-A) and 6G, NTN systems.

2.1.1 3GPP Release 17: The First 5G NTN-Based Standard Within 3rd Generation Partnership Project (3GPP) Rel. 15, two Study Items (SI) for NTN-based 5G systems were initiated addressing the deployment scenarios and

M. Marchese () · F. Patrone Dipartimento di ingegneria navale, elettrica, elettronica e delle telecomunicazioni (DITEN), University of Genova, Genova, Italy e-mail: [email protected]; [email protected] A. Guidotti Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), University of Bologna, Bologna, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_2

41

42

M. Marchese et al.

Fig. 2.1 Number of documents submitted to 3GPP RAN, SA, and CT meetings related to NTN per TSG (left) and per TSG WG (right)

the related challenges. Building on the outcomes of these analyses, two Work Items (WI) were approved, and then finalised, for Rel. 17 aiming at: (1) consolidating the preliminary performance assessment and potential impacts on the Physical (PHY) and Medium Access Control (MAC) layers; (2) analysing aspects related to Layers 2 and 3, including handover and Dual Connectivity (DC); and (3) identify potential requirements for the upper layers. In addition, also Narrow-Band IoT (NB-IoT) entered into the normative phase since Rel. 17; related to NB-IoT, in this chapter we provide some information on the type of service, with the general focus mainly on NTN. To understand the outstanding effort that has been put into the development of a NTN-based 5G and beyond network infrastructure, Fig. 2.1 shows the number of documents that were submitted for discussion during the Radio Access Network (RAN) and Service and system Aspects meetings between 2016 and 2022, at the time of writing this chapter;1 in this figure, we are considering the Radio Access

1 The values reported in the figure have been obtained by web scraping the public information available on the 3GPP website in October, 2022, with a tool proprietary of the University of Bologna. As such, the exact values might be subject to variations, not impacting the general trends.

2 The Role of Satellite in 5G and Beyond

43

Network (RAN), Service and system Aspects (SA), and Core Network & Terminals (CT) Technical Specification Groups (TSG). In 2022, 3GPP finalised the specifications for Rel. 17 and initiated the studies for Rel. 18, the first release covering 5G-Advanced (5G-A) networks. The path to the revolutionary first standard integrating non-terrestrial and terrestrial communications is summarised below: • In Rel. 15, two SIs under RAN1 and SA2 where developed: (1) “Study on NR to support NTN,” within RAN; and (2) “Study on using Satellite Access in 5G,” in the framework of SA2. In the former SI, the objectives were to identify the deployment scenarios and the related system parameters, including the potential key impact areas for the NR standard; potential solutions to the key issues identified in this activity were also proposed. In the latter SI, the focus was on defining the use cases for the integration of the NTN component in the NR architecture and on identifying the potential services, which were then categorised for service continuity, ubiquity, and scalability. • Rel. 16, building on the outcomes of Rel. 15, witnessed the first WI for NTN on “Integration of Satellite Access in 5G,” supervised by SA1, which was then completed under Rel. 17. Moreover, three additional SIs were initiated: (1) “Study on architecture aspects for using satellite access in 5G,” under SA2; (2) “Study on management and orchestration aspects with integrated satellite components in a 5G network,” within SA5 (management); and “Study on solutions for NR to support NTN,” within RAN3. These studies addressed the impacted areas for the integration of the NTN component and the related solutions for two specific use cases, extending the activities initiated in Rel. 15 on this subject. Moreover, also the technical challenges related to the interaction between the NTN RAN and the Next Generation Core network (NGC) were discussed. The requirements for NR systems introduced in 3GPP TS 22.261, [7], were extended to include the nonterrestrial component both interms of RAN technology and connectivity models. Also aspects related to network management and orchestration were analysed. Within the RAN3 SI previously mentioned, a set of required adaptations enabling NR operations on NTN channels; in particular, LEO and Geostationary Earth Orbit (GEO) scenarios were assessed both at system and link level and the outcomes are reported in TR 38.821, [2], building on the assumptions and definitions provided in TR 38.811, [1]. • In Rel. 17, two WIs were started: (1) “Solutions for NR to support NTN”, under RAN2 (Layers 2 and 3) activities, but covering RAN1, RAN2, and RAN3 technologies and techniques; and (2) “Integration of satellite systems in the 5G architecture”, under SA2. The former focused on the consolidation of the performance assessment in TR 38.821 and on the potential impacts of the NTN characteristics at PHY and MAC level. Moreover, also layers 2 and 3 were addressed, in particular for handover and DC, and the potential requirements for the upper layers were defined. The WI under SA2 focused on: (1) the identification of impact areas of the NTN integration in NR systems, in particular aiming at minimising it; (2) the analysis of the issues related to the

44

M. Marchese et al.

Fig. 2.2 3GPP timeline for the standardization of NTN in Rel. 15, 16, and 17

interaction between the NTN-based RAN and the NGC; and (3) the identification of solutions for roaming between terrestrial networks (TNs) and non-terrestrial networks and backhauling solutions. For the sake of completeness, it shall also be mentioned that under the CT TSG the following studies were performed: (1) a SI on “Study on PLMN selection for satellite access”; and (2) a WI on “PLMN selection for satellite access”; moreover, the following studies were initiated for NB-IoT via NTN both under RAN1: (1) SI on “Study on NBIoT/enhanced Machine Type Communication (eMTC) support for NTN”; and (2) WI on “Solutions for NB-IoT and eMTC to support NTN.” Figure 2.2 shows the timeline of the SIs and WIs listed above. The main objective in Rel. 17 was to identify the impact on the New Radio standard related to the characteristics of the satellite channel, including large propagation delays and large Doppler shifts, as well as the impact of having large and moving cells. The main focus was on transparent payload architectures with Low Earth Orbit (LEO) and Geostationary Earth Orbit (GEO) satellites, implicitely covering Medium Earth Orbit (MEO) and with the potential support for High Altitude Platform Stations (HAPS). The considered spectrum is Frequency Range (FR) 1, i.e., below 6 GHz, for handheld and car or drone mounted terminals. In Rel. 17, the User Equipment (UE) is assumed to have Global Navigation Satellite System (GNSS) capabilities.

2 The Role of Satellite in 5G and Beyond

45

2.1.2 NTN Radio Access Network Based on 3GPP nomenclature, [2], NTNs refer to networks, or segments of networks, exploiting Radio Frequency (RF) resources that are available on-board a satellite or an Unmanned Aerial System (UAS). The latter can include both a High Altitude Platform (HAPS) or an Unmanned Aerial Vehicle (i.e., drone). It shall be mentioned that UAVs are typically intended with a twofold role in NTN as they can be: (1) the User Equipment (UE), for services in which the NTN is controlling or monitoring the Unmanned Aerial Vechile (UAV); or (2) part of the NTN infrastructure, in which the UAV is providing connectivity to the onground UEs. In general, all types of orbit are possible, i.e., LEO, MEO, GEO, and HEO. However, up to Rel. 17, only GEO and LEO satellite systems have been considered, i.e., UAS or MEO/HEO orbits have not yet been assessed. When considering Non-Geosynchronous Orbits (NGSO), the coverage is classified as: (1) Earth-fixed beams, when the on-ground beams are fixed, i.e., the satellite digitally steers the beams towards the desired service area; or (2) Earth-moving beams, when the satellite is not steering the beams and, thus, they move on-ground along with the satellite’s movement on its orbit. Building on the analyses performed in the SIs within Rel. 15 and Rel. 16, 3GPP defined a set of architecture options for GSO and NGSO NTNs. In TR 38.821, these RAN structures, as well as the related radio protocols, interfaces, and the challenges to be coped with to adapt the New Radio (NR) procedures to the NTN characteristics, [1], are extensively discussed. The impact on the technologies initially developed for terrestrial 5G systems is deeply related to the type of NTN node (satellite or UAS), its capabilities, the constellation, and the use cases. It is also worthwhile highlighting that also how the space/aerial, user, and ground segments are interconnected and mapped to the 5G network elements plays a pivotal role. The architecture options for the NTN RAN can be broadly classified depending on: (1) the type of payload, which can be either transparent or regenerative; and (2) the type of user access link, which can be either direct or relay-based. With respect to regenerative payloads, it shall be mentioned that the space-/aerial-borne node can contain either the entire gNB protocol stack or only the Distributed Unit (gNBDU), in case functional split is implemented. As for relay-based access, the UE is connected to an Integrated Access and Backhaul (IAB) node, and not directly to the space-/aerial-borne node. Finally, in terms of UE, both handheld and satellitespecific equipment, as a Very Small Aperture Terminal (VSAT), are envisaged. Below, we provide an in-depth view on the different NTN-based RAN options.

2.1.2.1

Direct User Access

Direct access RAN architectures can be implemented with both transparent or regenerative payloads, as shown in Figs. 2.3 and 2.4, respectively.

46

M. Marchese et al.

Fig. 2.3 NTN-based RAN architecture with transparent payloads and direct user access

Fig. 2.4 NTN-based RAN architecture with regenerative payloads and direct user access. Functional split not implemented

Focusing on Fig. 2.3, the payload implements only frequency conversion, filtering, and amplification, and the gNB serving the on-ground UEs is conceptually located at the system gateway (GW). Thus, aiming at the full compatibility with the 3GPP NR standard, both the feeder and the user links shall be implemented by means of the NR-Uu Air Interface. In fact, a satellite equipped with a transparent payload can neither terminate the NR-Uu procedures nor manage the Quality of Service (QoS) flows. As a consequence, the RAN contains the gNB and the Remote Radio Unit (RRU), with the latter including the satellite and the GW to which the gNB is connected. It shall be mentioned that the gNB might be either located at the GW premises, i.e., it can directly connect to the satellite, or it can be located at a different position and, thus, it shall be connected to the gNB through the NG interface; then, the connection to the NGC and, from it, to other potentially required data networks is implemented through the N6 interface, [6]. It is worthwhile noticing that the architecture shown in Fig. 2.3 reports a single gNB; however, each gNB is capable of managing a few tens of beams, while in multi-beam NTN systems each satellite might be generating even hundreds of beams. Thus, depending on the number of on-ground beams, in order to manage the NTN node multiple gNBs might be needed. In case the satellites generate a reduced number of beams, then a single gNB can serve multiple payloads. Since the NR-Uu Air Interface implemented on the user and feeder links was specifically designed for

2 The Role of Satellite in 5G and Beyond

47

terrestrial systems, it is of paramount importance to assess the impact of the NTN channel impairments on the various radio protocol layers, including the Physical (PHY), Medium Access Control (MAC), Radio Link Control (RLC), Packet Data Convergence Protocol (PDCP) layers, for both the User Plane (UP) and Control Plane (CP). These aspects, addressed in 3GPP TR 38.821, [2], and TR 38.811, [1], are discussed in Sect. 2.1.2.4. Figure 2.4 shows the architecture option with a regenerative payload and without functional split, i.e., the entire gNB is implemented on-board. In this case, the NRUu protocols are terminated on the satellite and, thus, the GW acts as a Transport Network layer node, which terminates all of the transport protocols and is connected to the NGC through the NG interface, as already discussed. Having an on-board gNB, the user access link is still implemented as a NR-Uu Air Interface, while the feeder link shall carry the NG interface. This Air Interface is a logical interface, i.e., as long as it guarantees to perform a set of baseline protocol operations, it can be implemented by means of any Satellite Radio Interface (SRI), as, for instance, the DVB-S2, [8], DVB-S2X, [9], or DVB-RCS2, [10]. This architecture option allows to significantly reduce the over-the-air latency, since all NR-Uu protocols are deal with by the regenerative payload; however, is is also more complex and the cost of the satellite, and, thus, of the overall NTN system, is increased. Figure 2.5 still shows the architecture with the full gNB on-board; however, in this case an InterSatellite Link (ISL) to connect two (or more, if needed) satellites. The ISL acts as a transport link between the satellites and it shall be implemented by means of the Xn Air Interface. As for NG, also the Xn is a logical interface for NR systems and, as such, it can thus be implemented by means of any SRI. It is also worthwhile highlighting that, in Fig. 2.5 we are representing two on-board gNBs connected to two separate GWs and the same NGC. In fact, coordination through the Xn interface is needed within the same core network; however, it is possible to have the satellites connected to different NGCs and, in that case, no Xn would be needed. In terms of

Fig. 2.5 NTN-based RAN architecture with regenerative payloads, ISL, and direct user access. Functional split not implemented

48

M. Marchese et al.

Fig. 2.6 NTN-based RAN architecture with regenerative payloads and direct user access. Functional split implemented

the latency that impacts the NR procedures, when ISLs are considered, the over-theair latency on these links and the routing processing delays shall be considered. As a consequence, the impact of latency on the protocols will be a function of both the routing scheme and the number of hops. Finally, Fig. 2.6 depicts the architecture with regenerative payload and functional split. On the one hand, this allows a scalable solution based on Network Function Virtualisation (NFV) and Software Defined Networks (SDN) concepts so as to tailor the system to different use cases and vertical services, in addition to an overall improved performance in terms of network management. On the other hand, the overall system cost and complexity are increased. As extensively discussed in 3GPP TS 38.401, [11]: (1) a gNB can be split into a Centralised Unit (gNB-CU) and one or more Distributed Units (gNB-DUs); (2) each gNB-DU can be connected to a single gNB-CU, while a single gNB-CU can manage multiple gNB-DUs; and (3) the gNB-CU and the gNB-DU(s) are connected through the F1 and E1 Air Interfaces for the UP and CP, respectively. This interface is again logical and, thus, can be implemented by means of any SRI as long as specific signalling operations are guaranteed, [12]. This architecture poses a challenge related to the F1 interface. In fact, this interface requires a stable connection between the gNB-DU and the gNB-CU and it cannot be closed and re-activated on-demand; as such, with moving satellites as in a LEO scenario, all of the connections towards the served UEs would be dropped once the LEO satellite is outside of the visibility of the current gNBCU. Thus, smart implementations of the F1 interface and/or the functional split in NTN shall be designed. As per 3GPP NR, the functional split can be implemented at different layers or even within a single layer, separating the lower and higher sub-protocols therein, [6], as shown in Fig. 2.7. However, currently only option 2 is supported by 3GPP, which includes the F1/E1 interfaces; the other options are clearly possible, but they are not yet 3GPP-compliant up to Rel. 17 and they shall be implemented by means of Open Radio Access Network (O-RAN) or Common Public Radio Interface (CPRI) solutions. As for option 2, it can be noticed that: (1)

2 The Role of Satellite in 5G and Beyond

49

Fig. 2.7 Possible functional split options in 3GPP NR

the gNB-DU implements PHY, MAC, and RLC; and (2) the gNB-CU implements the PDCP and the Service Data Application Protocol (SDAP) for the UP and the CP, respectively.

2.1.2.2

Relay-Based User Access

In relay-based access solution, the UEs are not directly connected to the gNB or gNB-DU, but to an IAB. This network element is basically an evolution of the not particularly successful Long Term Evolution (LTE) Relay Node (RN) and it has been introduced in Rel. 16 as a flexible and scalable solution for multi-hop backhauling and to address dense deployment scenarios. Currently, indirect access solutions based on IAB is considered for further study within 3GPP; however, it is worth to explore also this type of solution for NTN, since it can indeed provide a viable option for NTN-based backhauling and transport networks. As detailed in TR 38.809, [13], and TR 38.874, [14], in the less complex implementation, a IAB, denoted as IAB-Donor, acts as a gNB and it is connected to the NGC through NG Air Interface. As shown in Fig. 2.8, it includes the CP/UP of the IAB-Donor gNBCU and then one or more DUs which can be connected to the IAB-Donor gNB-CU by means of a wired connection. Each IAB DU: (1) is connected to one or more IAB-nodes, in particular to their Mobile Termination (MT), through the F1 logical interface on the CP and the NR-Uu interface on the UP; and/or (2) is connected to the UEs to be served via the NR-Uu Air Interface. Thus, the introduction of the IABs allows to implement an extremely flexible and hierarchical backhaul infrastructure. It is also worthwhile mentioning that the IAB-Donor implements PDCP/SDAP and upper layers, while the IAB-nodes only implement the PHY, MAC, and RLC layers. Figures 2.9 and 2.10 show an example of NTN-based architectures with transparent and regenerative payloads, respectively. In both cases, and based on the discussions related to direct access solutions, it can be observed that the IABDonor and the NGC are always connected through the NG interface, which, with regenerative payloads, can be implemented by means of a SRI. Then, as previously mentioned, the connection between the IAB-Donor and the IAB-nodes is given by the F1-SRI and the NR-Uu interface, for the CP and UP, respectively. Based on this, the following cases arise: (1) with transparent payloads or regenerative paylods with on-ground IAB-Donor, both the feeder link and the user link shall

50

M. Marchese et al.

Fig. 2.8 Reference diagram of the IAB hierarchical structure

Fig. 2.9 NTN-based RAN architecture with IAB-based access and transparent payload

implement the F1/NR-Uu interfaces; (2) with regenerative paylods and on-board IAB-Donor, only the user link implements the F1/NR-Uu interfaces, while the feeder link shall be carrying the NG-SRI so as to connect to the satellite to the NGC. In terms of the challenges related to the radio protocols and procedures, the same considerations on the NR-Uu and F1 logical interfaces on the user and feeder links previously introduced hold and they will be discussed in Sect. 2.1.2.4. Finally, it shall be mentioned that the on-board IAB-node in Fig. 2.10, based on the above characteristics of IABs for NR systems, might also be serving one or more onground UEs. However, this situation is not considered here since it can be considered equivalent to a direct access architecture, discussed in Sect. 2.1.2.1.

2 The Role of Satellite in 5G and Beyond

51

Fig. 2.10 NTN-based RAN architecture with IAB-based access and regenerative payload. The IAB-Donor can be on-ground (top) or on-board (bottom)

2.1.2.3

Multi-Connectivity

Multi-Connectivity (MC) allows the UE to be connected to multiple transmitters at the same time. These transmitters simultaneously configure the radio resources to serve the target UE, thus introducing path diversity and significantly enhancing the capacity and service reliability. At the moment of writing this chapter, the focus in 3GPP for NTN is mainly directed towards Dual-Connectivity (DC) solutions in which two radio accesses can be used in parallel by the generic UE. As for the above architectures, both transparent and regenerative payloads can be implemented, and the UE can be served by: (1) a NTN-based RAN and a terrestrial RAN; or (2) two NTN-based RANs. It shall be noticed that, in the former case, the network complexity is larger due to the need for properly synchronise and align the transmissions over two very different channels. When regenerative payloads are assumed, as previously discussed, the gNB can be split into an on-board gNBDU and an on-ground gNB-DU. These observations lead to multiple architecture options, which are shown in Fig. 2.11, with connectivity from a terrestrial and a NTN infrastructure, and Fig. 2.12, with connectivity from two satellites. In both cases, a Xn interface over SRI is required to tightly coordinate the master and secondary gNBs; when functional split is involved, it shall be mentioned that, as represented in Fig. 2.11, the Xn interface coordinates the CUs, and not the DUs. With DC with two satellites, we considered regenerative payloads with and without functional split, which require an Xn interface over SRI or on-ground, respectively. As previously highlighted, the presence of the Xn interface is needed when the same NGC is considered. Finally, it shall be mentioned that the RAN might flexibly

52

M. Marchese et al.

Fig. 2.11 RAN architecture with Dual-Connectivity provided through terrestrial and NTN access with transparent (top) and regenerative (bottom) payloads

select either the NTN or the terrestrial gNB as master node, with the other acting as secondary gNB.

2.1.2.4

Radio Protocol Issues and Adaptations

The implementation on NTN channels of the NR Air Interfaces initially designed for terrestrial networks poses many challenges, due to the larger latencies and to the peculiarities of satellite networks. These have been extensively assessed in the study phase of NTN and the identified solutions, although still being enhanced and/or discussed, are included in TR 38.821, [2]. Below, we provide a brief overview of the main challenges at PHY/MAC level: • Timing Advance (TA): this procedure allows the UE to align the uplink and downlink transmissions at the gNB based on the distance from the gNB. The maximum TA value in 3GPP specifications is currently defined based on terrestrial systems and, as such, is not sufficient for NTN implementations. The simplest solution is that of (perhaps significantly) increasing the maximum supported TA, so as to allow the UE to align the uplink and downlink transmissions at the gNB. Other options, including more complex approaches based on Global

2 The Role of Satellite in 5G and Beyond

53

Fig. 2.12 RAN architecture with dual-connectivity provided through two NTN accesses without (top) and with (bottom) functional split

Navigation Satellite System (GNSS) capabilities at the UE and on the knowledge of the satellite ephemeris, are being discussed. • Aging of the Channel State Information (CSI): notably, CSI estimates from the UEs are reported to the gNB for several techniques and procedures, e.g., precoding and Adaptive Coding and Modulation (ACM). In NTN, the large latency can lead to CSI aging and, thus, an overall performance loss. One proposed solution in TR 38.821 was that of CSI reporting with channel averaging to capture long-term fading. A second option is to implement prediction-based link adaptation algorithms. For the moment being, the NR procedures based on CSI defined in Rel. 15 are considered sufficient also for NTN in Line Of Sight (LOS) conditions; in other propagation scenarios, the above solutions and other potential approaches shall still be evaluated. • Random Access (RA): Rel. 15 formats for the RA preamble can be used also in NTN assuming that time and frequency pre-compensation techniques are available at the UE. If this is not the case, then the preamble sequences require adaptations, including the design of novel preambles or the inclusion of scrambling sequences. Moreover, the RA procedure includes two timers at the UE during which the terminal is continuously monitoring the downlink channel for the corresponding responses from the gNB: (1) the RA Response window,

54

M. Marchese et al.

which is in the order of a few milliseconds; and (2) the Contention Resolution Timer (CRT). Despite the CRT is large enough to cover NTN scenarios, also in this case it is desirable to find a solution to avoid draining the UE battery with the continuous monitoring of the gNB reply. In both cases, an offset can be applied to the start of the timers based on the NTN scenario and geometry; in addition, it is proposed, if needed, to also extend the timers. • Hybrid Automatic Repeat reQuest (): this procedure improves the link adaption capability of the Air Interface based on the exchange of ACK/NACK messages between the UE and the gNB. The large propagation delay can significantly increase the latency in exchanging the ACK/NACK and also the memory requirements for the procedure. Disabling the HARQ for NTN was discussed, but it is not suggested for GEO since it may pose issues for other procedures and protocol layers, e.g., RRC signalling. For LEO, this is still under discussion, taking into account the lower Frame Error Rate (FER) that can be expected. Thus, several options to optimise HARQ in NTN are being discussed and they are reported in TR 38.821. Additional details on the PHY/MAC procedures in NTN can be found in [2, 15, 16].

2.1.3 5G-Advanced 3GPP Rel. 17 provides the specifications for the first, true integration of a nonterrestrial component with terrestrial communications. However, the definition of 5G-A features has started in March 2022 with Rel. 18. The related SIs and WIs are listed in Table 2.1. These enhancements will further strengthen the integration between NTN and terrestrial communications and will serve as a baseline for 6G systems, which will see a real unification with a joint terrestrial-NTN optimised network. This will require a further step compared to 5G and 5G-A, which were mainly aimed at optimising the terrestrial network and minimising the impact on the NR specifications to support NTN. Within Rel. 18, which is the first release addressing 5G-A systems, the following main features have been prioritised for NTN:

Table 2.1 List of WIs and SIs for NTN in Rel. 18 Type WI WI SI SI SI

Title NR NTN enhancements IoT NTN enhancements 5GC enhancement for satellite access Phase 2 Study on satellite backhauling Enhanced location services

TSG RAN2 RAN2 SA2 SA2 SA2

Completion Dec 2023 Dec 2023 Jun 2023 Jun 2023 Jun 2023

2 The Role of Satellite in 5G and Beyond

55

• Network-based UE localisation. • Coverage enhancements. • NTN deployments above 10 GHz and support for VSAT or Earth Station In Motion (ESIM) terminals. • NTN-TN and NTN-NTN mobility and service continuity enhancements. For the sake of completeness, it is also worthwhile mentioning that the following features will be considered for NB-IoT and eMTC: (1) mobility enhancements; and (2) discontinuous coverage enhancements. For the moment being, no prioritisation is clearly defined for the techniques to be assessed in future releases, i.e., Rel. 19. However, an educated guess is that the features de-prioritised from Rel. 18 will be assessed, including: (1) NTN-TN and NTN-NTN asynchronous MC and Carrier Aggregation (CA); (2) minimisation of the downlink Peak-to-Average Power Ratio (PAPR); (3) coordinated transmissions; (4) beam management enhancements; (5) half duplex Frequency Division Duplexing (FDD). In addition to these, new capabilities might also be considered, as: (1) the exploitation of regenerative payloads; (2) relay-based architectures for VSAT and ESIM; (3) the operation of UEs without GNSS capabilities; (4) spectrum coexistence between terrestrial and NTN systems; and (5) architectures supporting Artificial Intelligence (AI) and Machine Learning (ML) techniques for NTN.

2.2 Services The envisioned 5G Satellite-Terrestrial Integrated Network (STIN) improves the capabilities of the only terrestrial 5G infrastructure offering new functionalities that enhancing the offered performance [17]: • Service ubiquity: terrestrial mobile communication networks are typically deployed focusing on covering the densely populated urban area. The number of users per cell can so be high enough to guarantee enough money income to cover the network production, deployment, and management costs and obtain valuable profits. In rural and remote areas, instead, the number of users per cell is typically lower so the network coverage is more spotty and unreliable. Satellites can help bring network access covering all the areas within their footprints and so offering network connectivity to multiple uncovered areas without additional production and deployment costs. Satellites can offer connectivity to all underneath users without requiring additional terrestrial infrastructure deployment, especially in case of direct access. A properly sized satellite communication network, especially a LEO satellite constellation network, can offer a worldwide 5G coverage to users even in densely populated areas as well as in rural and remote areas. • Service continuity: the incomplete worldwide coverage of the current only terrestrial mobile communication infrastructure can also lead to service discontinuities affecting moving users. High-speed moving users, such as those located in cars,

56

M. Marchese et al.

trains, ships, and planes, but also low-speed pedestrian users, may pass through covered and uncovered areas multiple times during their 5G connectivity period, with consequent disconnection subperiods and multiple unpleasant disconnection and re-connection phases. Satellites can contribute improving this situation by “bridging” the terrestrial network covered areas. This action can and should take place in a completely transparent way for the final users, who should not experience any tangible modifications to the used services. • Broadcasting: some services, such as TV and radio broadcasting, require sending the same contents to multiple users on a very high space scale, typically regional or national. This may require a high number of terrestrial transmitting stations that have to be widely spread and linked together. Satellites can help considering that generate wider cells than the ones generated by terrestrial base stations, whose size mainly depends on the satellite altitude, where a higher number of users can so be located. This aspect can also be exploited to deliver the same contents to all the users with a single downlink broadcast transmission. The advantages of this action are even more prominent in case of a satellite constellation network, where more users even located in areas covered by different terrestrial networks, such as in different countries, can be easily reached by the same message with a single transmission. • Emergency backup: the terrestrial networks can suffer from natural events, such as earthquakes, floods, and tornados, and human actions, such as bombing, that can damage the terrestrial infrastructure and temporary or permanently disrupt its offered connectivity. Part of the terrestrial infrastructure design is to make it more robust against these events, for example carefully selecting the locations where to place the terrestrial base stations, but it will always be prone to these possible events. Satellites can contribute by offering a space infrastructure more robust against these events that can support the terrestrial infrastructure offering emergency connectivity in case of disruptions due to unplanned events. The terrestrial infrastructure can operate as the primary network while the space infrastructure can enter into action only when needed, with no or minimal limitations and inefficiencies perceived by the users. • Data offloading: the drastic increase in the number of users foreseen for 5G and B5G networks requires drastic improvements to the terrestrial infrastructure in terms of higher supported user density and data traffic volumes. Even if these requirements have already driven the planned and standardized improvements of the terrestrial infrastructure, some temporary congestion events where the terrestrial infrastructure is not able to fulfil all users’ requests can still happen. Satellites can contribute decreasing the probability of these events by offering a parallel space infrastructure where delay tolerant data flows can be offloaded, even before congestion situations take place. This can be particularly helpful for high-volume delay-tolerant applications, such as video distribution and remote data storage, where a high amount of data may need to traverse the network without requiring users’ real-time interaction.

2 The Role of Satellite in 5G and Beyond

57

• Preventively move data at the edge: some applications related, for example, to IoT use cases and real-time monitoring and surveillance may require retrieving a high volume of data over time to properly perform their task. In traditional networks, every time a user needs information that is typically stored in centralized nodes, such as servers in datacenters, sends them a proper request and waits until the required information is forwarded backwards. However, most of the users have the same habits and typically require the same service at the same daily time slots (for example, users typically use video streaming services more often during the evenings than mornings or afternoons). This can be one of the causes of the temporary congestion events mentioned in the previous point. Satellites can contribute supporting content forecast strategies, accordingly move the contents that will be required by the users in the near future, and temporary store them in storage and processing nodes, called edge servers, that can be located at the network edge, i.e., on the terrestrial part of the network closer to the final users. In this way, these data do not need to traverse the all network when they will be required by the users but only the smaller and local portions between users and edge servers. Satellites can so speed up the forwarding process to the edge. • Orbital Edge Computing: having network nodes to the edge is not only useful for storage purposes. These nodes could be also equipped with processing capabilities particularly useful in case data generated by users need to be elaborated and sent back, implementing the so called edge computing concept. This distributed approach is an evolution of the centralized approach in use in traditional terrestrial networks where the main processing capability is located in a centralized/cloud platform at the network core and the data have to traverse the all network back and forth. Satellites can contribute by offering onboard additional storage and processing capabilities dedicated to users’ needs, so avoiding that pre- and post-processed data traverse the core part of the network and further reducing latencies.

2.3 System Architectures The typical 5G STIN system architecture is composed of three main segments, as shown in Fig. 2.13.

2.3.1 User Segment The User segment is composed of the 5G terminals, i.e., the 5G UEs. They are the sources or destinations of all data flows passing through the 5G network. They can generate different kinds of data with different statistics depending on their nature (smartphones, IoT devices, .. . .), applications, and required services. They can be

58

M. Marchese et al.

Fig. 2.13 5G STIN system architecture, multi-connectivity scenario

nodes with fixed positions or moving with different mobility models, such as people with portable UEs travelling inside cars or onboard ships or planes. As we explained in Sect. 2.1, they can get connectivity directly from satellites or IAB-nodes in case of satellite direct access or relay-based access, respectively, and also from terrestrial base stations in case of multi-connectivity.

2.3.2 Space Segment The Space segment is composed of all the satellites of the 5G STIN organized in a satellite constellation. They are part of the NG-RAN and can differ from different parameters related to the single used satellite, such as the operational altitude, supported frequency bands, and generated cell(s), and to the satellite constellation, such number of satellites, satellite orbital plane parameters, and presence of InterSatellite Links (ISL). Satellite constellation design for 5G STIN is driven by the 5G Key Performance Indicators (KPIs). High supported user density per cell, high guaranteed data rate per user, and latencies as low as possible are leading the Space segment design process to consider the following aspects: • Additional communication resources: the increasing demand for data rate per user, bandwidth per user, and user density is posing new challenges to spectrum utilization. The utilization of new frequency bands, such as Q, V, and W, is one of the solutions under investigation to achieve this goal. Higher frequency bands allow having available and so allocating more bandwidth per user but lead to higher attenuations proper of satellite links that harm its benefits. • Multi-plane constellations: satellites are typically categorized in three classes depending on the operational altitude: GSO or GEO: 35,786 km; MEO: 6000–

2 The Role of Satellite in 5G and Beyond

59

35,786 km; LEO: 200–2000 km. A high number of satellites is not the only parameter to focus on to improve the 5G STIN coverage up to reaching the worldwide level. Multi-plane satellite constellations may be a valuable solution to cope with this aspect. Multiple orbital planes can guarantee a better redundancy of the cell coverage with a consequent higher acceptable user density. They also allow a possible smart distribution of the entire constellation resources and tasks to perform aiming to optimize the achievable performance. For example, LEO satellites can take care of the user access since they are the closest to the ground while MEO and GEO satellites can take care of the data traffic distribution through the network and the network control operations, such as routing and resource allocation, considering their wider visibility of the underneath network. • Inter-Satellite Links: satellites equipped with ISLs can allow end-to-end communications even if a satellite is not currently connected to both a served UE and a satellite gateway. In this case, a satellite can take care of the UE connectivity and access and forward the received data to other neighbouring satellites that can take care of the data forward to the Ground segment. This sort of role split will not be static but it will depend on the network status over time and, in particular, on the ground nodes (UEs and satellite gateways) located within each satellite’s covered area over time. Typically, each LEO satellite can exchange data through 4 LEO ISLs, two with the adjacent LEO satellites in the same orbital plane (intraplane ISLs) and the other two with the adjacent LEO satellites in the adjacent left and right orbital planes (inter-plane ISLs). These LEO ISLs are typically always active except when satellites are travelling at high latitudes due to the very close distance among them and the very high relative speed. In case of multi-plane constellations, satellites should be equipped also with inter-plane ISLs among different altitude satellites, such as LEO-MEO, LEO-GEO, and MEO-GEO ISLs, to allow data and control information exchange throughout the constellation.

2.3.3 Ground Segment The Ground segment consists of all the network nodes that are physically located on the ground. Part of them, i.e., the terrestrial gNBs and the satellite gateways, are part of the NG-RAN, while the others compose the 5G NGC. Also for the Ground segment, the design process has to be properly tuned by considering the 5G KPIs and the services that the 5G Core network has to manage. A key aspect is to allow a highly dynamic allocation of the communication resources, such as computational power, data storage space, and link bandwidth in order to guarantee to each UE the proper required amount of resources depending on the required application for the proper amount of time. The process of resource allocation and deallocation has to be automatic, dynamic, flexible, precise, and involve all the communication entities in the end-to-end path between each UE and the access to the data network, such as the Internet. The three main paradigms involved in this process are service orchestration, slicing, and virtualization implemented through the SDN, NFV, and

60

M. Marchese et al.

Software-Defined Radio (SDR) technological frameworks. Their combination can allow achieving this goal and make satellite systems integrable with terrestrial ones for any connectivity service. Separate the Management plane, i.e., the rules and strategies that manage the resource application to provide to each user the required QoS, from the Control plane, i.e., all the operations to enforce the resource allocation and the control packets related to the network management, and the Data/User plane, i.e., the data packet flows sent and received by the UEs, is the main aim of SDN. NFV allows abstracting the available resources from the physical nodes and offers them “as a Service”. Virtual Network Functions (VNFs) to assemble and chain are the means to achieve this goal. The slicing approach allows creating and managing multiple VNFs devoted to the most efficient implementation of specific services based on the available virtualized resources. The presence of a service Manager and Orchestrator allows operators and service providers to have a complete view of the resources and, consequently, capabilities of the 5G STIN network and coordinate the effort of the various network entities, elements, and layers to setup, configure, and manage multiple types of service. The Ground segment will be the focal point to allow such integration. The orchestration and virtualization approach based on softwarization techniques and cloud/edge processing capabilities will drive the evolution of both space and ground segment elements to achieve the high flexibility level of resource allocation required to match the main 5G KPIs in STINs. However, several aspects have to be taken into account, such as the integration with traditional satellite networks which are usually based on proprietary solutions concerning both physical (hardware) and operational (protocols) viewpoints. These hinder the sharing of their resources and capabilities with the 5G terrestrial networks and limit the adoption of new protocols and algorithms. In such conditions, 3GPP is considering the integration with not-5G native environments as a first step to allow deploying and expanding 5G networks without waiting to deploy native 5G elements, while new 5G-compliant satellite systems will be designed and made operative.

2.4 Applications Another advantage of the integration between terrestrial and satellite networks within the 5G framework is the enabling of applications and scenarios that can benefit from this integration from a better spread and improved performance viewpoints. Some of them are applications already consolidated but whose spread in particular environments, such as rural and remote areas, have been hindered so far by the lack of a proper network coverage, while others are emerging applications where satellites have a primary role in their operations and users exploitation: • Internet of Things (IoT) [18]: IoT is one of the key technologies and emerging applications in our future. It typically involves a very high number of smart and

2 The Role of Satellite in 5G and Beyond

61

connected devices able to massively collect data for different purposes, such as supporting informed decisions, reducing operational costs through automation, tracking objects and materials, monitoring assets and environmental parameters, and enabling more effective and innovative healthcare solutions. The main used wireless IoT communication technologies can be based on two solutions: mobile cellular network and Low-Power Wide-Area Network (LPWAN). Satellite network, and in general NTN, support is seen as a viable alternative to overcome the missing terrestrial infrastructure by using flying communication nodes such as satellites and UAVs [19]. Besides, IoT applications typically involve also data analysis and processing processes that cannot take place onboard the resource-constrained IoT devices but require nodes with higher computational capabilities. Satellites can contribute by offering not only data collection and forward capabilities but also allowing raw data to be processed onboard and then send the processing output back with only a two-hop (device or gateway-satellite and satellite-device or gateway) satellite link instead of involving a ground processing node with a four-hop (device-satellite, satellitesatellite gateway, satellite gateway-satellite, and satellite-device) satellite link [20]. • Moving platforms: one of the 5G concerns is to allow connectivity to UEs which are moving at high speed (up to 500 km/h) and are located everywhere. This is not trivial from a terrestrial infrastructure viewpoint. High-moving terminals means frequent terrestrial cell handovers and users that can be located in areas not covered by the terrestrial infrastructure, such as oceans. Satellites are the only solution that can guarantee high-speed links to On-the-Move terminals, such as onboard planes, trains, and ships. Satellites’ coverage cone area guarantees connectivity also to users not located on the Earth’s surface and moving at a higher speed than the traditional users whose position is fixed, slightly moving (pedestrians), or moving up to 150 km/h (onboard road vehicles). • Smart city [21]: most of the human population is nowadays living in cities, with an estimated increasing trend. The high technology level offered in urban areas is one of the parameters that is driving this huge amount of people to migrate from rural and remote zones. Satellites can contribute, among the improvements that ICT technologies can bring, to turn cities into “Smart cities”. Even if the terrestrial infrastructure will always be present and have the primary role in connecting the typically huge amount of users and devices, the satellite contribution can be manyfold. Outdoor localization and navigation still benefit from satellites. New generation satellite localization constellations, such as Galileo, can offer a much higher precision than the previous systems, such as GPS. Besides, they offer integrated solutions based on multiple satellite localization systems, such as GPS, Galileo, Beidou, and Glonass working together to estimate and retrieve very precise (accuracy below 1 m) positions. A further integration with the 5G terrestrial infrastructure in order to have positioning solutions based on both terrestrial and satellite systems are under standardization and foresees even higher accuracy and reliability for both outdoor and indoor environments.

62

M. Marchese et al.

Satellite data offloading capability can be fully exploited in smart city environments to further improve the supported user density, reduce the probability of possible congestion situations letting delay tolerant data to be forwarded through the Space segment, and keep the 5G UEs connected to the Internet even in case of damage of the terrestrial infrastructure, where satellites can also help rescuers, police officers, firefighters, etc., coordinate with civilians and spread emergency messages. • Smart farm/agriculture [22]: with the rapid growth of the world population, food production and animal breeding worldwide had to be increased rapidly. The smart farming/agriculture concept, based on using modern ICT technologies, is the solution to increase the quantity and quality of agricultural products with minimal loss and labour. A huge amount of devices, such as sensors, can be deployed in vast agricultural areas to better monitor crop growth and further increase food production. Satellites can help collect and forward all the data coming from or destined to these devices without the need of a terrestrial communication infrastructure, that in most rural areas is poor due to the traditionally low number of devices to connect. Different architectural solutions have been designed considering the agricultural devices as 5G UEs, so able to get direct connectivity to the satellites, or as not-5G nodes but directly linked with terrestrial 5G gateways that, on one hand, collect data from the agricultural devices and, on the other hand, forward these data through 5G interfaces to the satellites, operating as 5G UEs from the 5G network viewpoint. In the latter solution, further versions have been designed considering also the integration with aerial components, i.e., UAVs and HAPSs, that can operate as flying 5G gateways [23]. • Smart grid [24]: The electric grid is currently facing massive changes in light of a combination of interrelated technical and economic drives. The most important of these changes deals with making the grid “smarter” by enabling the communication of relevant information across the entire network. This will help toward the creation of new services and applications with the goal of a more efficient, reliable, secure, and cost-effective system in all the phases from generation to consumption. A huge amount of new data is being produced by new intelligent devices, such as smart meters, and have to be collected and forwarded where they can be processed, analysed, and stored. It is extremely important to build an information and communication overlay that is able to support the dictated requirements. Besides, the electrical grid is typically spread throughout a country connecting all houses and buildings, with hundreds of thousands kilometer long power lines whose conditions could be better monitored to check their current status and possibly estimate maintenance operations. Satellites can help provide this service to the smart grid, especially in areas with poor terrestrial network connectivity. Additional data sources can be deployed along all the electrical grid without, or with minimal, modifications of the 5G terrestrial infrastructure [25]. • Smart logistics and transportation [26]: logistics plays a vital role in economic growth and it is a driver of countries’ and firms’ competitiveness. However, on

2 The Role of Satellite in 5G and Beyond

63

account of the complex supply chains and high labour costs, the costs of logistics are still very high. There is no doubt that developing smarter approaches to improve logistics efficiency and reduce logistics costs is nowadays a timely and important topic. The concept of smart is based on modern ICT solutions and it foresees realizing a modern integrated logistics system in an intelligent way by real-time processing and comprehensively analyzing the information of all aspects of logistics. Smart logistics can bring end-to-end visibility, improve the way of logistics transportation, warehousing, distribution processing, distribution, and information services, also contributing to time and cost savings. Satellites can have a role in different stages of the logistic chains and help improve the efficiency of different tasks. They can help monitor the goods in transit thanks to their wider visibility of the entire Earth’s surface, manage the moving vehicles and the paths they need to follow, and offer a more precise and real-time monitoring of drivers thanks to onboard sensors that can collect and send data to processing centers through satellites [27].

2.5 Research and Development Activities Below, we provide an educated guess at the technologies that might constitute the building blocks to achieve the unification of terrestrial and NTN networks in 6G systems, [28]: • Regenerative payloads and active antennas: in the near future, satellite payloads will be equipped with advanced On-Board Processors (OBPs), which will allow to improve the link budget, reduce the latency, implement edge computing, and move on-board network features based on the selected functional split option. The combination of such advanced processors with active antenna arrays will allow the implementation of Cell-Free Multiple Input Multiple Output (CF-MIMO) or other advanced resource management techniques to allocate the power/frequency/time resources based on the users’ need. Notably, these techniques are non-trivial for NTN systems, since they typically require accurate CSI at the transmitter, [29]. However, recently also location-based techniques, which infer the CSI vectors based on the users’ locations and the satellite ephemeris, are gaining an increasing interest. These approaches will allow to realise Cell-Free paradigm, in which the beams are not a priori defined anymore, and the resources are actually tailored to the users’ locations and requests. In this framework, it is worthwhile also highlighting that distributed MIMO solutions are possible, in which the antenna arrays implemented on-board a subset of flying nodes, i.e., a swarm or formation, are jointly optimised. These approaches pose significant challenges in terms of the inter- and intra-swarm synchronisation and signalling, which shall be coped with in the near future. • AI and ML: focusing in particular on NGSO systems, which are characterised by a large speed of the flying nodes and by a significant heterogeneity (LEO/MEO

64

M. Marchese et al.

satellites, HAPS, drones), the real-time system optimisation is challenging. As such, AI and ML solutions are globaly recognised as a viable application for resource management, including CF-MIMO and beam-hopping, [30]. In this case, AI/ML algorithms can be exploited for both predictive channel estimation and user-centric resource scheduling. Both supervised (i.e., regression) and unsupervised (i.e., clustering) can be exploited. AI solutions can be also implemented aimed at detecting the ionospheric scintillations, forecasting of network traffic, remote sensing, and telemetry mining, among the others. • Next generation waveforms: notably, the high PAPR is one of the major challenges in implementing the NR CP-OFDM based waveforms in the NTN context. In particular, it was shown in [31] that: (1) the efficiency of the onboard amplifiers decreases for increasing carrier frequencies; and (2) transponder configurations with a limited number of OFDM channels per HPA e.g., 3–4 or less) experience a severe throughput degradation. In addition to PAPR, also the Out-Of-Band Emissions (OOBE) play a major role. While the NR uplink already foresees, optionally, a DFT-s-OFDM waveform to tackle this issue, it can be expected that future 6G systems will at least explore the implementation of other OFDM variants, including DFT-W-s-OFDM, W-OFDM, DT-s-f-OFDM, and fOFDM. In addition, also the implementation of Orthogonal Time Frequency Space (OTFS) is gaining attention. • Reflecting Intelligent Surfaces (RIS): a RIS is a thin meta-surfaces integrated with passive components allowing to manipulate the wireless signals impinging on its surface, [32]. As such, they allow to adjust the amplitude and phase of the signals and to re-direct them towards desired directions. Their application to NTN systems is still in its infancy; however, in [33–35], some interesting considerations related to link budget, refracting capabilities, and benefit for GEO communications are discussed. It can be expected that the research in this field will grow in the next few years to actually assess whether RIS can represent a viable solution for indoor coverage via NTN.

References 1. 3GPP TR 38.811, Study on New Radio (NR) to support non-terrestrial networks (Release 15) (2020) 2. 3GPP TR 38.821, Solutions for NR to support non-terrestrial networks (NTN) (Release 16) (2021) 3. 3GPP TR 22.822, Study on using Satellite Access in 5G; Stage 1 (Release 16) (2018) 4. 3GPP TR 23.737, Study on architecture aspects for using satellite access in 5G (Release 17) (2021) 5. 3GPP TR 28.808, Study on management and orchestration aspects of integrated satellite components in a 5G network (Release 17) (20210 6. 3GPP TR 38.801, Study on new radio access technology: Radio access architecture and interfaces (Release 14) (2017) 7. 3GPP TS 22.261, Service requirements for the 5G system; Stage 1 (Release 18) (2022)

2 The Role of Satellite in 5G and Beyond

65

8. ETSI EN 302 307-1, Digital Video Broadcasting (DVB); Second generation framing structure, channel coding and modulation systems for Broadcasting, Interactive Services, News Gathering and other broadband satellite applications; Part 1: DVB-S2 (2014) 9. ETSI EN 302 307-2, Digital video broadcasting (DVB); Second generation framing structure, channel coding and modulation systems for broadcasting, interactive services, news gathering and other broadband satellite applications; Part 2: DVB-S2 extensions (2021) 10. ETSI EN 101 545-1, Digital Video Broadcasting (DVB); Second Generation DVB Interactive Satellite System (DVB-RCS2); Part 1: Overview and System Level specification (2020) 11. 3GPP TS 38.401, NG-RAN; Architectur edescription (Release 17) (2022) 12. 3GPP TS 38.470, NG-RAN; F1 general aspects and principles (Release 17) (2022) 13. 3GPP TR 38.809, NR; Background for Integrated access and backhaul radio transmission and reception (Release 16) (2022) 14. 3GPP TR 38.874, NR; Study on Integrated Access and Backhaul; (Release 16) (2018) 15. A. Guidotti, S. Cioni, G. Colavolpe, M. Conti, T. Foggi, A. Mengali, G. Montorsi, A. Piemontese, A. Vanelli-Coralli, Architectures, standardisation, and procedures for 5g satellite communications: a survey. Comput. Netw. 183, 107588 (2020) 16. A. Guidotti, A. Vanelli-Coralli, M. Conti, S. Andrenacci, S. Chatzinotas, N. Maturo, B. Evans, A. Awoseyila, A. Ugolini, T. Foggi, L. Gaudio, N. Alagha, S. Cioni, Architectures and key technical challenges for 5g systems incorporating satellites. IEEE Trans. Veh. Technol. 68(3), 2624–2639 (2019) 17. L. Boero, R. Bruschi, F. Davoli, M. Marchese, F. Patrone, Satellite networking integration in the 5G ecosystem: research trends and open challenges. IEEE Network 32(5), 9–15 (2018) 18. M. Centenaro, C.E. Costa, F. Granelli, C. Sacchi, L. Vangelista, A survey on technologies, standards and open challenges in satellite IoT. IEEE Commun. Surv. Tutorials 23(3), 1693– 1720 (2021) 19. M. Marchese, A. Moheddine, F. Patrone, IoT and UAV integration in 5G hybrid terrestrialsatellite networks. Sensors 19 (2019) 20. P. Cassará, A. Gotta, M. Marchese, F. Patrone, Orbital edge offloading on mega-LEO satellite constellations for equal access to computing. IEEE Commun. Mag. 60(4), 32–36 (2022) 21. D. Minoli, B. Occhiogrosso, Practical aspects for the integration of 5G networks and IoT applications in smart cities environments. Wiley Wirel. Commun. Mobile Comput. 2019(5710834), 1–30 (2019) 22. Y. Tang, S. Dananjayan, C. Hou, Q. Guo, S. Luo, Y. He, A survey on the 5G network and its impact on agriculture: challenges and opportunities. Comput. Electron. Agric. 180 (2021) 23. M. Bacco, P. Barsocchi, E. Ferro, A. Gotta, M. Ruggeri, The digitisation of agriculture: a survey of research activities on smart farming. Wiley Array 3–4 (2019) 24. A. Meloni, L. Atzori, The role of satellite communications in the smart grid. IEEE Wirel. Commun. 24(2), 50–56 (2017) 25. K. Sohraby, D. Minoli, B. Occhiogrosso et al., A review of wireless and satellite-based M2M/IoT services in support of smart grids. Springer Mobile Netw. Appl. 23, 881–895 (2018) 26. Y. Song, F.R. Yu, L. Zhou, X. Yang, Z. He, Applications of the Internet of Things (IoT) in smart logistics: a comprehensive survey. IEEE Int. Things J. 8(6), 4250–4274 (2021) 27. E.J. Khatib, R. Barco, Optimization of 5G networks for smart logistics. Energies 14(6) (2021) 28. A. Guidotti et al., The path to 5G-advanced and 6G non-terrestrial network systems, in 2022 11th Advanced Satellite Multimedia Systems Conference and the 17th Signal Processing for Space Communications Workshop (ASMS/SPSC) (2022), pp. 1–8 29. A. Guidotti, C. Amatetti, F. Arnal, B. Chamaillard, A. Vanelli-Coralli, Location-assisted precoding in 5G LEO systems: architectures and performances, in 2022 Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit) (2022), pp. 154–159 30. F. Fourati, M.-S. Alouini, Artificial intelligence for satellite communication: a review. Intell. Converged Netw. 2(3), 213–243 (2021) 31. 3GPP TR 38.808, Study on supporting NR from 52.6 GHz to 71 GHz (Release 17) (2021)

66

M. Marchese et al.

32. M. Di Renzo, A.I. Aravanis, Catching the 6G Wave by using metamaterials, in Chapter in Shaping Future 6G Networks: Needs, Impacts, and Technologies (2022), pp. 69–87 33. S. Alfattani, W. Jaafar, Y. Hmamouche, H. Yanikomeroglu, A. Yongaçoglu, Linkbudget analysis for reconfigurable smart surfaces in aerial platforms. IEEE Open J. Commun. Soc. 2, 1980–1995 (2021) 34. Z. Lin, H. Niu, K. An, Y. Wang, G. Zheng, S. Chatzinotas, Y. Hu, Refracting ris aided hybrid satellite-terrestrial relay networks: joint beamforming design and optimization. IEEE Trans. Aerosp. Electron. Syst. (2022) 35. W.U. Khan, E. Lagunas, A. Mahmood, S. Chatzinotas, B. Ottersten, When RIS meets GEO satellite communications: a new sustainable optimization framework in 6G (2022) [Online]. Available: https://arxiv.org/abs/2202.00497

Chapter 3

Futuristic Satellite Scenarios in 6G Justine Cris Borromeo, Koteswararao Kondepu, Mauro De Sanctis, Luca Valcarenghi, Riccardo Bassoli, and Frank H. P. Fitzek

3.1 Vision of 6G and Non-Terrestrial Networks 5G broadband cellular networks started its worldwide deployment in 2019, which supports new services based on three major scenarios: (1) enhanced mobile broadband (eMBB); (2) massive machine-type communications (mMTC); and (3) ultra-reliable low-latency communications (URLLC) [1]. However, there are still some areas and scenarios that experience cellular connectivity issues. People living in rural areas of Low- and Middle-Income Countries (LMIC) are 37% less likely to use mobile internet compared to those living in urban areas, with the largest ruralurban gap reported in Sub-Saharan Africa [2]. Another concern is the lack of internet access during post-disaster recovery: communication towers usually get damaged depending on the intensity of the disaster, while the internet connectivity to access Social Network Sites (SNS) like Whatsapp, Skype, etc. for residents on disaster-hit

J. C. Borromeo · L. Valcarenghi Scuola Universitaria Superiore di Sant’Anna, Pisa, Italy e-mail: [email protected]; [email protected] K. Kondepu () Department of Computer Science and Engineering Indian Institute of Technology Dharwad, WALMI Campus, Dharwad, India e-mail: [email protected] M. D. Sanctis University of Rome “Tor Vergata”, Rome, Italy e-mail: [email protected] R. Bassoli · F. H. P. Fitzek Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany Centre for Tactile Internet with Human-in-the-Loop (CeTI), Dresden, Germany e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_3

67

68

J. C. Borromeo et al.

Public Defense

Emergency and Safety Rural Areas

Disaster Areas

Crowded Areas Maritime

Fig. 3.1 5G used cases in non-terrestrial network

areas to communicate their situation and needs is very important [3]. Finally, using internet connectivity to live-stream users’ experience in a concert, football match, or festivals is very common nowadays. However, a limited available spectrum cannot serve the crowd in these types of events [4]. The use of Non-terrestrial Network (NTN) or Three-Dimensional Network (3D Network) will play an important role in beyond 5G and 6G communication networks to address the network connectivity issues that are experienced on some use case scenarios as shown in Fig. 3.1 [5, 6]. 3D networks are expected to improve the network availability on areas with low Average Revenue per User (ARPU), and temporarily provide cellular connectivity on areas that are affected by natural disasters where communication towers are affected. In crowded areas like concerts, festivals or game match, 3D network can also operate when terrestrial networks are already overloaded. Moreover, 3D networks are also expected to play an important role in maritime, emergency and safety, and public defense applications. This chapter will discuss the feasibility of exploiting UAV as an aerial network to provide cellular communication in unserved/underserved, disaster-hit, and hotspot areas. Two different implementation scenarios are considered in this chapter. The first implementation is using UAV as a relay system between the satellite and end devices. In this scenario, two different relay protocols which are Amplify-andForward and Decode-and-Forward will be investigated. Another implementation is by utilizing UAV as a Radio Unit (RU). Due to the high bandwidth requirement of the fronthaul link, physical layer specifications are varied to achieve the data

3 Futuristic Satellite Scenarios in 6G

69

rate that can be supported by the satellite communication when considering an implementation that uses Ka-band frequency. Then, different physical layer functional split options (e.g., 7-1, 7-2, 7-2x, and 7-3 functional split options) are evaluated to determine which one is better suited for 3D networks. Localisation and RF sensing services will also be defined in single satellite network.

3.2 Architectural Perspectives of 3D Network in 6G Three-dimensional networks in 5G are characterized by hierarchical structure which is composed of satellite, aerial and ground stations [7]. In this section, different 3D Network platforms will be discussed together with the frequency bands that are used for satellite communication, and relay protocols to be used in the aerial networks.

3.2.1 3D Network Platforms and Frequency Bands Figure 3.2 shows the satellite and aerial platforms that can be integrated to the terrestrial network for cellular connectivity in 3D networks. • Geostationary Earth Orbit (GEO) satellites are the farthest from the terrestrial station with an altitude of around .35,800 km, which is why the experience a very huge attenuation and propagation delay [8]. However, they can cover a large geographical area and are continuously visible from terrestrial terminals. • Medium Earth Orbit (MEO) satellites have an altitude between 7000 km to .25,000 km, which is lower than GEO satellites resulting to a lower propagation Fig. 3.2 Satellite and aerial platforms for 3D network

G ~ 36000 km GEO

MEO ~ 7000 km

LEO ~ 300 km HAP ~ 20 km

LAP ~ 1 km

70

J. C. Borromeo et al.

delay and better signal strength. However, they must operate in constellation to achieve service continuity since they are non-stationary. • Low Earth Orbit (LEO) satellites has the lowest altitude among the three satellite platforms with an altitude from 300 km to 1500 km, which can achieve a better signal strength and lower propagation delay than MEO and GEO satellites but with a trade-off of covering a smaller geographical area. LEO satellites are also non-stationary and must operate in constellation to provide a continuous service. • High Altitude Platforms (HAPs) are aerial networks that are deployed in the stratosphere with an altitude of .≈ 20 km. They can be hot-air balloons or airships, and can support a more flexible and cost effective implementation than satellites. • Low Altitude Platforms (LAPs) on the other hand like Unmanned Aerial Vehicles (UAVs) can fly from 100 m to 1 km altitude, which provides a high signal strength and low propagation delay due to their short range line-of-sight deployment. They can also be deployed faster and cheaper but with the disadvantage of limited power supply. In order to support the communication in 3D networks, different frequency bands that support satellite communication are considered for 6G NTN implementation namely S-band (.2.17–.2.2 GHz), C-band (.3.4–.3.7 GHz), X-band (.7.25–.7.75 GHz), Ku-band (.10.7–.12.75 GHz), and Ka-band (.≈ 20 GHz) [9].

3.2.2 Proposed Implementation Two different types of proposed implementation are considered to evaluate the feasibility of UAV-assisted 3D networks to provide cellular connectivity specifically on rural, post disaster, and hotspot areas. In this implementation, UAVs will be exploited as an aerial network since they can be deployed faster making them more cost-effective with lesser energy consumption [10]. The first proposed implementation scenario is shown in Fig. 3.3 where the UAV is used as a relay protocol implementing either Amplify-and-Forward (AF) or Decodeand-Forward (DF). The role of UAV as a relay protocol between satellite and mobile terminal is relevant since frequency bands that are used by the current service providers operated on the Sub-6GHz frequency (e.g. 3.6 GHz is used in major cities while 2.1 GHz are used in other areas of Germany), which cannot be compatible to satellite operating in higher frequencies(X-, Ku- and Ka-band) unless a transponder is used in between satellite-UE to convert these frequencies to Sub-6GHz [9]. In this scenario, the 5G core network is deployed in the ground station while the baseband or gNB functions are deployed in the satellite. The 5G core is connected to the gNB through an NTN gateway using the NG Satellite Radio Interface (SRI). After performing the gNB functions, the satellite then sends the data to the UAV through any of the satellite communication frequency bands (e.g., S-band, C-band, X-band, Ku-band, and Ka-band). The UAV then performs either AF or DF and converts the frequency to S-band to be connected to the handheld and IoT terminals.

3 Futuristic Satellite Scenarios in 6G

71

gNB CU/DU/RU NTN Gateway S-band C-band X-band Ku-band Ka-band

5G Core

NG

Amplify-and-Forward Decode-and-Forward Sub-6 GHz

Fig. 3.3 3D network implementation using relay system PDCP RLC

CU/DU

MAC High-PHY

Option 7-2 Option 7-2x Option 7-3

Fronthaul

Option 7-1

5G Core

NTN Gateway

RU

Low-PHY RF

NG

Fig. 3.4 3D network implementation using lower layer functional split options

Another considered implementation is where gNB functions are distributed between UAV and Satellite as shown in Fig. 3.4. In this implementation, the 5G core is still implemented in the ground station and is connected to the satellite via an NTN gateway using the NG SRI. In this scenario, some of the baseband functions are implemented in the UAV using physical layer split options (e.g., Option 7-1, 7-2, 7-2x, and 7-3). Which means that the Packet Data Convergence Protocol (PDCP), Radio Link Control (RLC), Medium Access Control (MAC) and Higher Physical Layer (High-PHY) functions are implemented in the satellite while the Lower Physical Layer (Low-PHY) and Radio Frequency (RF) functions are deployed in the UAV. Lower layer split options are considered in this implementation to reduce the number of 5G functions implemented in the UAV since energy consumption is very critical when exploiting UAVs. Increasing the number of functions deployed

72

J. C. Borromeo et al.

in the UAV will just increase its processing power resulting to a shorter flying time since they are battery powered [11].

3.3 Localisation and RF Sensing in Satellite Networks Satellite communications, localisation and RF sensing services have different requirements and the integration of such services in a single satellite network dramatically increases the design complexity. The set of design drivers of a satellite communication network includes: global and continuous coverage (maximum availability time over any ground station), high mean elevation angle (multipath minimization), low number of handovers, low latency. For what concerns localization services, the set of design drivers includes: high number of simultaneously visible satellites (minimum of 4), high mean elevation angle (multipath minimization), high separation between visible satellites, frequent update of measurements, low Geometric Dilution of Precision (GDOP). For what concerns RF sensing services, the design phase should consider: high precision of estimates, frequent updates of estimates, rich set of estimates including e.g. user speed, group mobility, user density in areas of interest. Security and user privacy protection are requirements common to all the services.

3.3.1 Satellite Network-Based Localisation The common view of satellite-based localisation/navigation refers to GNSS systems such a GPS, Galileo, GLONASS, Beidou where the user terminal localise itself through a GNSS receiver which processes the RF signals received by the satellites of the GNSS constellation and computes its geographical location. This type of approach is called user-based localisation: the user terminal collects several measurements from RF signals transmitted by several sources in known locations. An alternative type of approach is called network-based localisation; in this case the nodes of the localisation/communication network receive the signal transmitted by the user terminal and a central network controller processes the signal received by the different network nodes (again positioned in known locations) to compute the user terminal position. Network-based localisation is of interest in terrestrial cellular networks where network operators would estimate the geographical distribution of users to optimise the efficiency of network operation [12]. However, network-based localisation may be of interest also in satellite communication networks, especially in LEO satellite constellation networks, for the management of basic techniques such as handover and paging. Currently, the most common option is to require each user terminal to carry a GNSS receiver jointly with the satellite radio communication interface so that the user terminal position is communicated to the satellite network by the user

3 Futuristic Satellite Scenarios in 6G

73

terminal itself. Unfortunately, it is not always feasible to include a GNSS receiver in a user terminal because of limitations in terms of terminal cost and energy. For instance, in specific scenarios such as the Internet of Remote Things (IoRT), the user terminal has severe limitations in terms of energy as it may only use battery power and energy harvesting from e.g. sunlight, thermal gradients, or mechanical motion [13]. Furthermore, most of the energy stored by the batteries of IoRT terminals are needed to transmit RF signals towards the satellites with relatively high power. In fact, IoRT user terminals are located in remote areas of the Earth where neither the electricity grid nor a wireless terrestrial network are available. Satellite network-based localization may benefit of many of the results in the area of emitter/jammer geolocation using a single satellite or a swarm/formation of satellites for signals intelligence (SIGINT) or electronic intelligence (ELINT) applications. Such systems exploit well known location methods based on measurements/estimates of Time of Arrival (TOA), Time Difference of Arrival (TDOA), Frequency of Arrival (FOA), Frequency Difference of Arrival (FDOA), Angle of Arrival (AOA). For what concerns the geolocation service from a cluster of satellites, the following works are mentioned. A brief presentation of the ELectronic Intelligence by SAtellite (ELISA) mission, of the operational organization, and of the satellites characteristics is provided in [14] together. The ELISA mission involves a fleet of four microsatellites which were launched in December 2011 with the objective to characterize and localize RF emitters on ground. The orbital requirements of the mission are discussed and in particular the need to keep a specific satellite geometrical pattern. HawkEye 360 (HE360) Pathfinder mission is a formation-flying cluster of three microsatellites in Sun Synchronous Orbit (SSO) between 550 and 650 km for highprecision RF geolocation using TDOA and FDOA measurements [15]. For what concerns the geolocation service from a single satellite, notable examples are the following. A new algebraic solution for the Doppler positioning problem which allows to estimate the position of a stationary emitter is estimated from Doppler frequency measurements collected by a single LEO satellite was presented in [16]. The work in [17] addresses high-accuracy geolocation of a ground-level, uncooperative RF emitter using a single LEO satellite through a realtime, single-pass solution that requires only a few seconds of Doppler and Doppler Rate measurements and resistant to erroneous ephemeris readings and oscillator errors.

3.3.2 Satellite Network-Based RF Sensing In addition to the classic application of Earth Observation (also Remote Sensing), which refers to the technique of observing and analyzing passive objects from space without being in direct contact with them, satellite communication systems may be also exploited for two additional applications in the RF device-based sensing

74

J. C. Borromeo et al.

domain: (1) RF spectrum sensing with jammer detection and localization; (2) user terminal RF sensing and mobility detection. The first application refers to the utilization of on-board broadband RF receivers analyzing the RF spectrum with the objective to search for jamming signals or unknown RF sources. This application usually foresees two steps, i.e. detection and localization. The second application refers to the analysis and processing of user terminals’ signals received by the satellite nodes to provide higher level information with respect to simple geographical location, i.e. users’ terminals speed and type (including the classification as pedestrian or vehicular on car, train, airplane), users’terminal geographical concentration and group motion. This application has the objective to monitor users’ activities and may use directly the localization service, e.g. compute user speed from multiple positions, or may apply additional processing to the received signals to extract useful information.

3.4 Design Parameters This section will discuss different design parameters that will be evaluated considering the two different implementation scenarios in Sect. 3.2.2.

3.4.1 UAV-Based Implementation of Relay Protocol in 3D Networks When using UAV as a relay protocol, four different parameters can be considered which are the transmission delay, session time, signal-to-noise ratio, and throughput. These parameters will be used in evaluating and comparing the performance of AF and DF together with the direct link from satellite to ground station.

3.4.1.1

Transmission Delay

One of the necessary parameters when using UAV as a relay system is the transmission time between the satellite and UAV. To measure the transmission delay, an acceptable delay budget .φ is considered to satisfy the required end-toend latency [18]. It is the processing and transfer time of data from UE to the gNB and vice versa, which can be computed using [19]: φ = ttx + tbp + tq + tother

.

(3.1)

3 Futuristic Satellite Scenarios in 6G

75

with .ttx as the transmission delay, .tbp as time to process the baseband unit, .tq as queuing latency introduced by each node involved in the communication, and .tother as the time to perform other functions like inverse/fast fourier transform (IFFT/FFT). Assuming .tbp as the time to perform Low-density parity-check code (LDPC) decoding, then the allowed transmission budget .ttx can be expressed as follows: ttx = φ − tq − tother −

.

LF k pO

(3.2)

where .LF k/pO is the time to process LDPC decoding with k as the number of decoding iterations, L the code block size in bit, F the decoder complexity in .operations/bit, p the processing unit’s (PU) clock rate in H z and O the processor efficiency in .operations/cycle [19]. The transmission delay can also be expressed in terms of distance d, which is the slant range, i.e., the wireless path connecting the UAV and CubeSat depending on the elevation angle . and the CubeSat altitude h [20]. In that case, .ttx = d/c, which can also be formulated as follows [21]:  ttx =

.

(REarth +h)2 (REarth +a)2

 − cos 2 () − sin() · (REarth + a) c

(3.3)

with .REarth as the radius of the Earth, a as the UAV’s altitude and c as the speed of light. In the next sub-section, the estimation of session time will be computed by combining Eqs. (3.2) and (3.3).

3.4.1.2

Session Time

Another parameter considered is the session time, which is defined as the amount of time it takes for the satellite to communicate with the UAV. In this implementation, we considered a LEO satellite that orbits around the Earth with a speed of v[22]. Considering a UAV hovering over the UE, the LEO satellite will be able to communicate to to the UE through the UAV which performs either AF or DF relay protocol with a session time .ts . The minimum elevation angle . min is retrieved by fixing the maximum number of LDPC decoding iterations .k max in Eq. (3.2), max . For this purpose, which leads to the maximum allowed transmission delay .ttx maximizing .ts is preferable to utilize a slow handover strategy between the satellite max is and UAV [23]. The UAV re-establishes a link to the CubeSat as soon as .ttx exceeded. The session time .ts can be computed as follows: ts =

.

θ max · (REarth + h) v

(3.4)

76

J. C. Borromeo et al.

considering .θ max as the maximum Earth’s central angle and can be computed using: θ

.

  dmax · cos( min ) = arcsin REarth + h

max

(3.5)

max . Considering an elevation angle ranging between . = with .dmax = c · ttx min [ , π/2], .ts is a lower bound where we consider the LEO satellite at the maximum distance .d max and another one approaching the Zenith, thus at .d min = h. The upper bound of .ts considers a CubeSat at .d max ( min ) and another one at max (π −  min ), thus roughly doubling the session time. .d

3.4.1.3

Signal-to-Noise Ratio

Signal-to-Noise ratio (SNR) is the measurement of strength of the desired signal with respect to the noise of the channel. The SNR when using an AF relay protocol can be computed using SNRAF =

.

 n  1+ i=1

 −1 1 −1 SNR (n)

(3.6)

while the SNR when using DF is SNRDF = min(SN R (n) )

.

(3.7)

where .SNR (n) is the SNR of each hop. In this case the SNR from satellite-UAV and UAV-ground. In order to compute for the SNR of each hop, we can use this formula [24, 25] SN R(dB) = EI RP +

.

G − P L − k − B − NF T

(3.8)

where EI RP is the effective isotropic radiated power, .G/T is the receiver antennagain-to-noise temperature, P L is the path loss, k is the Boltzmann’s constant which is around .1.3806485210−23 J /K, B is the channel bandwidth, and NF is the noise figure. The EI RP can be calculated using EI RP (dBW ) = PT − LC + GT

.

(3.9)

where .PT is the antenna transmission power, .LC is the cable loss, and .GT is the transmit antenna gain. While .G/T can be computed using G/T (dB) = GR − NF − 10log10 (T0 + (Ta − T0 )10−0.1N F )

.

(3.10)

3 Futuristic Satellite Scenarios in 6G

77

Satellite Network

TB 1 TB 2 Random Bits

DL-SCH

PDSCH

CPOFDM

Precoding

Doppler PreCompensation

NTN Fading Channel Model

Throughput Estimation

DL-SCH

Decoding

Ground Network

PDSCH

CPE

MMSE Equalization

Channel Estimation

Decoding Estimation and Correction

CPOFDM

Timing Synchronization

Demodulation

Fig. 3.5 PDSCH throughput simulation using MATLAB satellite communication toolbox

with .GR as the receive antenna gain, .T0 as the ambient temperature, and .Ta as the antenna temperature. Considering a parabolic antenna on the receiver, .GR can be achieved using  GR = 10log10

.

D2 η∗π ∗ 2 λ 2

 (3.11)

with .η as the antenna aperture efficiency, D as the equivalent antenna diameter in meters, and .λ as the wavelength.

3.4.1.4

Throughput

Throughput is the measurement of data that can be relayed from the satellite to the UE deployed in the ground station. The satellite communication toolbox of Matlab is exploited to measure the Physical Downlink Shared Channel(PDSCH) from the satellite to the UE [26]. Based on the specifications defined in [24, 25], the satellite communication toolbox can measure the PDSCH throughput of an NTN channel using the 5G NR with varying SNR values. Figure 3.5 shows the processing chain of the PDSCH throughput simulation. It features DL-SCH transport channel coding with up to 2 codewords and 8 layers, PDSCH precoding using singular value decomposition (SVD), cyclic prefix orthogonal division multiplexing (CP-OFDM) modulation, practical synchronization and channel estimation, and a single bandwidth part (BWP) across the whole carrier.

78

J. C. Borromeo et al.

Table 3.1 System parameters for LEO, UAV and ground architectures Parameter Frequency band Carrier frequency (GHz) EIRP (dBW) G/T (dB/K) Receiver gain (dBi) System bandwidth (MHz) Noise figure (dB)

LEO-UAV S 2 45 −0.95 24.87 30 1.2

Parameter Frequency band Carrier frequency (GHz) EIRP (dBW) G/T (dB/K) Receiver gain (dBi) System bandwidth (MHz) Noise figure (dB)

UAV-ground S 2 27.9 −6.75 24.87 same as LEO-UAV 7

Parameter Frequency band Carrier frequency (GHz) EIRP (dBW) G/T (dB/K) Receiver gain (dBi) System bandwidth (MHz) Noise figure (dB)

LEO-ground S 2 45 −6.75 24.87 30 7

3.4.1.5

C 3.6 45 4.16 29.98 100 1.2

X 7.5 45 12.51 36.35 100 1.2

Ku 11.3 45 16.06 39.91 200 1.2

Ka 20 45 21.03 44.87 400 1.2

C 3.6 45 −1.64 29.98 100 7

X 7.5 45 7.02 36.35 100 7

Ku 11.3 45 10.58 39.91 200 7

Ka 20 45 15.54 44.87 400 7

System Parameters

The system design parameter for the LEO satellite, UAV and Ground station are summarized on Table 3.1. These parameter are based on specifications of satellite, aerial and ground stations reported in [24, 25, 27–30]. Three different tables are shown considering system parameters for LEO-UAV, UAV-Ground, and LEOGround scenarios. The first two tables will be used for the AF and DF relay system implementation, while the third table will be for a direct link implementation from LEO to ground station. In this research, a LEO Satellite is considered with an altitude of 300 km and can support frequencies on the S-band at 2 GHz, C-band at 3.6 GHz, X-band at 7.5 GHz, Ku-band at 11.3 GHz, and Ka-band at 20 GHz with a channel bandwidths of 30 MHz, 100 MHz, 100 MHz, 200 MHz, and 400 MHz, respectively. The .G/T and the receiver gain is achieved considering an antenna diameter of 1 m.

3 Futuristic Satellite Scenarios in 6G Fig. 3.6 Fronthaul interface between the satellite and UAV with different lower layer functional split options

79

Low-PHY

High-PHY 7-3

7-2 7-2x 7-1

3.4.2 UAV-Based Implementation of Radio Unit in 3D Networks Another proposed scenario in implementing 3D Networks in 6G is by performing RU functions in the UAV using lower layer functional split options (e.g., 7–1, 7–2, 7–2x, and 7–3). The focus of this implementation scenario is the data transmission between the satellite and UAV considering the Ka-band frequency. In this section, 5 different performance parameters are considered to evaluation the satellite-UAV communication through 6G in a 3D Network scenario.

3.4.2.1

Fronthaul Bandwidth

The first parameter considered is the fronthaul bandwidth, which is defined as the data rate between the DU(satellite) and the RU(UAV). Figure 3.6 shows the fronthaul interface between the satellite and UAV considering different lower layer split options. According to 3GPP [31, 32], lower layer functional split option have different fronthaul bandwidth requirements. The fronthaul bandwidth requirements for option 7-1 can be computed using: F H = NSC ∗ NSY ∗ NAP ∗ NBT W ∗ 2 ∗ 1000 + MAC

.

(3.12)

while the fronthaul bandwidth for options 7-3, 7-2 and 7-2x can be determined using the formula below: F H = NSC ∗ NSY ∗ NLA ∗ NBT W ∗ 2 ∗ 1000 + MAC

.

(3.13)

with .NSC as the number of subcarriers of each OFDM symbol, .NSY as the number of OFDM symbols in a subframe, .NLA as the number of layers, .NAP as the number of antenna ports, .NBT W as IQ bitwidth, and MAC as the Medium Access Control(MAC) layer information. The number of subcarriers depends on the channel bandwidth used for transmission, while the subcarrier spacing determines the number of OFDM symbols in each subframe. The IQ bitwidth is considered to have 16-bits, and the MAC layer information is assumed to consume .121 Mbps for 7-1, and .713.9 Mbps for 7-2, 7-2x, and 7-3 [32].

80

J. C. Borromeo et al.

3.4.2.2

Theoretical Throughput

Another parameter considered is the throughput which is the amount of information that the UAV can process in a given time. The theoretical throughput in Mbps can be computed using the formula [33]:

.

T hroughput = 106

J

j

j

j

NLA ∗ Qm ∗ f j ∗ Rmax ∗

j =1

j

NSC ∗ NSY μ

Ts

∗ (1 − OH j ) (3.14)

where J is the number of aggregated component carriers, .NLA is the number of layers, .Qm is the maximum modulation order, f is the scaling factor (where values can be taken from 0, .0.8, .0.75 or .0.4), .Rmax is the code rate (.= 948/1024), .NSC is the number of subcarriers of each OFDM symbol, .NSY is the number of OFDM μ symbols in a subframe, .Ts is the time duration of each subframe in seconds, and OH is the overhead which takes the following values: .[0.14] for FR1 frequency range in DL; .[0.18] for FR2 frequency range in DL; .[0.08] for FR1 frequency range in UL and; .[0.10] for FR2 frequency range in UL.

3.4.2.3

Connection Density

Connection density is the average number of user/devices that can be connected to the UAV. In this case we can consider the required user experience throughput in Average Revenue Per User (ARPU) and rural areas. According to [34], areas with low ARPU require around 10 Mbps of user experience throughput while 50 Mbps is required in rural areas. The connection density can then be achieved by dividing the cell throughput in Sect. 3.4.2.2 to the required user experience throughput.

3.4.2.4

Number of Functions

The number of functions to be deployed in the RU is another important parameter to be considered since energy consumption is very important in UAV implementation. In this case, the least number of gNodeB functions, the better the energy efficiency in the UAV resulting to a longer flying time. Figure 3.7 shows the specific functions that will be deployed in the UAV considering different lower layer functional split options.

3.4.2.5

Fronthaul Energy Consumption

The energy consumption of the data reception on the fronthaul interface is measured in terms of energy per bit .(Eb ), which is the signal power over the user bit rate, as shown in Eq. (3.15) below:

3 Futuristic Satellite Scenarios in 6G

81

Modulation Layer Mapping Precoding RE Mapping

RE Mapping

Beamforming

Beamforming

Beamforming

IFFT and CP Addition

IFFT and CP Addition

IFFT and CP Addition

IFFT and CP Addition

RF

RF

RF

RF

Option 7-3

Option 7-2

Option 7-2x

Option 7-1

Fig. 3.7 Number of functions implemented in the RU for different lower layer functional split options

Eb = PR /RB

.

(3.15)

with .Eb as the energy per bit, .PR as the received power, and .RB as the bitrate between the satellite and UAV. The received power .PR can be achieved using: PR = EI RP + GR − Losses

.

(3.16)

with EI RP as the Equivalent Isotropic Radiated Power, .GR as the receiver gain, and Losses in the signal that can be due to free space path loss (FSPL), atmospheric loss (AL), and scintillation loss (SL). Typical values of EI RP , .GR , and losses are defined in [25] and [27] where .EI RP = 66dBW , .GR = 35.3dBi, .F SP L = 184.6dB, .AL = 0.5dB, and .SL = 0.3dB.

References 1. A. Ghosh, A. Maeder, M. Baker, D. Chandramouli, 5G evolution: a view on 5G cellular technology beyond 3GPP release 15. IEEE Access 7, 127639–127651 (2019) 2. GSMA, The state of mobile internet connectivity. https://www.gsma.com/r/wp-content/ uploads/2021/09/The-State-of-Mobile-Internet-Connectivity-Report-2021.pdf (2021). Last accessed 14 Jun 2022

82

J. C. Borromeo et al.

3. N. Dragovic, D. Vasiljevic, U. Stankov, M. Vujicic, Go social for your own safety! Review of social networks use on natural disasters – case studies from worldwide. Open Geosci. 11(1), 352–366 (2019) [Online]. Available:https://doi.org/10.1515/geo-2019-0028 4. Forbes, Why don’t cell phones work at music festivals? https://www.forbes.com/sites/ quora/2012/07/31/why-dont-cell-phones-work-at-music-festivals/?sh=7b7f10a17202 (2012). Last accessed 31 Aug 2022 5. F. Rinaldi, H.-L. Maattanen, J. Torsner, S. Pizzi, S. Andreev, A. Iera, Y. Koucheryavy, G. Araniti, Non-terrestrial networks in 5G and beyond: a survey. IEEE Access 8, 165178–165200 (2020) 6. Hexa-X Project, Expanded 6g vision, use cases and societal values. Deliverable D1.2 (2021) 7. H. Nishiyama, Y. Tada, N. Kato, N. Yoshimura, M. Toyoshima, N. Kadowaki, Toward optimized traffic distribution for efficient network capacity utilization in two-layered satellite networks. IEEE Trans. Veh. Technol. 62(3), 1303–1313 (2013) 8. E. Cianca et al., Integrated satellite-HAP systems. IEEE Commun. Mag. 43(12), 33–39 (2005). https://doi.org/10.1109/MCOM.2005.1561921 9. ESA, Satellite Frequency Bands. https://www.esa.int/Applications/Telecomm-unications/ Integrated/Applications/Satellite\T1\textbackslashfrequency/bands (2022). Last accessed: 15 Jun 2022 10. M. Giordani, M. Zorzi, Non-terrestrial networks in the 6G era: challenges and opportunities. IEEE Netw. 35(2), 244–251 (2021) 11. R. Bassoli, F. Granelli, C. Sacchi, S. Bonafini, F.H. Fitzek, CubeSat-based 5G cloud radio access networks: a novel paradigm for on-demand anytime/anywhere connectivity. IEEE Veh. Technol. Mag. 15(2), 39–47 (2020) 12. R. Margolies, R. Becker, S. Byers, S. Deb, R. Jana, S. Urbanek, C. Volinsky, Can you find me now? Evaluation of network-based localization in a 4G LTE network, in IEEE INFOCOM 2017 - IEEE Conference On Computer Communications (2017), pp. 1–9 13. M. De Sanctis, E. Cianca, G. Araniti, I. Bisio, R. Prasad, Satellite communications supporting internet of remote things. IEEE Int. Things J. 3, 113–123 (2016) 14. C. Delmas, C. Hourtolle, Y. Prevot, Formation flight handling with 4 satellites – Electronic Earth Observation as an application, in 2018 SpaceOps Conference (2018) 15. D. CaJacob, N. McCarthy, T. O’Shea, R. McGwier, Geolocation of RF emitters with a formation-flying cluster of three microsatellites, in In 30th Annual AIAA/USU Conference on Small Satellites (2016), pages 1–13 16. N.H. Nguyen, K. Dogancay, Algebraic solution for stationary emitter geolocation by a LEO satellite using Doppler frequency measurements, in 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2016), pp. 3341–3345 17. P.B. Ellis, F. Dowla, Single satellite emitter geolocation in the presence of oscillator and ephemeris errors, in 2020 IEEE Aerospace Conference (2020) 18. S. Bonafini, C. Sacchi, R. Bassoli, K. Kondepu, F. Granelli, F.H. Fitzek, End-to-end performance assessment of a 3D network for 6G connectivity on Mars surface. Comput. Netw. 213, 109079 (2022) [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S1389128622002171 19. M.A. Marotta, H. Ahmadi, J. Rochol, L. DaSilva, C.B. Both, Characterizing the relation between processing power and distance between BBU and RRH in a cloud RAN. IEEE Wirel. Commun. Lett. 7(3), 472–475 (2018) 20. S. Cakaj, B. Kamo, A. Lala, A. Rakipi, The coverage analysis for low earth orbiting satellites at low elevation. Int. J. Adv. Comput. Sci. Appl. 5, 07 (2014) 21. S. Bonafini, C. Sacchi, R. Bassoli, F. Granelli, K. Kondepu, F.H. Fitzek, An analytical study on functional split in martian 3D networks. IEEE Trans. Aerosp. Electron. Syst. 1–10 (2022) 22. J. Lissauer, I. de Pater, Fundamental planetary science: physics, chemistry and habitability (Cambridge University Press, Cambridge, 2013) [Online]. Available: https://books.google.it/ books?id=0iggAwAAQBAJ 23. A. Al-Hourani, Session duration between handovers in dense LEO satellite networks. IEEE Wirel. Commun. Lett. 10(12), 2810–2814 (2021)

3 Futuristic Satellite Scenarios in 6G

83

24. 3GPP, Radio Access Network, Study on New Radio (NR) to support non-terrestrial networks). Technical Report (TR) 38.811, version 15.4.0 (2021) 25. 3GPP, 3rd generation Partnership Project, Technical Specification Group Radio Access Network; Solutions for NR to support non-terrestrial networks (NTN). Technical Report (TR) 38.821, version 2.0.0 (2021) 26. Matlab, Satellite communications toolbox: NR NTN PDSCH throughput (2022). https://de. mathworks.com/help/satcom/ug/nr-ntn-pdsch-throughput.html. Last accessed 12 Oct 2022 27. ITU-R, Deployment and technical characteristics of broadband high altitude platform stations in the fixed service in the frequency bands 6440–6520 MHz, 21.4–22.0 GHz, 24.25–27.5 GHz, 27.9–28.2 GHz, 31.0–31.3 GHz, 38.0–39.5 GHz, 47.2–47.5 GHz and 47.9–48.2 GHz used in sharing and compatibility studies. Technical Report (TR) F.2439-0 (2018) 28. Satbeams, Satellite footprints (2009) [Online]. Available: https://www.satbeams.com/satellitefootprints/page-64, Last accessed 30 Aug 2022 29. A. Gomez, Versatile and robust X-band: impressive performance for MILSATCOM (2019) [Online]. Available: https://www.worldteleport.org/news/442469/Versatile-and-Robust-XBand-Impressive-performance-for-MILSATCOM-.htm, Last accessed: 20 Jun 2022 30. ITU-R, Satellite system characteristics to be considered in frequency sharing analyses between geostationary-satellite orbit (GSO) and non-GSO satellite systems in the fixed-satellite service (FSS) including feeder links for the mobile-satellite service (MSS). Technical Report (TR) S.1328-3 (2001) 31. 3GPP, NTT DOCOMO, CU-DU split: refinement for Annex A (Transport Network and RAN internal functional split). Technical Report (TR) R3-162102 (2016) 32. 3GPP, CMCC, Transport requirement for CU and DU functional split options. Technical Report (TR) R3-161813 (2016) 33. 3GPP, 3rd generation partnership project: user equipment (UE) radio access capabilities. Technical Report (TR) 38.306, version 15.2.0 (2018) 34. NMGN, NMGN Alliance. NMGN 5G White Paper. White Paper (2015)

Chapter 4

Quantum Satellite Communications Sonai Biswas, Riccardo Bassoli, Janis Nötzel, Christian Deppe, Holger Boche, and Frank H. P. Fitzek

4.1 Introduction During 1900s a series of experiments and corresponding theories like wave mechanics, matrix mechanics, particle-wave duality etc. gradually shaped the foundations of Quantum Mechanics. Although Max Planck was the one who first coined the term quantum, there was significant contributions from several other scientists like Albert Einstein, Schrodinger, Otto Stern, Walther Gerlach and many others. In order to enable new applications like ultra-secure communication and a new generation of high-performance computing, quantum communication makes use of the special qualities of quantum entanglement. The range of current quantum networks is constrained. At distances greater than a few hundred kilometers, noise weakens and destroys fragile quantum states. Quantum repeaters can expand quantum networks locally by establishing long-range entanglement over a number of hops. A complementary method is provided by satellites, which transmit qubits

S. Biswas Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany e-mail: [email protected] R. Bassoli () · F. H. P. Fitzek Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany Centre for Tactile Internet with Human-in-the-Loop (CeTI), Dresden, Germany e-mail: [email protected]; [email protected] J. Nötzel · C. Deppe · H. Boche Department of Electrical and Computer Engineering, Technische Universität München, München, Germany e-mail: [email protected]; [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_4

85

86

S. Biswas et al.

straight through the atmosphere. We will cover developments toward satellite-based quantum networks in this post. Along with the advent of quantum, there have been some theoretical advancements across the domain in the quantum communication also. Quantum communication can be established via satellites across considerably greater distances than are now possible on the ground. Photons are a type of quantum state that can be faded easily. Photons are often sent along links in fiber optic cables, the same kind of fiber that is used in the modern internet, to convey them on the ground. A few hundred kilometers is how rapidly photons in fiber are lost to noise. Photons can travel far farther through the atmosphere, commonly referred to as “free space,” before becoming lost.

4.2 Introduction to Quantum Communications Contrary to classical computing the qubits or quantum bits does not necessarily stay in 0 or 1 state and can exist in 0 and 1 simultaneously. The qubits have 2 eigenstates represented as 0 and 1, but there can be states with more than 2 eigenstates like qutrits with three eigenstates (.0, 1 and 2) or ququarts (.0, 1, 2 and 3) or in general n eigenstates which are called ququarts (.0, 1, 2, . . . n − 1, n). Generally most of the experimental simulations deal with qubits and thus we mostly deal with problems of quadratic order and theoretically we can solve problems reducing the higher order problems to quadratic problems [1] and fit it in a quantum computer. Qubits can be represented as follows, .

|q = α|0 + β|1

(4.1)

where .α, β ∈ C and .|α|2 + |β|2 = 1. Here .|. and ..| denotes  the bra-ket notation 1 as described in [2]. Here .|0 represents the column vector . , and .|1 represents 0   0 the column vector . and the .| of any state represents the complex conjugate of 1 the .| state. Also these two states forms the basis of the single qubit state and this basis is called the computational basis and most commonly used. Also the states 1 1 + − .|φ  = √ (|0 + |1) and .|φ  = √ (|0 − |1) forms another basis for single 2 2 qubit states is called Hadamard basis. This basis is used in some cases as we will see in the following sections.

4 Quantum Satellite Communications

87

4.2.1 Superposition The state of a qubit can be written as a column vector in 2-dimensional complex vector space .C2 as described in Eqs. (4.1). Equation (4.1) also describes one of the most fundamental and powerful tool of quantum mechanics i.e. superposition in the sense that the qubit can simultaneously exist in both .|1 and .|0 state. The state of the qubit collapses to one of these two states upon measurement and with repeated experiment we can also determine the component .α and .β by measuring the probability associated with each state. The probability to measure the state .|0 is .|α|2 and that of .|1 is .|β|2 .

4.2.2 Multiple Qubits and Entanglement When there is more than one qubit the combined stated of the qubits can be (not always) written as the Kronecker product of the each of the qubit state. For example if the two qubits are in states .α1 |0 + β1 |1 and .α2 |0 + β2|1 then the two qubit state can be written as, |qq = (α1 |0 + β1 |1) ⊗ (α2 |0 + β2|1) .

= α1 α2 |00 + β1 α2 |10 + α1 β2 |01 + β1 β2 |11

(4.2)

Here .|00, |01, |10 and .|11 represents the basis vectors for the 2 qubit systems and any 2 qubit state can be defined using linear combination of these and upon measuring the above state we will obtain .00, 01, 10 or 11 state according the ⎡ ⎤ 1     ⎢0⎥ 1 1 ⎥ probability associated with it. Here .|00 is denoted as . ⊗ =⎢ ⎣0⎦. The other 0 0 0 basis vectors are also denoted as .4 × 1 column vector like this. The interesting thing however is that not every two qubit state can be explicitly written as a kronecker product of two one qubit system, For example the following state, .

1 qqentangled = √ (|00 + |11) 2

(4.3)

One interesting thing to note here that not only the state can’t be decomposed into Kronecker product but also measuring one of the qubit will affect the state of another qubit. Here the only two outcomes are .|00 and .|11. Note, how the first and second qubits are perfectly correlated, in the sense that when we obtain state 0 in first qubit the state in the second is same i.e. 0 and same goes for state 1. This gives rise to the most powerful tool of quantum mechanics ‘entanglement’.

88

S. Biswas et al.

Fig. 4.1 Bloch sphere representation

4.2.3 Bloch Sphere There is a geometric representation of qubit, in which any single qubit state can be imagined on a sphere called Bloch sphere. The north pole of sphere represents the qubit state .|0, the south pole represents qubit state .|1 and all other points on the sphere represent some linear combination of these two states. As shown in Fig. 4.1 the axis defining .|0 and .|1 is the Z axis and X and Y axis are perpendicular on the equatorial plane. The vector that points to the state on the Bloch sphere from the centre is called Bloch vector. Any single qubit state is related to each other by the rotation of this vector. The angle this vector makes with the Z axis is denoted as .θ and the angle it makes with X is .φ. The most general single qubit state can be represented by these two angles as follows, θ θ | = cos |0 + eiφ sin |1 2 2

.

(4.4)

This can be thought equivalent to the Eq. (4.1) where the arbitrary qubit state depends on the two parameters .α and .β. However in this case the parameters .θ and .φ are not linear coefficients as in the former case.

4.2.4 Quantum Computing Gates There are quantum computing gates, some of which are analogous to classical computing and some are completely unique and gives the quantum computing an competitive edge over the classical counterpart. We will start with the single qubit gates, then look at some 2 qubit gates and give a hint of more complex gates.

4 Quantum Satellite Communications

4.2.4.1

89

Single Qubit Gates

X, Y and Z are three single qubit gates that rotates .π around the axis X, Y and Z respectively around the Bloch sphere. For example if X gate is applied the qubit state changes from .|0 to .|1 and vice versa. So this gate can be thought to be analogous to NOT gate in classical computing. On the other hand we have a special gate called hadamard (denoted by H ) which creates equal superposition state . √1 (|0 + |1) 2 which is equivalent to a .π/2 rotation around Y axis followed by a .π rotation around X axis. Apart from these basic gates there are parametrized gates like Rx, Ry and Rz which rotates the Bloch vector for an arbitrary angle around X, Y and Z axis respectively, these gates particularly important for Quantum Machine Learning instances, where one needs to find optimized angle for the gates based on a certain model. All quantum gates can also be visualized as matrix operating on qubit state  .0 1 column vectors. For instance X gate is denoted by, . . Following table shows 1 0 some of the basic single quantum gates with their matrix representation, Quantum gate X

Y

X

H

Matrix

 01 . 10 

0 −i . i 0

 01 . 10

 01 . 10

4.2.5 2 Qubit and Multi-Qubit Gates Apart from creating superposition another powerful tool of quantum computing is to create entanglement. To create entanglement between qubits, a combination of H gate and a special two-qubit gate called CNOT is used. CN OT gate is defined as follows, it has a control qubit and a target qubit, depending on the value (.|1) of control qubit it flips the state in the target qubit. Due to this flipping operation it is also called as CX or controlled X gate. The matrix of the CNOT is as follows,

90

S. Biswas et al.

⎡

1 ⎢0 . CN OT := ⎢ ⎣0 0

0 1 0 0

0 0 0 1

⎤ 0 0⎥ ⎥ 1⎦ 0

(4.5)

In a quantum circuit the gates are read from left to right and if parallel to single qubit  gate there is no gate applied on other qubit, it represents an identity 10 I gate . and the two qubit gate can be considered as a tensor product of these 01 two single qubit gates and any multi-qubit gate can be constructed similar way. Now note the entanglement creating circuit has an H gate (with I gate on the other) and CNOT gate, it is easy to carry out all the matrix multiplication and the final outcome would be . √1 (|00 + |11). Intuitively the CN OT gate alters the 2 second qubit whenever it sees .|1 and keeps it same otherwise and because of equal superposition we get the above output. There are more complex multi-qubit gates like controlled rotations, SWAP gates, controlled SWAP gates, controlled unitary, N-controlled NOT gate and many more all of which can be constructed by some combinations of the gates discussed here. [3] gives a very good detailed discussion on all sorts of complex and generalized quantum gates.

4.2.6 Bell State Measurement One of most powerful tool in quantum computing is Bell State Measurement (BSM) or Bell Basis Measurement. In the previous section an particular two qubit entangled state .qqentangled is discussed. There are in total 4 possible entangled states in 2qubits system listed as follows, 1 φ + = √ (|00 + |11) 2 1 φ − = √ (|00 − |11) 2 . 1 ψ + = √ (|01 + |10) 2 1 ψ − = √ (|01 − |10) 2

(4.6)

These states are readily used in various quantum cryptography and other protocols. To distinguish between these states a BSM is carried out at the reciever’s end. The BSM is simply a combination of CNOT and Hadamard gate (Fig. 4.2).

4 Quantum Satellite Communications

91

Fig. 4.2 Bell state measurement circuit

4.2.7 Entanglement Swapping If between two parties let’s say Alice and Bob both contain a .φ + state (also called Bell pair) each. They want to establish an entanglement among their qubits. Let’s denote the four qubits by .A1 , A2 , B1 and .B2 and the quantum states. The initial state is given by, +  = φA ⊗ φB+1 B2 1 A2

.

(4.7)

Here .φ.. state signifies the state . √1 (|00 + |11). 2 There is a third party let’s say Carol and both Alice and Bob send their .A1 and .B1 to Carol. Let’s denote qubits belonging to carol by C. The state is, + ψC = φA ⊗ φB+1 1

.

(4.8)

Next Carol performs a BSM on these two qubits and the resulting state will be φA2 B2 and the Alice and Bob now will have entanglement between qubits .A2 and .B2 . For other entangled states there will be a additional single qubit operations on the qubit .B2 . .

4.2.8 Measurement One of the crucial and most intriguing part of quantum mechanics is the measurement. The quantum state behaves quantum quantum mechanically until its measured, or in other words the quantum properties described above, of the qubit exists until it is measured. So let’s say we have a single particle in a state 1 . √ (|0 + |1), upon measuring only 0 or 1 will be observed. This concept is also 2 described as the collapse of the wavefunction.

4.2.9 Nondemolition Measurement Quantum Nondemolition (QND) measurement [4] is also a very interesting and sophisticated measurement technique in quantum mechanics which allows us to

92

S. Biswas et al.

measure the quantum system with minimally disturbing the quantum system. The necessary requirement for this to take place is that the variable on which we are making measurement is related with overall system. For example if we make a measurement on momentum of a free particle can be QND provided the energy of the free particle remains the same and it will make sure during the next measurement the momentum quantum system is undisturbed.

4.3 State of the Art in the Quantum Satellite Communications There are a few quantum protocols that have been experimentally realized in the field of quantum communications like Quantum Key Distribution (QKD), entanglement-based QKD, Entanglement Distribution (ED), ground-to-satellite quantum teleportation. Also there have been experiments in Space lab, microsatellite experiments like SOCRATES and CubeSats. In the following subsections each of these protocols are discussed in more details.

4.3.1 Quantum Key Distribution Although first proposed in 1970s the QKD did not catch much attention until 1980s. QKD is an information-theoretically secure way of sharing two random secret keys between two parties, which can be further used for encryption or decryption of classical communication messages. Contrary to the classical mechanics the security of this protocol does not depend on the computational complexity of the so-called one way function, which is computationally hard. The security of this protocol comes from the fundamentals of quantum mechanics. As described in the original paper of 1984 by Bennett and Brassard [5] the first QKD protocol, also known as BB84 protocol, utilizes the no-cloning theorem [6, 7] of quantum mechanics, which affirms that producing a identical and independent copy of an arbitrary quantum state is impossible. The BB84 protocol starts as follows, Between two parties (a sender and a receiver), let’s say Alice and Bob, Alice wants to share a private key with Bob. She has two bitstrings a and b each of length n. Now she will encode the these bitstrings in n qubits as follows,

.

=

n

|ψai bi 

(4.9)

i=1

Now one can notice that the .|ψai bi  will have four possibilities in the sense that each of .ai and .bi will be either 0 or 1, and these four states are encoded as follows,

4 Quantum Satellite Communications

93

|ψ00  = |0 |ψ10  = |1 .

1 |ψ01  = √ (|0 + |1) 2 1 |ψ11  = √ (|0 − |1) 2

(4.10)

Interesting thing here is that the bit .bi decides the basis of the qubit state, if .bi = 0 the qubit is in computational basis and if the .bi = 1 the qubit is in Hadamard basis. Also the four states are not orthogonal states and thus it is impossible to distinguish between them without knowing the bit value .bi . Now if there is a third party, let’s say Eve wants to copy the qubit by eavesdropping into the quantum channel it is impossible for her to recreate and send the exact qubit over the channel if she makes a measurement due to no-cloning theorem. In the next part of the protocol Bob randomly generate a bitstring .b of length n and measures the qubits to get the estimate of a, let’s say .a  . Next Alice shares the original bitstrings b over public channel and Bob discards the bits .ai where b and .b don’t match. Now, from the remaining k bits, where the b and .b match, Alice randomly chooses .k/2 bits and sends to Bob through public channel and both of them check if a certain number of bits have matched or not. Ideally there should be cent percent match in the absence of Eve but due to noise some qubits may be transformed and according to a pre-determined threshold if the results match then Alice and Bob proceeds with the protocol. Otherwise they confirm the channel is compromised and they discard it and start over.

4.3.2 Entanglement Distribution The experimental observation of the existence of entanglement itself was awarded with a Nobel price for Alain Aspect, John F. Clauser, and Anton Zeilinger in 2022. Entanglement is mathematically engraved into the mathematical formalism of quantum mechanics via pure states .|ψ ∈ H ⊗ H, where .H denotes a Hilbert space, which cannot be decomposed into the form |ψ = |u ⊗ |v.

.

(4.11)

A particular instance of such a state is given by 1 |ψ = √ (|00 + |11). 2

.

(4.12)

94

S. Biswas et al.

Due to the monogamy of entanglement, a perfectly entangled pure state which is shared between two parties is completely decorrelated from any other state. In addition, perfectly entangled states can be used to generate so-called non-local correlations. This type of correlation formed the core of theoretical arguments leading to the eventual proof of existence for entanglement. The prototypical example of such correlations is created by using a setup consisting of a shared entangled state, for example 1 |ψ = √ (|00 + |11), 2

.

(4.13)

and then acting on this state using, for each of the two systems, two possibly different measurement setups. Let these be indexed by .A1 , A2 and .B1 , B2 . If .A1 , A2 and .B1 , B2 are realized based on the use of local unitary rotations followed by measurements in the computational basis, then the resulting statistics of the experiment takes the form (x, y) = (2, 2) ⇒ ∀i = j ∈ {1, 2} : .

(4.14)

.

q(i, i|x, y) = t/2, q(i, j |x, y) = (1 − t)/2. x = y = 2 ⇒ ∀i = j ∈ {1, 2} : .

(4.15) (4.16)

q(i, i|x, y) = (1 − t)/2, q(i, j |x, y) = t/2.

(4.17)

√ where .t := (1 + 1/ 2)/2. The fact that it is impossible to create this conditional probability distribution between two remote parties holding the two separate parts of the quantum system without using communication has led to their contextualization in the area of low-latency [8, 9] and low-complexity multi-access data transmission systems. The derivation of the particular conditional probability distribution q can be picked up from [8, Lemma 11]. In more generality, a conditional probability distribution .q  is called a “local correlation” if there is a probability distribution p such that q  (i, j |x, y) =

.

p(e)q1 (i|e, x)q2 (j |e, y).

(4.18)

e

It is called non-signalling if ∀ i, x, y, y  :

.

q  (i, j |x, y) =

j

∀ j, x, x  , y :

i

q  (i, j |x, y  ).

(4.19)

q  (i, j |x  , y)

(4.20)

j

q  (a, j |x, y) =

i

An example for a non-signalling correlation is given by q.

4 Quantum Satellite Communications

95

4.3.3 Entanglement Based QKD In [10] the author has proposed a QKD protocol using entanglement. Here, we describe a slightly modified version of this protocol. In this protocol Axel and Bayley want to create a provably secrete key between them. Starting from a series of entangled states .|ψ, they measure using three different measurement devices: Axel uses .A1 , A2 , A3 where .A1 , A2 are the devices described above and Bayley uses .B1 , B2 , B3 where again .B1 , B2 are as described above. For the third setup it must hold .A3 = B3 , and for simplicity we can assume .A3 = {0|0, 1|1}. In this protocol, there exist four steps. Assuming N quantum states have been shared between them, Axel and Bayley first randomly select between the three measurement setups they have at their disposal. They then carry out the respective measurements, resulting in two strings N for Axel and .(b , n )N where the values .a , b ∈ {1, 2, 3} denote the .(ai , mi ) i i i=1 i i i=1 measurement setups and the values .mi , ni ∈ {1, 2} the respective measurement outcomes. In the second stage, Axel and Bayley publicly reveal the strings .(ai )N i=1 and N . .(bi ) i=1 In the third stage they reveal the measurement results for that part of the measurement result where they have both chosen one of the four possible combinations .(Ai , Bj ) for which .i, j ∈ {1, 2}. Let these measurement results and  corresponding measurement setups be labelled .M  = (ai , bi , mi , ni )N i=1 . Based on this announcement, they can each verify that the empirical average of these results is (sufficiently) close to .q(i, j |x, y). If the verification is successful, they proceed with step four. Otherwise, they abort. In the fourth and final step, Axel and Bayley calculate the secret key based on the measurement results where they both made the basis choice .2, 2. Key ingredients here are non-locality, which leads to the result that the function 

N

.f (M ) := (mi − 1)(ni − 1)s(ai , bi ) 

(4.21)

i=1

√ (where .s(a, b) = −1 only if .a = b = 2 and .s(a, b) = 1 else) converges to .2 2 if the shared quantum states are actually all equal to .|ψ. In sharp contrast, if an eavesdropping attack had taken place on all shared quantum states, this would have destroyed the shared entanglement. In this case, the value of .f (M  ) would converge to 2 when N (and .N  ) grows. If any large enough fraction of the qubits had been  corrupted √ by an eavesdropper, the value of .f (M ) would still differ from the ideal value .2 2. Since this makes it practically impossible for Axel and Bayley to control the success probability of their protocol, privacy amplification is used in conjunction with the quantum key distribution protocol.

96

S. Biswas et al.

4.4 Why Satellite-Based Quantum Communication Once we have understood the fundamentals of quantum mechanics the natural question that arises is why using satellite based quantum communication specially when there is a high cost related to launching and establishing a satellite in a orbit around the earth. We have other options like let us say, optical fibre based quantum communication where a photon’s orbital angular momentum or photon’s polarization can be used a resource for entanglement. Or using terrestrial free space for communicating seems to be cheaper and good option. However with ground based systems the entanglement distribution has achieved only few hundred kilometers [11] distances. For example with the use of telecommunication fibers for 19 MHz count rate and for a loss of .0.16 dB/km, one can distribute an entangled photon pair for a distance of 600 km with certainty .10−12 photons coincide per second. The entanglement can’t be established simply due to low signal to noise ratio. One natural solution could be using quantum repeater [12], which divides the entire transmission line into smaller segments and using the tools from entanglement swapping, entanglement purification [13] and quantum memory [14]. Although there has been some considerable research works already done in these components particularly as a proof-of-concept [15],[16], practically the quantum memories still suffer some critical drawbacks like low storage times, low retrieval efficiency [17]. With these lack of inabilities to procure large distances in the next section some of key achievements in recent quantum satellite communications has been highlighted.

4.5 Recent Trends in Quantum Satellite Communications Although some works like [18] had given the foundation of long distance communication using quantum technologies the actual experiments with satellites started after the year 2015. In the following flowchart one can find the timeline of the progress in the quantum satellite communications experiments and important breakthroughs.

4.5.1 Satellite-to-Ground Communication In [19] the authors have entangled the polarization of two photons in order to communicate between ground and satellite. Here, satellite corner cube retroreflectors were used as quantum transmitters in orbit and the Matera Laser Ranging Observatory of the Italian Space Agency as a quantum receiver to show quantum communications with polarization encoding from space to ground. As the violation of Bell inequalities or quantum key distribution, the Quantum Bit Error Rate

4 Quantum Satellite Communications

97

Fig. 4.3 Quantum satellite communications using quantum repeaters for global distances. QMs are quantum memories

(QBER) was maintained steadily low to a level suitable for QKD. In fact, the authors worked with an average value of QBER .4.6% for a total link duration of 85 s using data from various satellites.

4.5.2 Entanglement with Multiple Satellite Links In [20] the authors have proposed a quantum repeater based quantum satellite communications with multiple Satellite communication links using the QND property and quantum memories for storing quantum information. In the proposed architecture of the quantum repeater the distance between each ground station is .L0 and the satellites contain the source of entangled photon pair. The satellites are placed in a low earth orbit with a height of h and the complex quantum mechanical tools are situated in the ground stations. The communication takes place with the following recipe, first the entangled photons are transmitted to two distant ground stations and QND devices just detects the presence of photons without disturbing the quantum information stored in it and it is kept in the quantum memories until the quantum entanglement in two distant ground stations is achieved or both the QND devices detect photon. Next using the BSM one of the entangled state between the adjacent ground stations is transferred (Fig. 4.3).

98

S. Biswas et al.

4.5.3 Continuous Variable Entanglement Distribution In 2015, [21] have proposed two different protocols for Continuous Variable (CV) entanglement distribution, one with the entanglement source onboard the satellite and the other the ground stations equipped with the entanglement source. The first protocol the entanglement is created at a ground station and then the photon beam is reflected from the satellite through two independent atmospheric fading channel. The protocol has the potential benefit of reliability due to the source of entanglement being stationed on the ground. The protocol simply uses a Low Earth Orbit (LEO) satellite as a reflector to transfer the entanglement information to another ground station. After creating an initial Gaussian entanglement between two photons. The schematic for the same is as follows, In the second protocol the entanglement is produced on the satellite and then the photon is transmitted through two different ground station through fading downlink channels. The schematic is as follows, The third protocol the entanglement is generated at each of the ground station and then one of the photon from each entanglement pair is shared via entanglement swapping to the satellite. For the low loss fading channel the CV entanglement generation rate is very similar for both ground entanglement generation and satellite entanglement generation. On the contrary for high loss case generating entanglement at ground directly transmitting it via satellite reflection has the advantage. Only for the case when downlink channels are significantly better in terms of low loss and fading, entanglement generation on the satellite is favourable, provided the cost of the setup of quantum devices in the satellite. In [22] the author has also proposed an another model in which a different entanglement model is used to interpret the effect of atmospheric noise for distributing entanglement. Although the work done is more restricted to the model of generating entanglement in the satellite and then distributing it to two distant ground stations.

4.5.4 Quantum Information Transfer Through Free Space 4.5.4.1

Single Photon Transfer along 7000 km in Space

When it comes to quantum satellite communications, not only entanglement generation or processing, but also sharing the entanglement (here photon) over large distance in free space is also a very important component of the entire setup. In [23] a single photon exchange for a slant distance of 7000 km between ground station and satellite is achieved via Medium Earth Orbit (MEO) satellite. The experiment was particularly important as with increasing distance the effect of relativity becomes prominent and showed that there might be a way towards geostationary satellite to build the testbed for combined effect of quantum and general relativity.

4 Quantum Satellite Communications

4.5.4.2

99

Optical Signals Through 38,600 km into The Atmosphere

A very good follow up work can be found in [24], where a optical signal is transferred through the turbulent atmosphere from a GEO satellite to the ground station for over a distance of 38,600.

4.5.4.3

QKD Through Inter-Satellite Free-Space Links

On Earth, QKD is shown in free space. In [25] experiment, no satellites were used, but the researchers demonstrated that QKD could be carried out through freespace rather than a direct fiber connection. Since satellites also rely on free-space communication, this is an important step towards satellite communications.

4.5.5 Breakthrough in Quantum Satellite Communications Around 2017 Jian Wei Pan’s team’s work was a breakthrough in the quantum satellite communications in the sense that for the first time, in 2017 [26], the group experimentally used a satellite to demonstrate the quantum communications over a long distance.

4.5.5.1

Satellite-to-Ground QKD and Ground-to-Satellite Quantum State Transfer

The team led by Jian-Wei Pan used the Micius satellite [27] to demonstrate two crucial uses for a quantum communication satellite in addition to their entanglement demonstration. For secure long-distance communication, QKD was first demonstrated over a distance of 1200 km. Then, quantum teleportation, a crucial component required for real-world quantum networks and applications beyond QKD, was proven. This work made the path towards a complete quantum satellite communications by enabling first the satellite-to-ground QKD [27] and then groundto-satellite quantum teleportation[28].

4.5.6 Developments Towards a Global Network 4.5.6.1

Satellite Based Quantum Intercontinental Network

Jian-Wei Pan and their team continued to their work in quantum satellite communications by showing QKD between China and Austria in 2018 [29]. The researchers displayed a conference call that was QKD-secured. Since the satellite served as a

100

S. Biswas et al.

‘trusted’ node, the security of the communication assumes that the satellite is secure. In the future, the researchers are going to accomplish real entanglement-based QKD by removing this assumption.

4.5.6.2

Global Navigation Satellite System Using Quantum Communication

The majority of earlier satellite-based demonstrations relied on LEO satellites, which orbit the earth at a somewhat closer distance. Lower noise is produced by the closer distance, however there may be some downsides. For instance, a LEO satellite is only briefly visible to a particular ground station. In [30] researchers demonstrated single-photon communication utilizing a GNSS network satellite at a distance of 20,000 km. This provided a proof-of-concept for the use of high-orbit satellites in combination with GNSS network.

4.5.7 Secured Quantum Entanglement-Based Cryptography Demonstration Using Satellites Jian-Wei Pan and their team demonstrated QKD with the Micius satellite as a trusted node. This enabled secure long-distance communication, but it was predicated on the premise that the satellite was kept secure. The researchers demonstrated entanglement-based QKD between two ground stations separated by 1120 km at a secret-key rate of .0.12 bits per second in this work, eliminating the necessity for a reliable relay. A crucial use of Entanglement-as-a-Service networks is Entanglement-based QKD, which enables real physics-based security without the requirement for trusted relays.

4.5.8 Moving Boundaries from Lab to Real World In 2021, quantum satellite communications moved from the mere experiments in labs to real world scenarios.

4.5.8.1

A 4600 km Comprehensive Quantum Communication Network from Space to The Ground

Jian-Wei Pan and colleagues, who are still at the forefront of the development of satellite quantum communication, have demonstrated a fully integrated QKD network spanning 4600 km [31] with a mixture of satellite and ground nodes. The

4 Quantum Satellite Communications

101

network, which uses trusted relays to increase its range, is made up of two satelliteto-ground free-space QKD links in addition to a fiber network with more than 700 fiber QKD links. The satellite links demonstrated integration in a complete network and also attained greater key rates than had previously been demonstrated, up to .47.8 kb/s for a typical satellite pass.

4.5.9 An Overview of QKD experiments on-Orbit Over the Years

Year 2012 2013 2014 2016 2018 2019

Achievement Basic miniaturized SPDC SPDC intercorrelated system SPDC source who is associated and suited for space survives rocket explosion Correlated, space-qualified SPDC source in low-Earth orbit Demonstration of entangled photon pair source in space QKD experiments using optical networks in orbit

Mission Balloon test in high altitutde[32] High altitude Helium Balloon Test[33] GomSpace GomX-2 CubeSat[34] NUS Galassia CubeSat[35] CQT SpooQy-1 CubeSat[36] International collaborations[37]

4.6 Final Remarks The quantum satellite communications has seen a rapid improvement in the experiments carried out in last 5–6 years. Although entirely functioning end-toend quantum satellite communications are not yet implemented, the individual key components for quantum satellite communications has been well identified. The three major components of quantum satellite communications that is entanglement source, channel (free-space transmission of photon) and the receiver, have seen some major improvements in different scientific researches that have been carried out. Although long distance photon transmission, carrying quantum information earlier considered to be extremely hard, has been tackled up to a large extent recently [23, 24]. For entanglement generation multiple conflicting ideas are now being proposed. One being the entanglement source to be stationed at ground and the other being stationed in orbit inside the satellite. Despite the cost of maintaining the sophisticated quantum devices in orbit, this idea is mostly preferred in the current scientific community due to the shorter distance the photon needs to travel.

102

S. Biswas et al.

4.7 Current Limitations in Quantum Satellite Communications From the two different viewpoints of quantum satellite communications there are few obstacles to be overcome for implementing the quantum satellite communications in public networks. • Most of the quantum satellite communications protocols proposed recently have their entanglement source located in the satellite. It can be tricky and moreover costly to maintain when it comes to dealing with sophisticated quantum devices located high into the orbits. • Pirandola [38] have theoretically given an estimate on the geometric distance estimate for the satellites. In the same work the author have also formulated the bounds arising from the atmospheric disturbance, free space diffraction, background noise and fading in uplink and downlink satellites. • As quantum devices specifically generating entanglement is sophisticated, the dimension of the entire satellite to be launched (if quantum devices are sent in orbit) becomes huge. There have been some work in designing a nano-satellite [36, 39] for quantum communication but it is still far from being absorbed in real world satellite communication. This is big hindrance towards launching more quantum satellite missions as it increases the overall costs of the mission drastically. • Another major concern is decoherence due to atmospheric turbulence. One important thing to keep in mind that sending a quantum information using photon is completely different from sending a classical information over amplitude or frequency modulated waves. In classical communications also there are issues of fading of information due to atmosphere but in quantum domain, properties like polarization of a photon can be damaged or altered more likely due to inhomogeneous and medium. That is why most of the experiments so far only deals with low QBER.

4.8 Future Works in Quantum Satellite Communications The UK-based company Arqit recently disclosed plans to start constructing a QKD network utilizing satellites. The satellites will be launched by Virgin Galactic in 2023. With this, satellite quantum communication will make its way from the public to the private sphere. Apart from this there has been multiple initiatives from Germany, US and China in order to establish a full fledged quantum networks across large distances. Some of interesting directions in this regard is as follows, So far most of the in the domain of quantum satellite communications has been restricted to more or less 1 satellite, or sometimes 2 satellites. But in order to be an infrastructure for future generation networks, wider coverage around the globe is needed.

4 Quantum Satellite Communications

103

References 1. A. Mandal, A. Roy, S. Upadhyay, H. Ushijima-Mwesigwa, Compressed quadratization of higher order binary optimization problems, in 2020 Data Compression Conference (DCC) (2020), pp. 383–383. https://doi.org/10.1109/DCC47342.2020.00090 2. M. Nielsen, I. Chuang, Quantum Computation and Quantum Information, 10th Anniversary edn. (Cambridge University Press, Cambridge, 2010). https://doi.org/10.1017/ CBO9780511976667 3. A. Barenco et al., Elementary gates for quantum computation. Phys. Rev. A 52(5), 3457 (1995) 4. B.B. Vladimir, Y.I. Vorontsov, K.S. Thorne, Quantum nondemolition measurements. Science 209(4456), 547–557 (1980) 5. C. Bennett, G. Brassard, Proceedings of the IEEE International Conference on Computers, Systems and Signal Processing, Bangalore (IEEE, New York, 1984), p. 175 6. J.L. Park, The concept of transition in quantum mechanics. Found. Phys. 1, 23–33 (1970). https://doi.org/10.1007/BF00708652 7. W.K. Wootters, W.H. Zurek, A single quantum cannot be cloned. Nature 299(5886), 802–803 (1982) 8. A. Cacioppo, J. Nötzel, M. Rosati, Compound channel capacities under energy constraints and application, in IEEE International Symposium on Information Theory (ISIT), Melbourne, Australia (2021), pp. 640–645. https://doi.org/10.1109/ISIT45174.2021.9518144 9. S. Sekavˇcnik, J. Nötzel, Scaling of entanglement-assisted communication in amplified fiber links (2022). arxiv, https://arxiv.org/abs/2211.13296 10. A.K. Ekert, Quantum cryptography based on bell’s theorem. Phys. Rev. Lett. 67, 661–663 (1991) 11. J. Yin, J.G. Ren, H. Lu, Y. Cao, H.L. Yong, Y.P. Wu, C. Liu, S.-K. Liao, F Zhou, Y. Jiang, X.D. Cai, P. Xu, G.-S. Pan, J.-J. Jia, Y.-M. Huang, H. Yin, J.-Y. Wang, Y.-A. Chen, C.-Z. Peng, J.W. Pan, Quantum teleportation and entanglement distribution over 100-kilometre free-space channels. Nature 488(7410), 185–188 (2012) 12. H.J. Briegel, W. Dür, J.I. Cirac, P. Zoller, Quantum repeaters: the role of imperfect local operations in quantum communication. Phys. Rev. Lett. 81(26), 5932 (1998) ˙ 13. M. Zukowski, A. Zeilinger, M.A. Horne, A.K. Ekert, ‘Event-ready-detectors‘Bell experiment via entanglement swapping. Phys. Rev. Lett. 71(26), 4287 (1993) 14. A.I. Lvovsky, B.C. Sanders, W. Tittel, Optical quantum memory. Nat. Photonics 3(12), 706– 714 (2009) 15. Z.S. Yuan, Y.A. Chen, B. Zhao, S. Chen, J. Schmiedmayer, J.W. Pan, Experimental demonstration of a BDCZ quantum repeater node. Nature 454(7208), 1098–1101 (2008) 16. C.W. Chou, J. Laurat, H. Deng, K.S. Choi, H. De Riedmatten, D. Felinto, H.J. Kimble, Functional quantum nodes for entanglement distribution over scalable quantum networks. Science 316(5829), 1316–1320 (2007) 17. S.J. Yang, X.J. Wang, X.H. Bao, J.W. Pan, An efficient quantum light–matter interface with sub-second lifetime. Nat. Photonics 10(6), 381–384 (2016) 18. G. Ribordy, J. Brendel, J.D. Gautier, N. Gisin, H. Zbinden, Long-distance entanglement-based quantum key distribution. Phys. Rev. A 63(1), 012309 (2000) 19. G. Vallone, D. Bacco, D. Dequal, S. Gaiarin, V. Luceri, G. Bianco, P. Villoresi, Experimental satellite quantum communications. Phys. Rev. Lett. 115, 040502 (2015) 20. K. Boone, J.-P. Bourgoin, E. Meyer-Scott, K. Heshami, T. Jennewein, C. Simon, Entanglement over global distances via quantum repeaters with satellite links. Phys. Rev. A 91, 052325 (2015) 21. N. Hosseinidehaj, R. Malaney, Gaussian entanglement distribution via satellite. Phys. Rev. A 91, 022304 (2015) 22. N. Hosseinidehaj, R. Malaney, Entanglement generation via non-Gaussian transfer over atmospheric fading channels. Phys. Rev. A 92(6), 062336 (2015) 23. D. Dequal, G. Vallone, D. Bacco, S. Gaiarin, V. Luceri, G. Bianco, P. Villoresi, Experimental single-photon exchange along a space link of 7000 km. Phys. Rev. A 93(1), 010301 (2016)

104

S. Biswas et al.

24. K. Günthner, I. Khan, D. Elser, B. Stiller, Ö Bayraktar, C.R. Müller, K. Saucke, D. Tröndle, F. Heine, S. Seel, P. Greulich, H. Zech, B. Gütlich, S. Philipp-May, C. Marquardt, G. Leuchs, Quantum-limited measurements of optical signals from a geostationary satellite. Optica 4(6), 611–616 (2017) 25. S.K. Liao, H.L. Yong, C. Liu, G.L. Shentu, D.D. Li, J. Lin, H. Dai, S.Q. Zhao, B. Li, J.Y. Guan, W. Chen, Long-distance free-space quantum key distribution in daylight towards inter-satellite communication. Nat. Photonics 11(8), 509–513 (2017) 26. J. Yin, Y. Cao, Y.-H. Li, S.-K. Liao, L. Zhang, J.-G. Ren, W.-Q. Cai, W.-Y. Liu, B. Li, H. Dai, G.-B. Li, Q.-M. Lu, Y.-H. Gong, Y. Xu, S.-L. Li, F.-Z. Li, Y.-Y. Yin, Z.-Q. Jiang, M. Li, J.-J. Jia, G. Ren, D. He, Y.-L. Zhou, X.-X. Zhang, N. Wang, X. Chang, Z.-C. Zhu, N.-L. Liu, Y.A. Chen, C.-Y. Lu, R. Shu, C.-Z. Peng, J.-Y. Wang, J.-W. Pan, Satellite-based entanglement distribution over 1200 kilometers. Science 356(6343), 1140–1144 (2017) 27. S.K. Liao, W.Q. Cai, W.Y. Liu, L. Zhang, Y. Li, J.G. Ren, J. Yin, Q. Shen, Y. Cao, Z.P. Li, F.Z. Li, Satellite-to-ground quantum key distribution. Nature 549(7670), 43–47 (2017) 28. J.G. Ren, P. Xu, H.L. Yong, L. Zhang, S.K. Liao, J. Yin, W.Y. Liu, W.Q. Cai, M. Yang, L. Li, K.X. Yang, Ground-to-satellite quantum teleportation. Nature 549(7670), 70–73 (2017) 29. S.K. Liao, W.Q. Cai, J. Handsteiner, B. Liu, J. Yin, L. Zhang, D. Rauch, M. Fink, J.G. Ren, W.Y. Liu, Y. Li, Satellite-relayed intercontinental quantum network. Phys. Rev. Lett. 120(3), 030501 (2018) 30. L. Calderaro, C. Agnesi, D. Dequal, F. Vedovato, M. Schiavon, A. Santamato, V. Luceri, G. Bianco, G. Vallone, P. Villoresi, Towards quantum communication from global navigation satellite system. Quantum Sci. Technol. 4(1), 015012 (2018) 31. Y.A. Chen, Q. Zhang, T.Y. Chen, W.Q. Cai, S.K. Liao, J. Zhang, K. Chen, J. Yin, J.G. Ren, Z. Chen, S.L. Han, An integrated space-to-ground quantum communication network over 4,600 kilometres. Nature 589(7841), 214–219 (2021) 32. W. Morong, A. Ling, D. Oi, Quantum optics for space platforms. Opt. Photonics News 23(10), 42–49 (2012) 33. Z. Tang, R. Chandrasekara, Y.Y. Sean, C. Cheng, C. Wildfeuer, A. Ling, Near-space flight of a correlated photon system. Sci. Rep. 4(1), 1–5 (2014) 34. Z. Tang, R. Chandrasekara, Y.C. Tan, C. Cheng, K. Durak, A. Ling, The photon pair source that survived a rocket explosion. Sci. Rep. 6(1), 1–5 (2016) 35. R. Chandrasekara, Z. Tang, Y.C. Tan, C. Cheng, L. Sha, G.C. Hiang, D. Oi, A. Ling, Generation and analysis of correlated pairs of photons on board a nanosatellite, in Quantum Information Science and Technology II, vol. 9996 (SPIE, 2016), pp. 43–49 36. R. Bedington, X. Bai, E. Truong-Cao, Y.C. Tan, K. Durak, A. Villar Zafra, J.A. Grieve, D.K. Oi, A. Ling, Nanosatellite experiments to enable future space-based QKD missions. EPJ Quantum Technol. 3(1), 1–10 (2016) 37. D.K. Oi, A. Ling, G. Vallone, P. Villoresi, S. Greenland, E. Kerr, M. Macdonald, H. Weinfurter, H. Kuiper, E. Charbon, R. Ursin, CubeSat quantum communications mission. EPJ Quantum Technol. 4, 1–20 (2017) 38. S. Pirandola, Satellite quantum communications: fundamental bounds and practical security. Phys. Rev. Res. 3, 023130 (2021) 39. T. Jennewein, C. Grant, E. Choi, C. Pugh, C. Holloway, J.P. Bourgoin, H. Hakima, B. Higgins, R. Zee, The NanoQEY mission: ground to space quantum key and entanglement distribution using a nanosatellite, in Emerging Technologies in Security and Defence II; and QuantumPhysics-Based Information Security III, vol. 9254 (SPIE, 2014), p. 925402

Part II

Systems and Infrastructures

Chapter 5

Ground and Space Hardware for Interplanetary Communication Networks Paolo Tortora, Dario Modenini, Marco Zannoni, Edoardo Gramigna, Eliseo Strollo, Andrea Togni, Enrico Paolini, Lorenzo Valentini, Oreste Cocciolillo, and Lorenzo Simone

5.1 Introduction Ground infrastructures are critical for operations in space. From sending commands to remotely operated probes to downloading critical data from scientific payloads, the global networks of ground stations cooperate to guarantee almost continuous communication with deep space missions. Ground-based hardware, described in Sect. 5.2, is complemented by space-based equipment (on-board transponders, amplifiers, and antennas—described in Sect. 5.3) which represent the flying communication assets. However, when the distance between the transmitter and the receiver becomes extremely large, deep space missions represent a challenging wireless communication scenario due to the relative visibility geometry—which involves at least two major occulting bodies—and the additional attenuation of the signal power due to propagation through the Earth troposphere and ionosphere. EHF and/or optical frequencies may potentially increase the data throughput for a given error

P. Tortora () · D. Modenini · M. Zannoni Department of Industrial Engineering, Interdepartmental Center for Industrial Research in Aerospace, Alma Mater Studiorum - Università di Bologna, Forlì, Italy e-mail: [email protected]; [email protected]; [email protected] E. Gramigna · E. Strollo · A. Togni Department of Industrial Engineering, Alma Mater Studiorum - Università di Bologna, Forlì, Italy e-mail: [email protected]; [email protected]; [email protected] E. Paolini · L. Valentini Department of Electrical, Electronic and Information Engineering “Guglielmo Marconi”, Interdepartmental Center for Industrial Research in Aerospace, Cesena, Italy e-mail: [email protected]; [email protected] O. Cocciolillo · L. Simone Radio-Communication and TTC Products, Thales Alenia Space Italia, Rome, Italy e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_5

107

108

P. Tortora et al.

rate level. On the other hand, as frequency grows, so do atmospheric and ionospheric impairments. To partially overcome such a drawback, data relay communication architectures represent an attractive solution. In such a scenario, the link is split into multiple hops (two, as a baseline). The user node, say an orbiter around an inner or outer planet, would transmit the data to a relay spacecraft. This latter would then forward the link to an Earth-based ground station. The potential advantage of such an approach, presented in Sect. 5.4, is that the deep-space link would not suffer from atmospheric losses. The obvious drawback lies in that the intermediate relay node(s) would make use of antenna terminals significantly smaller than those of a ground station. Achieving the theoretical improvement brought by a multi-hop link demands that specific values of the system parameters, most notably the antenna size and its pointing accuracy, can be met. The sustainability of such demands in terms of engineering resources to be allocated onboard the space terminals is, however, debatable. In Sect. 5.5 we present our analysis on the subject, where we assume the ground segment part of the link to be kept the same between a direct link or a multihop one. Therefore, only resources allocation onboard the space terminals will be considered, assuming a link whether at EHF bands or optical frequencies.

5.2 Ground Infrastructure for Interplanetary Communication Network 5.2.1 Deep Space Antenna Architecture A Deep Space Antenna (DSA) is the essential infrastructure to create effective wireless telecommunication links between ground users and the spacecrafts traversing the Solar System. The role of the ground segment in any interplanetary mission is that of supporting the Tracking, Telemetry, and Command (TT&C), providing a gateway for the uplink of commands and the downlink of information on the status of the spacecraft and payload data [1]. In consideration of the importance of their scope to the success of deep space missions, a DSA must be designed and operated following strict standards and specifications. The design equations of DSAs are the Friis Transmission Formula (Eq. (5.1)) and the noise power equation (Eq. (5.2)). PR = PT GT GR (λ/4R)2 ,

(5.1)

N = kBTs ,

(5.2)

.

.

where .PT is the transmitted power, .GT and .GR the transmitter and receiver gains respectively, .λ the wavelength of the transmitted signal, R the distance between the antennas, k the Boltzmann’s constant, B the receiving bandwidth, and .Ts the system temperature at the receiver. The two equations come into play when computing the SNR, which can be written as .PR /N, and is the driving parameter of

5 Ground and Space Hardware for Interplanetary Communication Networks

109

Fig. 5.1 Main types of parabolic antenna feeds

the telecommunication link, as it encompasses all the information on the quality of the transmission. Specifically, by knowing the SNR of the transmission, the bit error rate of telemetry and command activities can be directly computed [2], and the achievable accuracy of tracking estimated [3–5]. In consideration of the interplanetary distances between the ground segment and spacecraft, the signals arriving at the receiver on the ground are extremely weak. To compensate for this effect, DSAs architecture relies on high directional gains .GR , proportional to the antenna aperture and capable of amplifying the downlink signals to a level above the receiver sensitivity. With these considerations in mind, it is possible to perform a high-level description of the most common architectures of DSAs. As high gains and directionality are required in the transmission link, virtually all DSAs employ a parabolic dish. Common main reflector sizes range from 34 m to 70 m diameters [1]. After the downlink signal reaches the main reflector, it is reflected on to the frontend of the receiver (called feed) directly or after passing through waveguides. A non-exhaustive schematic representation of common parabolic antenna feed configurations is shown in Fig. 5.1. Among all possible feed methods, the front feed configuration is the simplest, as it consists in placing the feed directly at the focus of the parabolic reflector. This architecture is common when complexity is to be limited, but the feed can be susceptible to adverse weather conditions, requiring more frequent maintenance. Additionally, as the feed and its supporting mechanical structure will block part of the signal from reaching or departing from the parabolic reflector, the aperture efficiency of the antenna is limited to values of 55%–60%, and the side lobe and polarization levels increase [6]. Dual-reflector antennas employ a second reflector to guide the received signal, generally through waveguides, and make it possible to position the feed in a dislocated position, therefore protecting it from adverse weather and increasing the antenna aperture efficiency. Dual-reflector antennas usually employ the Gregorian or Cassegrain configurations. The two architectures employ an elliptical or hyperbolic subreflector, respectively, and increase the antenna aperture efficiency by more than 70% [6]. Additionally, using an appropriate beam waveguide (BWG) subsystem, it is possible to quickly switch between different receivers during operations [7]. Figure 5.2 shows the general architecture of a 34-m Cassegrain BWG DSA of the DSN [1].

110

P. Tortora et al.

Fig. 5.2 General architecture of a 34-m Cassegrain BWG DSA of the Deep Space Network (DSN), courtesy of [1]

While the main concern of guiding the radio-frequency (RF) signal through the DSA is keeping a correct collimation, high illumination and spillover efficiencies, the mechanical design of the antennas is concerned with providing accurate pointing and maintaining deformations in the structure to a minimum. This concern arises from the fact that signals at high frequency bands are prone to gain losses due to errors in the reflecting surfaces. Such errors are canonically divided into timeinvariant errors due to mechanical manufacturing of the surfaces of the reflectors, and time-varying effects due to gravity deformations [1]. The first error source is usually well characterized and compensated thanks to holographic measurement, or by using electromechanical actuators to reduce inaccuracies. The strategy of

5 Ground and Space Hardware for Interplanetary Communication Networks

111

employing active surfaces, additionally, can also reduce the effects of gravity deformations and axes misalignments during tracking operations, and therefore compensate for time-varying deformations as well [7]. The pointing of the DSAs, from a practical point of view, is generally performed using an azimuth-elevation mechanical configuration, and in consideration of the inertia of the elements involved and required precision, multiple high-power gearmotors are connected to the structure to avoid issues associated to gear backlash and reduce the effect of mechanical elasticities during the transmission of motion.

5.2.2 Uplink and Downlink Chains After the signal is reflected into the feed, also called front-end, the DSA must perform appropriate signal processing on the analog (RF) signal before its digitalization. Every DSA has its own signal acquisition chain. The acquisition chain architecture for the 34-m BWG stations of the DSN can be found in [8], while the one for the 70-m antennas is described in [9]. Details about the signal processing happening in the DSN affiliated antennas DSS-17 and DSS-69 operating in deep space tracking are presented in [10] and [7] respectively. Generally speaking, essential components found in the acquisition chain are low-noise amplifiers, filters, isolators, and mixers. More in detail, the front-end is responsible of amplifying the RF signal to detectable levels after digitalization, and mixing it to an intermediate frequency (IF) in the order of 10–100 MHz, and transmitting it to the backend, usually through optical fiber to maintain frequency stability [11]. Figure 5.3 provides a high-level description of the components in the acquisition chain in a DSN station. During downlink, analog amplification is performed on the signal before mixing it to the IF and sending it to the receivers. Depending on the role of the antenna in the mission, both the open-loop and closed-loop receivers can be active. The first one is usually employed when radio science investigations are carried out, since unexpected power and frequency fluctuations can be overcome by postprocessing after the signal is recorded, differently from what happens when using a closed loop receiver [13]. The bitrate in the downlink acquisition chain depends greatly on the mission and the type of payload on board the spacecraft. While scientific payloads can increase the required downlink bitrate up to hundreds of Mbps, telemetry data rates usually range from kpbs to Mbps [14]. The uplink chain is used to send commands to spacecrafts and provide accurate frequency references when working in a coherent uplink-downlink configuration. This transmission relies on the frequency and timing subsystem of the station to guarantee appropriate stability in frequency to the transmission. Differently from the downlink, the uplink’s required data rate is limited to kbps or even bps [14].

112

P. Tortora et al.

Fig. 5.3 Generalized diagram of a Deep Space Network antenna’s uplink and downlink path, courtesy of [12]

5.2.3 DSN Capabilities The Deep Space Network is an international network of communication facilities that supports NASA’s interplanetary spacecraft missions and radio and radar astronomy observations. Managed by Jet Propulsion Laboratory, DSN consists of three deep space communication facility complexes placed approximately 120.◦ apart from each other around the world. What follows is the list of complexes and their location: • Goldstone Deep Space Communication Complex outside of Barstow, California; • Madrid Deep Space Communication Complex, 60 km west of Madrid; • Canberra Deep Space Communication Complex, 40 km southwest of Canberra. Each complex includes several deep space stations equipped with ultrasensitive receiving systems and large parabolic dish antennas. Stations are classified according to their architecture. The 34-m antennas are canonically labelled as BWG if the signal is redirected toward the base of the structure, and high efficiency (HEF) if they employ a front feed configuration. Table 5.1 summarizes each complex antenna composition, uplink and downlink capabilities and parabolic reflector size. Current plans are to have on each site four 35-m plus one 70-m antennas in the near future, while 70-m antennas will be gradually decommissioned [17]. 34-m and 70-m parabolic reflectors and powerful 20 kW HPA allow the achievement of Effective Isotropic Radiated Power (EIRP) values up to 115 dBW at X-band transmission in normal situations. Additionally, during unexpected or emergency situations, a transmitter power of up to 400 kW at S-band can be sent from the 70-m

5 Ground and Space Hardware for Interplanetary Communication Networks

113

Table 5.1 Characteristics of DSN deep space stations [15, 16] Complex Goldstone

Canberra

Madrid

∗

.

DSS ID 15 23.∗ 24,26 25 14 45 33.∗ 34,36 35 43 65 53 54,56 55 63

Size (m) 34 34 34 34 70 34 34 34 34 70 34 34 34 34 70

Type HEF BWG BWG BWG D.S. HEF BWG BWG BWG D.S. HEF BWG BWG BWG D.S.

Uplink X X S,X,K X,Ka S,X S,X X S,X,K.∗ X,Ka.∗ S,X S,X X S,X,K X,Ka .∗ S,X

Downlink S,X X,Ka S,X,K,Ka X,Ka L,S,X S,X X,Ka S,X,K,Ka X,Ka L,S,X S,X X,Ka S,X,K,Ka X,Ka L,S,X

EIRP (dBm) 109–140 134–145 109–145 134–145 132–146 109–140 134–145 109–145 134–145 132–146 109–140 134–145 109–145 134–145 132–146

G/T 39.1–56.2 51.3–62.5 40.6–62.5 51.3–62.5 48.3–61.7 39.1–56.2 51.3–62.5 40.6–62.5 51.3–62.5 48.3–61.7 39.1–56.2 51.3–62.5 40.6–62.5 51.3–62.5 48.3–61.7

Represents features that are planned in the near-term future

antenna. Many ground stations allow simultaneous transmission at S- and X-bands. At present, only DSS-25 at Goldstone Complex includes a fully operational Ka-band transmitter, originally developed for Cassini Radio Science experiments [18].

5.2.4 ESTRACK Capabilities ESTRACK represents the ESA’s capability to provide links between in orbit spacecrafts and the Operations Control Centre at ESOC. ESTRACK core network includes seven stations in different countries: Kourou (French Guiana), Cebreros (Spain), Redu (Belgium), Santa Maria (Portugal), Kiruna (Sweden), Malargue (Argentina), and New Norcia (Australia). The core network is complemented by commercially operated stations. ESTRACK is designed to provide global space link connectivity coverage for a wide range of space missions. At present, the only ESA ground stations that could support deep space missions are New Norcia (DSA-1), Cebreros (DSA-2), and Malargue (DSA-3), that compose the ESA Deep Space Network (EDSN). These stations are equipped with a 35-m parabolic reflector antenna which includes a full motion turning head pedestal with a feed system, a cryogenically cooled Low Noise Amplifier (LNA), HPA and auxiliary subsystems such as ranging calibration, frequency reference and power distribution. The frequency reference generation is based on a Hydrogen Maser with very high long term frequency stability. Radio frequency parameters for DSAs are shown in Table 5.2. Specifically,

114

P. Tortora et al.

Table 5.2 Characteristics of ESTRACK deep space antennas Complex New Norcia Cebreros Malargue

DSA ID 1 2 3

Size (m) 35 21 35

Type BWG BWG BWG

Uplink S, X X X, Ka

Downlink S, X X, Ka X, Ka

EIRP (dBm) 97–107 107 100–107

G/T (dB/K) 37–57 50–57 50–57

DSA-1 supports both S- and X-bands for uplink and downlink transmissions. DSA2 supports transmission only at X-band, and reception at X- and Ka-band. DSA-3 station is able to operate at X- and Ka-band for uplink and downlink. DSA-3 can also exploit the use of a new receiving band, namely the K-band, ranging from 25.5 to 27 GHz. The K-band strongly enhances the communication bit rate, thanks to the huge available bandwidth, which is three times larger than that used at Ka-band.

5.3 On-Board Equipment for Interplanetary Communication Networks 5.3.1 S/C TT&C Transponders Deep Space transponders are the heart of any interplanetary S/C telecom system. In the last two decades, the progress in digital implementation of complex architectures, semiconductor integration trends, and the availability of reliable device technologies led to a deep evolution of the onboard telemetry, tracking, and command (TT&C) equipment. Nowadays, digital technologies allow us to design and implement different modulation schemes, ranging functions and TM/TC formats leaving most of the hardware configuration unchanged [19]. The current trend is to abandon completely analog system, making use of the discrete nature of signal processing techniques which determine a conceptual breakthrough and make transponder architectures (see Fig. 5.4) very similar to any wireless transceiver, with the addition of navigation and radio-science functions. With this architecture in mind, several H/W solutions have been designed (and are currently in flight—or will be flown shortly—on ESA and NASA solar system exploration missions—see e.g. [20–23], which make use of 4 main blocks that can be adapted in a highly modular way: X-/Ka-band receiving section, X-/Kaband transmitting section, Digital section, and a common DC/DC Converter module (Fig. 5.5). The receiver section uses a low noise amplification stage and a single downconversion stage for translating the X-/Ka-band up-link carrier to a suitable Intermediate Frequency (IF). Down-conversion is obtained using a sub-harmonic mixer. The transmitter section is based on the direct synthesis of the X-/Ka-band downlink carrier. A wide band vector modulator can be used as an integrated solution

5 Ground and Space Hardware for Interplanetary Communication Networks

115

Fig. 5.4 Architecture of a digital transponder

Fig. 5.5 Modular architecture of a digital transponder

to perform the carrier modulation down-link signals. The baseband modulation (including TT&C coherency) is managed in the digital module by an FPGA that provides in-phase and quadrature modulating signals carrying TT&C modulating data. The DC/DC module is used to supply both the analog and the digital section, and copes with both the RX and TX section and can power both the X- and Kaband modules. Thus, the DC/DC converter shall be always ON and the TX ON/OFF function shall act directly on the TX secondary lines (note that the power supply post regulation is implemented inside each section). This architecture eases the use of higher (Ka-band) frequencies or the combined used multifrequency links at X- and Ka-band, and may offer several advantages. First, Ka-band yields a fourfold performance advantage over X-band due to the increased directivity of the RF beam, and this allows to increase the data return to the Earth by a factor of four or (alternatively) to reduce the S/C antenna size or

116

P. Tortora et al.

RF power. In addition, Ka-band allows reducing the plasma effects on the radio link with better performance of the TM/TC transmission because of an increased signal-to-noise ratio in the carrier tracking loop, which also yields a substantial improvement of the Doppler shift measurement accuracy [24]. From a technological point of view, the miniaturization and modularity of digital transponders makes it possible to bring mass and power reduction to the limit. These two factors turn out to be two crucial winning elements for the design of on-board equipment for deep space telecommunications.

5.3.2 S/C Solid State vs Travelling Wave Tube Amplifiers Both Solid State Power Amplifiers (SSPA) and Travelling Wave Tube Amplifiers (TWTA) have been employed in deep-space spacecrafts since the early days of space exploration. If TWTAs were the go-to solution for the very first deep-space probes (e.g. Pioneer 10 and Pioneer 11), the first application of an SSPA in a deep-space mission can be dated back to 1977 with the launch of the Voyager 1 and Voyager 2 probes, which employed two TWTA at X-band and one TWTA and one SSPA at S-band [25]. Some of the major figure of merits that are of interest for amplifiers in satellite applications, are the total mass of the system, the overall output RF power and the Power-Added efficiency, which is computed as: P AE = (Pout − Pin )/PDC ,

.

(5.3)

with .Pout being the output RF power, .Pin the input RF power and PDC the supplied power. TWTA consists of a Travelling Wave Tube coupled with a high-voltage power supply. The RF signal amplification is obtained by guiding the electromagnetic wave in close proximity to an electron beam. The beam travels at a velocity close to the phase velocity of the electromagnetic field. The interaction between the electromagnetic field and the beam results in an exchange of kinetic energy from the beam towards the electromagnetic wave, resulting in a signal amplification [26]. A schematic of the architecture for an helix TWTA is reported in Fig. 5.6. Among the possible designs of TWTA, the helix TWT is the most commonly employed in deep-space missions [27, 28] due to high PAE (in the order of 40– 50%) and the larger achievable bandwidth as compared to other designs such as the coupled-cavity TWT, which allows for higher output power levels at the expense of more narrow bandwidths. Among some examples of TWTA usage in deep-space missions it is possible to mention the 10 W Ka-Band amplifier used in Cassini mission for radio science experiments [29] in combination with 20 W Xband TWTA for data transmission toward earth [30] or the two X-band and one Ka-band amplifier for Mars Reconnaissance Orbiter [31].

5 Ground and Space Hardware for Interplanetary Communication Networks

117

Fig. 5.6 Architecture of an helix TWTA, courtesy of [27]

From a reliability stand-point, a review of in-flight data of commercial satellites conducted by Boeing Satellites Systems in early 2000s [32] on more than 100 million operating hours of their fleet equipped either with TWTA or SSPA has shown a higher failure rate in SSPA. Nevertheless, especially in recent years, with the advent of smaller sized deep-space spacecraft to be employed as secondary payloads in mission targeting outer space, the lower power consumption, the lower mass and the more compactness of SSP amplifiers, has renewed the attention towards these devices. SSPA are usually based on a combination of Monolithic Microwave Integrated Circuit (MIMC), High Electron Mobility Transistors (HEMT) and MEtalSemiconductor Field-Effect Transistors (MESFET). Both the achievable efficiency and maximum output power for an SSPA is related to the compound semiconductor technology and the overall architecture of the system [33]. Typically, deep-space applications have seen employment of Gallium Arsenide compounds (GaAs) which can achieve a PAE of 26%–28%. Examples of GaAs based SSPA are the one installed within the X-band Iris transponder employed on MarCO mission first and more recently on some of the CubeSats launched as secondary payloads for the SLS Exploration Mission 1 [34], or the SSPA installed both on Mars Exploration Rover and Mars Science Laboratory [35]. In recent years Gallium Nitride (GaN) based X-band SSPA with PAE up to 35% have been successfully employed within the PROCYON mission, a 50 kg class SmallCraft launched as secondary payload of Hayabusa 2 [36]. Depending on the desired RF output power, SSPA architectures can include one or more amplification stages, either in series (as per PROCYON’s PA) or in parallel (as per the final amplification stage of MER/MSL SSPA) by employing a cascade of Power Dividers and Power Combiners. An initial variable attenuation stage, aimed at compensating fluctuations due to temperature variation can also be included as in MER/MSL SSPA. An overview of most notable power amplifiers employed in deep-space missions is reported in Table 5.3,

118

P. Tortora et al.

Table 5.3 Main characteristics of power amplifiers employed in deep-space missions Mission Cassini Cassini (RSE) MRO JUNO MER/MSL MarCO/EM-1 PROCYON

Type TWTA TWTA TWTA TWTA SSPA (GaAs) SSPA (GaAs) SSPA (GaN)

Frequency band X-band Ka-band X-band X-band X-band X-band X-band

PAE 37% 40% 59% 44% 28% 23% 33%

Power output 20 W 10 W 102 W 25 W 17 W 4W 14 W

Reference [30] [29] [37] [20] [35] [37] [36]

5.3.3 High- Medium and Low Gain Antennas Antenna systems are critical components in enabling communications between deep-space satellites and Ground Stations on Earth. Three of the most commonly employed figure of merits in describing an antenna performances are: • The radiation pattern in the far field, which illustrates how the antenna irradiates/receives electromagnetic waves in the surrounding radiation sphere as a function of the polar .θ and azimuthal .φ angles; • The maximum antenna gain, which is the ratio between the maximum radiation intensity and the average radiation intensity over the radiation sphere; • The half-power beamwidth, which is the angle between the two directions in the polar and azimuthal plane at which the antenna gain is half of its maximum value; Given the above definitions, it is clear that an antenna characterized by higher maximum gain implies a smaller half-power beamwidth, and therefore the selected design strictly depends on the operative environment the antenna is targeted for [38]. Due to these reasons, usually deep-space probes are equipped with multiple antenna systems providing different levels of gains and radiation patterns, depending on the requirements of specific mission phases. An example of this approach is NASA’s JUNO telecom subsystem [39], which consists of 5 different antennas as follows: • two identical choked horn Low Gain Antennas (LGA) (peak gain >7.4 dBic in X-band) mounted on opposite sides of the platform in order to ensure contact with the DSN X-band antennas in early phases of the mission, even in case of not-nominal orientation of the spacecraft; • One toroidal LGA with a peak gain >4 dBic in X-band and a > 2 dBic gain at ◦ .+/−10. from peak; • one conical horn Medium Gain Antenna (MGA) with peak gain >18 dBic in Xband to be employed both in near-earth operation and in case of emergency in far-out ranges thanks to a 13.5 dBic gain over a 24.◦ cone; • An high gain dual-band (X-Ka band) dual reflector antenna, which achieves a gain >41 dBic in a cone of .+/−0.25.◦ from peak;

5 Ground and Space Hardware for Interplanetary Communication Networks

119

Another interesting example is represented by NASA’s MRO S/C, which, in addition to an HGA and two LGAs operating in X-band, is also provided with a quadri-helix UHF LGA providing a broad coverage of the Mars surface for data-relay link with rovers and landers operating on Mars [40, 41]. The concept of including small sized spacecrafts, either CubeSat or SmallSats, as secondary payload in deep space missions with technology demonstration purposes or secondary scientific objectives, has led to the applications in deep-space environment of novel technologies. A recent example as the deployable reflectarray HGA with deployable patch strip feed installed on Mars Cube One mission [42], which is aimed at serving as relay node towards Earth of probes operating on Mars surface, or in other cases, as in LICIACube spacecraft [43] to the usage of technologies that finds large applications in near earth satellites but have been employed more scarcely in deep-space missions, i.e. printed microstrip antennas (also referred as patch antennas), which despite lower gains provided, comes at lower manufacturing cost and lower weight as compared to other LGA layouts.

5.4 Direct Earth-to-Deep Space RF links vs Multi-Hop Links 5.4.1 Typical Performances and Limitations of a Direct Link Direct link between outer of inner planets of the Solar system end Earth has leveraged the use of very large ground antennas, as the ones from the NASA DSN or ESA Estrack. To date, X- and Ka-bands are the most widely employed frequencies in deep-space. Apart from the impact of atmospheric and ionospheric impairments preventing the growth of transmission frequency, a direct link to Earth also suffers from visibility problems due to the presence of the Sun as an occulting body. The evaluation of the adversarial effects that solar radiation has on RF signals has been widely addressed in the literature, see e.g. [44], resulting in an increase in system noise as a function of the intensity of the solar radiation at the operative frequencies and of the geometry of the occultation. This last is dictatated by the relative geometry between the Sun, the Earth, and the user probe at the target planet, described by the so-called SEP (Sun-Earth-Probe) angle, which eventually affects the link availability. A survey of both past deep space missions and current mission concepts performed in [45] allowed setting some reference scenarios in terms of achievable data-rate for standard single-leg links, which can then be used as a benchmark to evaluate the benefits of a multi-hop configuration. Such an exercise was performed in the framework of a was supported by European Space Research and Technology Centre (ESA/ESTEC) funded study, and the main outcome is displayed in Table 5.4. These values may be used as a benchmark to evaluate the feasibility and benefits of using a multi-hop link.

120 Table 5.4 Data rates achieved by a system based on a direct link (X-band: 95% availability, 10.◦ elevation; Ka-band: 90% availability, 20.◦ elevation)

P. Tortora et al. Scenario Venus (X-band) Mars (X-band) Uranus (X-band) Neptune (X-band) Uranus (Ka-band) Neptune (Ka-band)

Data rate 100 kbps 120 kbps 3.15 kbps 1.20 kbps 10 kbps 4 kbps

5.4.2 Data Relay Architectures The progress in space technologies and their continuous evolution, especially in the deep space context, requires the development of new ideas and solutions to support the endeavors of researchers and developers. A “space intelligence” is more than ever required in order to continue our exploration and understanding process, as well as to guarantee Human Beings’ constantly evolving needs. Among the others, one of the most important “glue” in space ICT services is the link, communication, and data exchange between different entities, which shall work in a seamless networked manner [46]. Thanks to data relays, space missions are no longer required to have direct line-of-sight to Earth ground based antennas, creating a continuous path for data flowing from missions in space to ground stations on Earth, and generating a complete end-to-end system. In this context, updating and developing data relay architectures is the first step towards an Earth to Space universal network, with the primary goal of satisfying the requirements in a fully sustainable manner. This section aims at investigating data relay architectures and to assess their potential advantages with respect to a classical space-to-Earth direct link (X-band in the range of 8–12 GHz or Ka-band 27–40 GHz. Furthermore, the goal is also to define a technology roadmap towards exploitation of these architectures in future deep space missions and technologies, based on the space domain sustainability, awareness and responsibility. To illustrate the benefit of data relaying, we consider a scenario involving a deep space link split in two legs [47]. The two links would be characterized by a spacecraft orbiting the Earth, or another central body (link 1), or in orbit at a Lagrange point, which would receive data from a deep space spacecraft (link 2) and relay them to the ground station on Earth or on another planet like Mars, see Fig. 5.7. The deep space to relay link will not be affected by the Earth atmosphere, hence it may take advantage of an extremely high frequency band, e.g., frequencies between Ka-band ones and 75 GHz (which include the Q/V band), or optical band, being then able of extremely higher data rates if compared to classical space-to-Earth direct links. Indeed, even if radio waves have been widely used in space communications since the beginning of space exploration with a proven track record of success, as space missions and exploration grow generating and collecting more data, the need for enhanced communications capabilities becomes paramount. On the other hand, the primary link may use a more classical RF band, such as the K-band (for a near-

5 Ground and Space Hardware for Interplanetary Communication Networks

121

Fig. 5.7 One-leg configuration (direct link from deep space to ground) versus two-leg configuration featuring a near-Earth relay, adapted from [47]

Earth relay) or the X/Ka band (for a relay in a deep space L4/L5 Lagrange point), benefiting from the shorter distance to the ground. The interest and the need to develop innovative data relay architectures is also reflected in one of the latest astronautical frontiers of the last two decades, namely the exploitation of small satellite platforms to explore near and deep space with reduced mission development time and costs. These small deep space platforms are characterized by reduced technical resources, hence the adoption of a data relay configuration, involving two cost-effective spacecraft, will potentially bring concrete advantages to the link design and the RF miniaturized technology roadmap definition. For example, reduced transmit power and/or reduced on-board/onground antenna size while maintaining the same data rate and performance, thus contributing to the fostering of the small satellites revolution in deep space, too.

5.4.2.1

Data Relay Architectures for Earth Orbiting Users

The very firsts data relaying systems for space applications, and in particular for users orbiting the Earth, date back in 1963 thanks to the envision of NASA. In order to overcome the issues related to the limited coverage of low-altitude Earth orbiting spacecraft by ground stations, they conceived the idea of Tracking and Data Relay Satellites (TDRS) [48]. Two decades later, in 1983, the first TDRS service ever became operational, with the aim of providing near continuous communications and tracking services to LEO spacecrafts, launch vehicles, and suborbital platforms in general. Data relay services to satellites orbiting in Earth proximity have been developed, both from national agencies and commercial companies (e.g., [49, 50], leveraging on GEO primarily, as well as on MEO satellites. The LEO spacecrafts represented the main use case for relay services since, from a purely coverage

122

P. Tortora et al.

LEO

GEO

Fig. 5.8 Concept of global coverage for a LEO user provided by a GEO tracking and data relay satellite service (figure not in scale)

point of view, the users orbiting below GEO altitude are the ones getting the maximum benefit. Indeed, thanks to GEO satellites, which are positioned at three distinct longitudes, global coverage for LEO users is always guaranteed, as shown in Fig. 5.8. Currently, there are three existing and successful relay satellite systems: • the aforementioned Tracking and Data Relay Satellite System (TDRSS—also known as the Space Network or Space Now), managed by the NASA GSFC’s space communication and navigation programme. This belongs to the NASA space communication networks, together with the Deep Space Network and the Near-Earth Network; • the European Data Relay Satellite System (EDRSS) developed within the ARTES programme [51]; • the Japanese Optical Data Relay System (JDRS) with performances similar to the one of EDRSS [52], a project adopting a “Laser Utilizing CommunicAtion System” (LUCAS) which is composed by a laser-communication terminal onboard both the Optical Data Relay Satellite and on the Advanced Land Observing Satellite “DAICHI-3” (ALOS-3), and a optical ground system; These programmes have been developed to answer the ever-growing need of a space communications infrastructure capable of supporting large volumes of data to be downloaded per day from space to ground. NASA’s TDRSS advocates the use

5 Ground and Space Hardware for Interplanetary Communication Networks

123

of optical links in addition to microwave communication, to enable a significant increase of achievable data rate, while the European and Japanese ones already implemented the optical technology.

5.4.2.2

Next Generation Data Relay Architectures for Earth Orbiting Users

NASA-TDRSS The next generation TDRSS will enhance its current communications capabilities by exploiting optical links [53]. NASA’s plan is to have the first optical relay satellite operating by 2025, and the optical relay communications will comprise both intersatellite and satellite-to ground links. As a first step towards NASA optical relay system commissioning, the Laser Communications Relay Demonstration (LCRD) mission was launched on 7 December 2021, built upon the success of the earlier Lunar Laser Communication Demonstration (LLCD) led by MIT, which successfully demonstrated 622 Mbps communications from lunar orbit to ground [54]. LCRD is expected to perform 2 years testing of laser communications capabilities to fully characterize these technologies allowing us to practice and evaluate potential disturbances, as well as to understand their potential future applications. During the initial experiment phase the system makes use of two simulated users, namely the Optical Ground Station 1 and 2, based in California and Hawaii. Then, after the experiment phase, LCRD is expected to support space missions, sending and receiving data to and from satellites over infrared lasers in order to demonstrate the benefits of a laser communications relay system. In particular, the first mission which will exploit LCRD will be ILLUMA-T (to be launched in 2023), which will gather data from the space station and send to LCRD at 1.2 Gbit/s. The key features of NASA LCRD, which may be considered as representative of the next generation TDRSS, are collected in Table 5.5 [53]. ESA-EDRSS The next generation enhanced EDRSS will implement the GlobeNet, a global laser-based network in space. GlobeNet will extend the current EDRSS by adding the EDRS-D node, as well as including additional security related upgrades and modifications to the EDRSS ground system, enabling new classes of security sensitive users such as UAVs and RPAS [55]. Thanks to the new EDRS-D node, Table 5.5 NASA LCRD technical specification

Subsystem Optical modules Laser wavelength Laser transmit power Modulation RF downlink Data rate

Description Two 10.8-cm telescopes, 2-axis 1550 nm 0.5 W OOK, DPSK, PPM, DPQPSK Host spacecraft provided Up to 1.244 Gbps

124 Table 5.6 NASA LCRD technical specification

P. Tortora et al. Subsystem Inter-satellite link Ground-satellite link Data rate Laser wavelength Laser transmit power Number of laser terminals Modulation

Description Optical or Ka-band Optical or Ka-band 3.6–10 Gbps (Optical) 3.6 Gbps (Ka-band) 1064 nm 1550 nm Up to 2.2 W 3 N/A

located between 120◦ and 155◦ East, data gathered over, e.g., the Pacific Ocean will be transmitted optically from the user platform to EDRS-D, re-routed via optical link to EDRS-A or EDRS-C, and then downloaded by a European ground station. Compared with the current EDRSS capabilities, the main technical improvements that GlobeNet will exhibit may be summarized as a dual wavelength optical communication capability, an increased data rate, and the capability to serve up to three users at the same time owing to a set of three laser communication terminals (see Table 5.6). Given that, to date, GEO spacecrafts represent the sole practical implementation of relay platforms, [50] proposed a comparative assessment between GEO, MEO (1/2 and 1/4 synced) and LEO relay architectures for the next generation TDRSS, in terms of both technical and cost metrics. Targeted users encompassed from suborbital to the edge of Near-Earth space (2 · 106 km). The main outcome of [50] is that the GEO option offers the most significant benefits for several reasons, among which: • • • •

fewer number of relays and ground terminals needed; no inter-relay crosslink required; lower operational complexity; the highest heritage.

ESA-HydRON Despite the very successful introduction of the EDRS, the capabilities of optical technologies are not yet fully exploited, due to the fact that the optical payloads are mostly used in non-optimized SatCom systems. To this end, ESA in 2016 opened a dedicated programme for Optical Communication Technologies called ScyLight (SeCure and Laser communication technology). Furthermore, in order to address and develop optical technologies for SatCom systems in all system level aspects, ESA recently launched a new innovative project (including many internal and external projects and initiatives) called HydRON (High Throughput Optical Network) [56].

5 Ground and Space Hardware for Interplanetary Communication Networks

125

HydRON aims to [56]: 1. investigate end-to-end system architectures needed to get Terabit/second throughput systems, in order to get a “fibre in the sky”, a natural extension of the terrestrial fibre networks in space; 2. identify the key aspects and elements of the system architecture as well as provide possible solutions and overviews of potential enabling technologies required; 3. define an in-orbit demonstration mission to fully characterize the key elements in terms of performances, but also cost analysis and technology development schedules. Commercial Many commercial entities are now involved in this increasingly crowded field, with the goal to address a growing demand from satellite operators for high-speed data transfer. Both Government agencies and commercial companies are looking for solutions and new architectures to exchange data more quickly and securely. Optical communications offer higher data transfer while, at the same time, laser beams require precise pointing, making signals difficult to access for anyone other than the intended recipient. Between 2024 and 2026 many more GEO and MEO data relays services are expected to be operational.

5.4.2.3

Data Relay Architectures for Deep Space Users

Deep space or planetary exploration missions have received less attention so far as potential users of relay constellations. However the continuous growth of deep space exploration, together with the needs of space missions to generate and collect more and more data, are pushing forward the quest for enhanced communications architectures in deep space, too. The availability of wide bandwidth relay satellites for deep space missions would be a critical asset for enhancing a mission scientific return by achieving higher data rates. For example, an image taken by the HiRISE camera onboard the Mars Reconnaissance Orbiter (up to 28Gb/image) takes about 1.5 h to be transmitted back to Earth at the maximum data rate of 6 Mbps [57]. Most of the available studies on data relaying for deep space considered an optical link between the user and the relay satellite: optical communications may indeed provide deep space probes with higher data rates. The main drawback of a direct space-to-ground optical link is the higher vulnerability to weather and atmospheric impairments than RF communication. One way to overcome this issue relies on optical relay satellites, potentially offering 24-h coverage without being affected by weather [58]. In this context, data relay architectures could become of paramount importance for deep space users too [47]. The first assessments of the feasibility of communication relay satellites for deep-space users are from NASA. After Hunter’s 1978 seminal work, envisaging an Earth orbiting satellite with RF receive-only capability [59], a Deep Space Relay Satellite System (DSRSS) was investigated 15 years later. In 1998 the Jet Propulsion Laboratory’s Advanced Project Design Team studied both direct and

126

P. Tortora et al.

coherent-detection configurations for an Earth-Orbiting Optical Communication Relay Transceiver (EOORT). More recently, a cost and performance comparison between an Earth-orbiting optical communication relay versus a network of groundbased optical receivers has been presented in [58]. Most of the above-mentioned investigations, as well as [60], suggested that, even though from a data link performance point of view a relay satellite is to be preferred, when considering economical aspects (i.e., cost per deep space mission user), a network of ground stations would outperform a deep-space relay satellite. Nevertheless, recently [58] showed that the life cycle cost of a 7 m optical relay station is comparable to that of an 8-station network of 10 m optical ground stations. However, it is important to notice that the cost estimates from which these conclusions are drawn are some 20 years old, at least, and [60] recognized that this conclusion may change over time if access to space becomes less costly. Application of the EDRS system for establishing data relay with near-Earth and deep space was considered in 2009, in the so-called DROM (Data Relay for Moon) mission concept, with possible extensions to Mars orbiters [61]. The most recent studies involving relay satellites for deep space are primarily focused on optical links (e.g., [62–65]), where performance requirements, candidate configurations, and future trends are discussed. For example, the next generation NASA GEO optical relay satellites are aimed at offering data rates up to 2.88 Gbps to Lunar users and data rates of 100 Mbps or more to a deep space mission to Psyche asteroid [64]. To conclude, it is also important to mention another deep space data relay application, which is foreseen to become more and more popular in the next years of planetary exploration: the use of Inter-Satellite Link systems between mothercrafts and CubeSats [66]. These systems will allow the CubeSats to relay their data to the mothercraft (usually at S or X-band), which will then transmit all the information on the Earth via X or Ka-band links. Once the conceptual advantages offered to deep-space users are established, the actual improvement that data relaying can bring compared to a direct link shall be addressed from a link budget perspective, as discussed in the next paragraphs.

5.4.3 Multi-Hop Links Advantages and Architectures One of the key challenges of the coming decades is to develop deep space communication systems with orders of magnitude greater capacity than present ones. The conventional approach, still widely adopted, to long-haul communication in deep space is characterized by architectures made by single-hop links, where the two elements are a ground-based antenna on Earth and a smaller antenna onboard a spacecraft [67]. This is mainly due to the higher costs to place instruments and technologies in deep space rather than at a site on Earth, namely the cost per unit mass lifted from the Earth to deep space. As a result, the more cost-effective approach for single-hop communications appears to be the one having a reliable spacecrafts with low complexity, while putting more advanced, complex and heavy

5 Ground and Space Hardware for Interplanetary Communication Networks

127

Fig. 5.9 Single-hop telecommunication link architecture (adapted from [67])

Fig. 5.10 Multi-hop telecommunication link architecture (adapted from [67])

systems at the Earth’s antenna site [67]. A valid alternative would be that of adopting multiple-hop links [67, 68]. Even though additional space units would seem to increase the overall complexity, costs, noise and decrease the reliability, multi-hop systems can be much more cost-effective than the single-hop ones, depending on the selected architecture and technology. The basic schematic architectures for singlehop and multi-hop solutions are reported in Figs. 5.9 and 5.10, where .Pt and .Pr are the transmitted and received powers, respectively, while R is the distance. The basic idea would be to split the long-haul link of distance R using several terminals, obtaining in this way n hops covering the same distance. The theoretical advantage of multi-hop systems appears when looking at the ratio of power received to power transmitted. In a single-hop link in vacuum, the ratio is namely [67]: α = Pr Pt =

.

ηt ηr At Ar = Gt Gr λ2 R 2



λ 4π R

2 ,

(5.4)

where .λ is the wavelength of the transmitted signal, .ηt and .ηr are the antenna efficiencies for the transmitter and receiver, respectively, .At and .Ar are the antennas physical areas, .Gt and .Gr are the antennas gains. If one now assumes the simple scenario where the overall distance is split into n identical equidistant hops of length .R/n, the previous equation becomes [67]:

128

P. Tortora et al.

α  = Pr Pt =

.

ηt ηr At Ar λ2 (R/n)2

= n2 α .

(5.5)

Suggesting that the ratio of power received to power transmitted increases by a factor of .n2 for the multi-hop system. Furthermore, even assuming that the cost of a multi-hop architecture would scale linearly with the number of terminals, its efficiency-to-cost ratio will be n times better than a single hop, thus retaining a conceptual advantage. The optimization of the multi-hop architecture is not trivial. Several factors must be taken into account when designing such systems, and the solution depends on the particular application, so it must be evaluated on a case by case basis. Breidenthal [67] proposed several strategies, varying the number of hops, their orbits as well as their relative positions. Additional characteristics and pros about the multi-hop solutions include [67]: • noise introduced on the additional hops can be overcome adopting modem errorcontrol coding; • more robust and stable than single-hop with respect to failures. If a unit will fail, the multi-hop will degrade its performances, while the single-hop would be completely unusable; • If designed using “glue technologies” such as SDR systems, their reconfigurability and reprogrammability allow to minimize eventual losses of performances, but also to scale the entire system (after initial installation, too), by adding more hops. As multiple units start being connected together, there is also the need to think and develop their network, traffic architecture, protocols, as well as the algorithms which manage all the links within the trees (as we do for terrestrial mobile networks, for example). In this context, [69] presented a comprehensive study regarding Space-based multi-hop networking, providing network algorithms and solutions for efficiently scheduling the communications resources to satisfy network traffic with minimum latency and maximum throughput. They also have shown a test-case satellite constellation with multiple ground stations to illustrate the algorithm and related performances, demonstrating and confirming again the overall benefits in designing multi-hop architectures for space applications. Despite the apparent theoretical advantage of a multi-hop configuration, the practical application of such a concept is far more complicated than what appears by Eqs. (5.4) and (5.5). Indeed, the theoretical .n2 scaling is achieved only when the product of the receiving and transmitting gains are maintained constant across the hops. If no losses other than the free-space ones are considered, this amounts to requiring the product of the antenna areas to be kept the same. Since one cannot assume to place a greater than 34-m diameter antenna onboard a relay spacecraft, the constant areas product can be achieved only by significantly upsizing the deepspace terminal antenna. When additional losses are included at the receiver and at the transmitter, the scaling is preserved only if their product is maintained constant

5 Ground and Space Hardware for Interplanetary Communication Networks

129

as well. Most notably, this is hardly the case for the pointing losses. Indeed, while it is customary to neglect pointing losses at the ground station terminal of a direct link, in a space-to-space link the pointing losses shall be accounted for at both terminals. To this end, the analysis in [70] introduced the concept of effective system gain as the product of the transmit and receive antenna gains, each multiplied by the corresponding pointing loss. Looking at the effective system gain reveals the existence of an optimal value for the antennas’ size, beyond which the overall system performance degrades instead of improving, the optimum depending on the assumed level of antenna pointing accuracy. The concept of effective system gain was applied in [45] to a two-hop, deep-space communication scenario, to assess the potential benefits of using EHF at Q/V band or optical frequencies for the space-tospace leg. The study outcome suggests that, when assuming reasonable values for the antenna size and pointing losses at the deep space and data relay terminals, the fourfold increase in the system efficiency as predicted by the simplified analysis in [67] is not supported. Requiring only to match the data-rate achieved by a direct link already set quite stringent constraints in terms of antenna size and pointing accuracy. These are equal, respectively, to .≈ 5 m and .≈ 10−2 degree for a regenerative Q/V band link and to .≈ 1.4 m and .≈ 10−5 degree for an optical link. For all scenarios considered in the analysis, the bottleneck of the overall system performance turns out to be the space-to-space leg. As a result, the main advantage of splitting the link in two hops becomes that of using significant smaller ground antennas in the space-to-ground leg without impacting on the performance.

5.5 Guidelines for System Resources Allocation 5.5.1 Hardware Resources for a Traditional Single-Hop RF Link Allocating the mass, volume, and power to the different subsystems is one of the first actions in the preliminary design phase of a spacecraft. When no a-priori information is available, the task can be accomplished using sizing equations based on historical data [71]. To gain some insight in this respect, let us apply the method in [71] to allocate the communication subsystem mass as a percentage of the total dry mass depending on the spacecraft. For a planetary exploration spacecraft, this fraction is 6%–7%, and increases to 28% for a communication satellite. The former can be assumed as the reference class for the deep-space terminal, while the latter could be assumed as the reference for a data relay node, to be used in the next session. Mass estimates can also be split according to different assumed S/C size ranges. For a planetary exploration spacecraft, light, medium, and heavy classes are considered, which yields the following mass allocation to the communication subsystem:

130

P. Tortora et al.

• Light: S/C Dry Mass = 400–600 kg, Comm. SS Mass = 24–42 kg; • Medium: S/C Dry Mass = 1000–1200 kg, Comm. SS Mass = 60–84 kg; • Heavy: S/C Dry Mass = 2000–2200 kg, Comm. SS Mass = 120–154 kg. Brown [71] reports analogous historical data also for allocating DC power, showing that for a planetary exploration spacecraft, the percentage of the total subsystems DC power, typically required by the communication equipment is approximately 23%. The DC power available to the bus subsystems, .Pss , is given by Pss = Pt − Ppl

(5.6)

.

where .Pt is the spacecraft total DC power and .Ppl is the amount of DC power allocated to the payloads. A relationship between .Pt and .Ppl was also derived in [71] by fitting previous data for different mission categories. For a planetary exploration spacecraft the following equation is proposed:   Pt = 332.93 · log Ppl − 1046.6

.

[W ]

P lanetary S/C

(5.7)

Substituting the difference .Pt − Pss to .Ppl and assuming a 23% ratio of Pss to be allocated to the telecom subsystem, allows deriving the following power-estimating relationship: 

Pcom

.

Pt + 1046.6 = 0.23 · Pt − exp 332.93

 [W ]

P lanetary S/C

(5.8)

5.5.2 Hardware Resources for a Relay Satellite For preliminary sizing, the data-relay node of a multi-hop link can be assimilated to a communication satellite, for which the average communication subsystem mass can be assumed as the 28% of the total dry mass. Incidentally, a mass fraction of about 28% is also corroborated by a study on a light-TDRS concept [72]. However, on a relay satellite, the mass fraction devoted to the communication subsystem should serve, most likely, two different frequency bands, one for the deep-space link and the other for transmission to the ground. Indeed, the benefits of splitting a deep space link in two or more hops result from the exploitation of higher frequencies for the deep space hop than those used for the ground hop. For example, in [45], the benefits of using Extremely High Frequencies or optical links in deep space were assessed in a two-hop scenario. Thus, one may wish to conservatively adopt a lower fraction, say 20%, for the leg-1 data relay subsystem. We further assume as reference designs for the data relay spacecraft the operative TDRS-M spacecraft and the novel light-TDRS concept, resulting in the following mass allocation for the communication subsystem:

5 Ground and Space Hardware for Interplanetary Communication Networks

131

• TDRS-M: S/C Dry Mass = 1800 kg, Comm. SS Mass = 504 kg, Deep Space Leg Comm. Mass = 360 kg; • Light-TDRS: S/C Dry Mass = 745 kg, Comm. SS Mass = 208.6 kg, Deep Space Comm. Mass = 149 kg. As per the power to be allocated to the communication subsystem, the same approach used in the previous paragraph is used, starting from the total-vs-payload power for a Communication S/C: Pt = 1.157 · Ppl + 55.50

.

[W ]

Communication S/C ,

(5.9)

Which leads to the following communication subsystem power estimate: Pcom = 0.033 · Pt + 11.0

.

5.5.2.1

[W ]

Communication S/C ,

(5.10)

HW Resources for a Q/V EHF Band Link

To date, the number of missions that hosted a Q/V band communication payload on board is still small, and most of them were aimed at propagation experiments. The most relevant non-military applications are two technological demonstration GEO satellites, Italsat F1 (ASI) [73] and Alphasat (ESA) [74]. For what concerns the High-Power Amplifier (HPA) technology, the power conversion efficiency of the amplification technology available for Q/V bands of modern Solid-State Power Amplifiers (SSPA) [75] is between 17% and 30% while it reaches 45%–50% in Travelling Wave Tube Amplifiers (TWTA) [76]. These values are consistent with the performance of equipment working at more traditional frequencies (L-Band, S-Band, C-Band and X-Band), which are 25%–40% for a SSPA and 40%–70% for a TWT-based power amplifier, meaning that no significant differences are expected for the onboard mass/power resources allocation with respect to what needed by the hardware used for traditional X/K/Ka-band direct links. Recent studies [75, 77, 78] push the SSPA technology forward as a valid alternative to the TWTA satellite-based transmitters, which are considered too expensive, large and not enough reliable. Alphasat was indeed equipped with a Gallium Arsenide-based Monolithic Microwave Integrated Circuits SSPA, whose power output was between 5 and 10 W, with a required DC power of 64 W. The device had a mass of 3 kg and was 28 cm .× 18 cm .× 7 cm. Clearly, application of a similar technology for a deep space communication scenario require a significant increase in both the power output and power density. According to [77], using highly efficient power combining techniques, it is possible to achieve, with the same mass and size of that used in Alphasat, an amplification stage based on SSPAs with an output power of 80 W against 220 W of DC power, i.e. with a 36% efficiency. This output level would be in line with typical levels used in interplanetary communication.

132

5.5.2.2

P. Tortora et al.

HW Resources for an Optical Link

Following [79], the mass of an optical transceiver can be estimated as the sum of: • .Mopt , the mass of the optical head of the telescope, which shows a dependency on the telescope aperture diameter; • .Mlaser , the mass of the transmitter equipment, which depends on the laser average power. The empirical equations derived in [79] for .Mopt and .Mlaser are: Mopt = K · D 2.57

.

[kg] ,

Mlaser = 1.152 · Pavg + 3.168

.

(5.11) [kg] ,

(5.12)

where .K = 0.00181, D is the telescope diameter in centimeters, and .Pavg is the laser average transmitted power. Assuming a mass for the deep space optical communication segment equal to 6– 7% of the total dry mass for the deep-space terminal and 20% of the total dry mass for the data relay node, and a reasonable amount of 5 Watts as .Pavg , the equations above can be solved for the maximum allowable diameter of the optical assembly. Performing such exercise for the 3 size classes of the deep-space node leads to: • Light: D = 32–42 cm; • Medium: D = 54–63 cm; • Heavy: D = 73–81 cm. Likewise, for the data relay spacecraft, the equivalent diameters would be: • TDRS-M: D = 114 cm; • Light-TDRS: D = 80 cm. These values shall be compared with the diameters required to match, or exceed the data rate of a reference direct link. The analysis in [45] shows that this is obtained with an optical antenna size of about 1.4 m onboard both the terminals, when Pavg .= 5 W and provided that a sub-microradian pointing accuracy can be guaranteed. The diameter value is remarkably higher than those estimated with the preliminary sizing equations, especially for the deep-space probe. Compatibility of the DC power required by an optical communication link with the estimates from preliminary allocation can be assessed once the wall-plug efficiency if the system is known. Following [80], this last can be split into three contributions, namely (1) the electrical-to-optical conversion efficiency of the pump semiconductor laser diode; (2) the transfer efficiency of the pump light into the active medium; (3) the optical-to-optical conversion efficiency of the media with active gain. Hemmati et al. [80], also provides an estimate of the efficiency factors associated with a) highly mature design and b) optimized design for two different transmitter architectures, namely Laser Pumped Nd:YAG and Fiber Amplified Pump Yb: Glass.

5 Ground and Space Hardware for Interplanetary Communication Networks

133

Table 5.7 NASA LCRD technical specification Transmitter architecture Laser pumped Nd:YAG fiber Amplified pump Yb: glass

Demonstrated (pulsed) = 0.08 .ηtot = 0.1 .ηtot

Optimized design = 0.20 .ηtot = 0.42 .ηtot

Table 5.8 NASA LCRD technical specification Transmitter architecture Laser pumped Nd:YAG fiber Amplified pump Yb: glass

Demonstrated (pulsed) 62.5 W 50 W

Optimized design 17.86 W 11.90 W

Upon combination of the three efficiency factors, estimates summarized in Table 5.7 are obtained for the wall-plug efficiencies. In Table 5.8 are then summarized the estimated input DC power for each of the cases above, assuming a reference average laser power equal to 5 W. These values are to be compared with those from the empirical equations for .Pcom derived in Sects. 5.4.1 and 5.4.2. The DC power allocation calculated for the optical transmitter is compatible even with demonstrated design with the typical power levels reserved by traditional architectures, especially for large spacecraft (.Pt >≈ 500 W). Lighter spacecraft (.Pt 64 Hz) are adopted for the deep space hop, provided that specific values of the system parameters can be supported in terms of (1) transmitted power, (2) onboard antenna size and (3) pointing accuracy. Out of those, the two latter are the most demanding when compared to current technology status and represent the main setback with respect to a direct link for which, at the ground terminal, the antenna size and pointing losses are not a limiting factor. When analyzing the system resources required by a two-hop configuration it appears that, keeping the same mass allocation considered for traditional RF communication segments, the maximum diameter obtainable for the optical terminal onboard the deep-space spacecraft is smaller than that required by the optimal effective system gain in [45]. Besides, the optical terminals must be pointed with sub-microradian accuracy. Likewise, opting for a Q/V-band link would be beneficial if an antenna having a diameter equal to, or larger than, 5 m can be hosted onboard and can be pointed with .10−2.◦ accuracy. From a technology roadmap point of view, the following actions are thus recommended to enhance the current capabilities of deep space communication via a multi-hop network:

134

P. Tortora et al.

• to pursue novel optical antenna designs featuring increased gain/mass ratios; • to investigate innovative foldable antenna configurations to enhance the gain-tostowed size ratio for EHF bands; • to enhance the attitude control performance of the hosting platform, to limit the pointing losses.

5.6 Conclusions Existing hardware technologies and development trends for interplanetary communication have been reviewed, covering both the space and ground equipment. A survey has been provided of the two largest existing ground infrastructures for Interplanetary Communication, namely NASA’s Deep Space Network and ESA’s Estrack, focusing on the current capabilities for the uplink and downlink chains. Afterward, a review of on-board communication equipment was given, covering deep space transponders—in particular highlighting the advantages of digital vs analogue architectures, RF amplifiers, describing the two most commonly employed technologies—namely Solid State Power Amplifiers and Travelling Wave Tube Amplifiers, followed by a review of the typical configurations for the High- and Medium- gain antennas. Finally, a discussion on the current trends for improving the availability and performance of deep-space links through data relaying was provided. After looking at the conceptual advantages of multi-hop links and at the history of data relaying systems for users at or beyond Earth orbits, an outline of their expected performance was given. In doing so, the scenario of a deep space link relying on extremely high radio or optical frequencies was considered, since a performance increase with respect to a direct link would be more easily achieved at the highest frequencies. A review of recent literature on two-hop deep-space links highlighted that outperforming a direct link is feasible when optical or EHF are adopted for the deep space hop, provided that specific values of the transmitted power, onboard antenna size, and pointing accuracy can be met. In this respect, an estimate of system resources allocation suggested that significant efforts shall be devoted to pursuing novel antenna designs to enhance whether the gain-to-mass ratio (for optical terminals) or the gain-to-stowed size ratio (for foldable EHF antennas). In addition, the attitude control performance of the hosting spacecraft shall also be carefully considered, to avoid the antenna pointing losses from becoming a bottleneck to the overall effective system gain. Acknowledgments The authors are grateful to the Italian Space Agency (ASI) for their continuos support which helped developing significant expertise in radio tracking and radio science techniques, spanning from the data analysis domain to the communication systems analysis and design, including hardware and software components. PT, DM, EP and LV aknowledge support from the European Space Research and Technology Centre (ESA/ESTEC), under Grant 4000132053/20/NL/FE, for their work on two-legs deep space communication systems.

5 Ground and Space Hardware for Interplanetary Communication Networks

135

References 1. W.A. Imbriale, Large antennas of the deep space network, in Large Antennas of the Deep Space Network (2003) 2. J.H. Yuen, Deep Space Telecommunications Systems Engineering (Springer Science & Business Media, New York, 2013) 3. D. Buccino, J.S. Border, W.M. Folkner, D. Kahan, S. Le Maistre, Low-snr doppler data processing for the insight radio science experiment. Remote Sens. 14(8), 1–9 (2022) 4. A. Togni, M. Zannoni, L.G. Casajus, P.Tortora, An fft-based method for doppler observables estimation in deep space tracking, in 2021 IEEE International Workshop on Metrology for AeroSpace, MetroAeroSpace 2021 - Proceedings, vol. 5 (2021), pp. 294–299 5. Z.L. Mou, Y.Q. Tu, P. Chen, K. Wang, Accurate frequency estimation of multiple complex and real sinusoids based on iterative interpolation. Digital Signal Process. A Rev. J. 117, 103173 (2021) 6. H. Lehpamer, Microwave Transmission Networks: Planning, Design, and Deployment. (McGraw-Hill Education, New York, 2010) 7. G. Valente, M.N. Iacolina, R. Ghiani, A. Saba, G. Serra, E.Urru, G. Montisci, S. Mulas, S.W. Asmar, T.T. Pham, J. De Vicente, S. Viviano, The sardinia space communication asset: performance of the sardinia deep space antenna x-band downlink capability. IEEE Access 10, 64525–64534 (2022) 8. S.D. Slobin, T. Pham, 34-m bwg stations telecommunications interfaces, in DSN Telecommunications Link Design Handbook (2010), pp. 810–005 9. S.D. Slobin, T. Phan, 34-m hef subnet telecommunications interfaces. DSN Telecommunications Link Design Handbook (2008), pp. 810–005 10. T. Pham, J. Liao, E.J. Wyatt, B. Malphrus, J. Kruth, R. Kroll, M. Combs, T. Gedenk, S. Wilczewski, J. Lewis, A. Roberts, E. Mattle, C. Hart, System development and spacecraft testing of the Morehead state university ground station (2019). Pasadena, CA: Jet Propulsion Laboratory, National Aeronautics and Space 11. M. Calhoun, S. Huang, R.L. Tjoelker, Stable photonic links for frequency and time transfer in the deep-space network and antenna arrays. Proc. IEEE 95(10), 1931–1946 (2007) 12. D. Buccino, M. Soriano, K. Oudrhiri, S. Finley, D. Kahan, O. Yang, A. Jongeling, Detecting juno’s ’heartbeat’: communications support during critical events of the juno mission. IEEE Aerospace Conference Proceedings, vol 818 (2020) 13. C.L. Thornton, J.S. Border, Radiometric tracking techniques for deep space navigation. Radiometric Tracking Techniques for Deep Space Navigation (2005) 14. A. Modenini, B. Ripani, S. Member, For space missions a tutorial on the tracking , telemetry , and command ( tt&c ) for space missions (2022), pp. 0–30 15. W. Tai, Dsn frequency transition from l- to ka-band (Jet Propulsion Laboratory, Pasadena) (2023), p. 8. The source is: https://gsaw.org/wp-content/uploads/2019/10/2005s09a_tai.pdf. Retrieved on May 3rd 2023. 16. D.S. Network, Dsn telecommunications link design handbook, no. 810-005 (JPL, Pasadena, 2000) 17. R. Martin, M. Warhaut, Esa’s 35-meter deep space antennas at new norcia/western australia and cebreros/spain, in 2004 IEEE Aerospace Conference Proceedings (IEEE Cat. No. 04TH8720), vol. 2 (2004), pp. 1124–1133 18. P. Tortora, L. Iess, J. Bordi, J. Ekelund, D. Roth, Precise cassini navigation during solar conjunctions through multifrequency plasma calibrations. J. Guidance Control Dyn. 27(2), 251–257 (2004) 19. M.C. Comparini, F. De Tiberis, R. Novello, V. Piloni, L. Simone, D. Gelfusa, F. Argentieri, D.J. Fiore, F. Barletta, M. Delfino, S. Cocchi, F. Marchetti, A. Delfino, A. Rapposelli, G. Basile, Advances in deep-space transponder technology. Proc. IEEE 95(10), 1994–2008 (2007) 20. S.W. Asmar, S.J. Bolton, D.R. Buccino, T.P. Cornish, W.M. Folkner, R. Formaro, L. Iess, A.P. Jongeling, D.K. Lewis, A.P. Mittskus, R. Mukai, L. Simone, The juno gravity science instrument. Space Sci. Rev. 213(1–4), 205–218 (2017)

136

P. Tortora et al.

21. P. Cappuccio, V. Notaro, A. di Ruscio, L. Iess, A. Genova, D. Durante, I. di Stefano, S.W. Asmar, S. Ciarcia, L. Simone, Report on first in flight data of bepicolombo’s mercury orbiter radio science experiment. IEEE Trans. Aerospace Electron. Syst. 56(6), 4984–4988 (2020) 22. L. Simone, S. Ciarcia, G. De Angelis, F. Argentieri, D. Gelfusa, O. Cocciolillo, Ka-band transponders for radioscience: Product family evolution and in-flight operations, in TTC 2016 - 7th ESA International Workshop on Tracking, Telemetry and Command Systems for Space Applications, (2016) 23. L. Simone, O. Cocciolillo, D. Gelfusa, M. Belardinelli, M.G. Campagna, Novel x-band transponder for Euclid mission, in TTC 2016 - 7th ESA International Workshop on Tracking, Telemetry and Command Systems for Space Applications, (2016) 24. L. Iess, M. Di Benedetto, N. James, M. Mercolino, L. Simone, P. Tortora, Astra: interdisciplinary study on enhancement of the end-to-end accuracy for spacecraft tracking techniques. Acta Astronautica 94(2), 699–707 (2014) 25. E.C. Posner, L.L. Rauch, B.D. Madsen, Voyager mission telecommunication firsts. IEEE Commun. Mag. 28(9), 22–27 (1990) 26. C. Paoloni, D. Gamzina, R. Letizia, Y. Zheng, N.C. Luhmann, Millimeter wave traveling wave tubes for the 21st century. J. Electromag. Waves Appl. 35(5), 567–603 (2021) 27. J.D. Wilson, E.G. Wintucky, K.R. Vaden, D.A. Force, I.L. Krainsky, R.N. Simons, N.R. Robbins, W.L. Menninger, D.R. Dibb, D.E. Lewis, Advances in space traveling-wave tubes for NASA missions. Proc. IEEE 95(10), 1958–1967 (2007) 28. R. Strauss, J. Bretting, R. Metivier, Traveling wave tubes for communication satellites. Proc. IEEE 65(3), 387–400 (1977) 29. A.N. Curren et al., The Cassini mission Ka-band TWT, in Proceedings of 1994 IEEE International Electron Devices Meeting, (1994), pp. 783–786. https://doi.org/10.1109/IEDM. 1994.383307 30. W. Harvey, J. Shell, L.L. Burke, The flight operations history of the cassini x-band twtas, in 2018 IEEE International Vacuum Electronics Conference, IVEC 2018 (2018), pp. 11–12 31. M. Daniel Johnston, D. Martin, R.W. Zurek, W. Richard, The Mars reconnaissance orbiter mission: 2018 status, in Proceedings of the International Astronautical Congress, IAC, vol. 2018 (2018) 32. J.M. Weekley, B.J. Mangus, Twta versus sspa: a comparison of on-orbit reliability data, in 5Th IEEE International Vacuum Electronics Conference, IVEC 2004, vol. 52, no. 5 (2004), p. 263 33. V. Camarchia, R. Quaglia, A. Piacibello, D.P. Nguyen, H. Wang, A.V. Pham, A review of technologies and design techniques of millimeter-wave power amplifiers. IEEE Trans. Microwave Theory Tech. 68(7), 2957–2983 (2020) 34. M.M. Kobayashi, S. Holmes, A. Yarlagadda, F. Aguirre, M. Chase, K. Angkasa, B. Burgett, L. McNally, T. Dobreva, E. Satorius, The iris deep-space transponder for the sls em-1 secondary payloads. IEEE Aerosp. Electron. Syst. Mag. 34(9), 34–44 (2019) 35. W. Boger, D. Burgess, R. Honda, C. Nuckolls, X-band, 17 watt, solid-state power amplifier for space applications, in IEEE MTT-S International Microwave Symposium Digest, 2005 (2005), pp. 1379–1382 36. Y. Kobayashi, A. Tomiki, S. Kawasaki, Annual deep-space flight operation verification of xband gan sspa. IEEE Trans. Aerosp. Electron. Syst. Mag. 55(2), 930–938 (2018) 37. M.M. Kobayashi, State-of-the-Art Status for Deep-Space Smallsat Telecom 2018 Int’l Planetary Probe Workshop Short Course on Small Satellites Boulder (2018) 38. W.A. Imbriale, L. Boccia, Space antenna handbook. (2012). https://doi.org/10.1002/ 9781119945147 39. J.D. Vacchione, R.C. Kruid, A. Prata, L.R. Amaro, A.P. Mittskus, Telecommunications antennas for the juno mission to jupiter, in IEEE Aerospace Conference Proceedings (2012) 40. C.D. Edwards, T.C. Jedrey, E. Schwartzbaum, A.S. Devereaux, R. De Paula, M. Dapore, T.W. Fischer, The electra proximity link payload for mars relay telecommunications and navigation, in 54th International Astronautical Congress of the International Astronautical Federation (IAF), the International Academy of Astronautics and the International Institute of Space Law, vol.2 (2003), pp. 3505–3515

5 Ground and Space Hardware for Interplanetary Communication Networks

137

41. M.D. Johnston, J.E. Graf, R.W. Zurek, H.J. Eisen, B. Jai, The mars reconnaissance orbiter mission, in 2005 IEEE Aerospace Conference (2005), pp. 447–464 42. R.E. Hodges, N.E. Chahat, D.J. Hoppe, J.D. Vacchione, The mars cube one deployable high gain antenna, in 2016 IEEE Antennas and Propagation Society International Symposium, APSURSI 2016 - Proceedings (2016), pp. 1533–1534 43. E. Dotto, V. DellaCorte, M. Amoroso, I. Bertini, J. Brucato, A. Capannolo, B. Cotugno, G. Cremonese, V. DiTana, I. Gai et al., Liciacube-the light italian cubesat for imaging of asteroids in support of the nasa dart mission towards asteroid (65803) didymos. Planetary Space Sci. 199, 105185 (2021) 44. S.D. Slobin, T. Pham, Atmospheric and environmental effects, in DSN Telecommunications Link Design Handbook, DSN No. 810-005, Space Link Interfaces, Module 105, Rev. E (Jet Propulsion Laboratory, Pasadena, 2015) 45. D. Modenini, A. Locarini, L. Valentini, A. Faedi, P. Tortora, D. Rovelli, N. Mazzali, M. Chiani, E. Paolini, Two-leg deep-space relay architectures: Performance, challenges, and perspectives. IEEE Trans. Aerosp. Electron. Syst. 58(5), 3840–3858 (2022) 46. C. Sacchi, C. Schlegel, T. Rossi, M. Noble, M. Ruggieri, K.M. Cheung, M. Marchese, F. Granelli, V. Popescu, M. Rice, M. Murroni, N. Conci, Glue technologies for space systems: an introduction to a new aess technical panel. IEEE Aerosp. Electron. Syst. Mag. 35(1), 46–54 (2020) 47. U. di Bologna interdepartmental center for industrial researchin aerospace. Tesla two leg deep space relay – system requirements, architecture and performance. Technical Report, ESA Contract n 4000132053/20/NL/FE (2021) 48. R.A. Stampfl, A.E. Jones, Tracking and data relay satellites. IEEE Trans. Aerosp. Electron. Syst. 6(3), 276–289 (1970) 49. B. Kopp, J. Harris, C. Lauand, Utilizing existing commercial geostationary earth orbit fixed satellite services for low earth orbit satellite communication relays with earth. New Space 7(1), 19–30 (2019) 50. K. Hogie, E. Criscuolo, A. Dissanayake, B. Flanders, H. Safavi, J. Lubelczyk, Tdrss demand access system augmentation, in IEEE Aerospace Conference Proceedings, vol. 2015 (2015) 51. O. Laux, D. Poncet, R. Mager, K. Schoenherr, Status of the European data relay satellite system, in 2012 International Conference on Space Optical Systems and Applications (ICSOS), (2012), pp. 9–12 52. Y. Satoh, Y. Miyamoto, Y. Takano, S. Yamakawa, H. Kohata, Current status of japanese optical data relay system (jdrs), in 2017 IEEE International Conference on Space Optical Systems and Applications, ICSOS 2017 (2018), pp. 240–242 53. D.J. Israel, H. Shaw, Next-generation nasa earth-orbiting relay satellites: fusing optical and microwave communications, in IEEE Aerospace Conference Proceedings, vol. 2018 (2018), pp. 1–7 54. D.M. Boroson, B.S. Robinson, D.A. Burianek, D.V. Murphy, A. Biswas, Overview and status of the lunar laser communications demonstration, in Free-Space Laser Communication Technologies XXIV, vol. 8246 (2012), p. 82460C 55. H. Hauschildt, S. Mezzasoma, H.L. Moeller, M. Witting, J. Herrmann, European data relay system goes global, in 2017 IEEE International Conference on Space Optical Systems and Applications, ICSOS 2017 (2018), pp. 15–18 56. H. Hauschildt, C. Elia, H.L. Moeller, W. El-Dali, T. Navarro, M. Guta, S. Mezzasoma, J. Perdigues, Hydron: high throughput optical network, in 2019 IEEE International Conference on Space Optical Systems and Applications (ICSOS) (2019), pp. 1–6 57. B.L. Edwards, D. Israel, K. Wilson, J.D. Moores, A.S. Fletcher, International Conference on Space Optical Systems the Laser Communications Relay Demonstration, vol.12 (2012), pp. 1–9 58. K.E. Wilson, M. Wright, R. Cesarone, J. Ceniceros, K. Shea, Cost and performance comparison of an earth- orbiting optical communication relay transceiver and a ground-based optical receiver subnet, (2003)

138

P. Tortora et al.

59. J.Hunter, The orbiting deep space relay station-a study report, in Conference on Large Space Platforms: Future Needs and Capabilities (1978), p. 1639 60. H. Hemmati, Deep Space Optical Communications (John Wiley & Sons, New York, 2006) 61. M. Wittig, Data relay for earth, moon and Mars missions, in 2009 International Workshop on Satellite and Space Communications (2009), pp. 300–304 62. H. Hemmati, A. Biswas, I.B. Djordjevic, Deep-space optical communications: Future perspectives and applications. Proc. IEEE 99(11), 2020–2039 (2011) 63. R. Cesarone, D. Abraham, S. Shambayati, J. Rush, Deep-space optical communications, in 2011 International Conference on Space Optical Systems and Applications (ICSOS) (2011), pp. 410–423 64. D.M. Cornwell, Nasa’s optical communications program for 2017 and beyond, in 2017 IEEE International Conference on Space Optical Systems and Applications (ICSOS) (2017), pp. 10–14 65. W.J. Hurd, B.E. MacNeal, G.G. Ortiz, R. Moe, J.Z. Walker, M. Dennis, E. Cheng, D. Fairbrother, B. Eegholm, K.J. Kasunic, Exo-atmospheric telescopes for deep space optical communications, in 2006 IEEE Aerospace Conference (2006), p. 12. 66. E. Gramigna, J.G. Johansen, R.L. Manghi, J. Magalhães, M. Zannoni, P. Tortora, E. Le Bras, A. Togni, Hera inter-satellite link doppler characterization for didymos gravity science experiments, in 2022 IEEE 9th International Workshop on Metrology for AeroSpace (MetroAeroSpace) (2022), pp. 430–435 67. J.C. Breidenthal, The merits of multi-hop communication in deep space, in IEEE Aerospace Conference Proceedings, vol.1 (2000), pp. 211–221 68. M.O. Hasna, M.-S. Alouini, End-to-end outage probability of multihop transmission over lognormal shadowed channels (2003). Citeseer 69. L.P. Clare, J.L. Gao, E.H. Jennings, C. Okino, Space-based multi-hop networking. Comput. Netw. 47(5), 701–724 (2005) 70. L. Valentini, A. Faedi, E. Paolini, M. Chiani, Analysis of pointing loss effects in deep space optical links, in 2021 IEEE Global Communications Conference (GLOBECOM) (2021), pp. 1–6 71. C.D. Brown, Elements of Spacecraft Design (AIAA, Reston, 2002) 72. K. Bhasin, J. Warner, S. Oleson, J. Schier, Design concepts for a small space-based geo relay satellite for missions between low earth and near earth orbits, in 13th International Conference on Space Operations, SpaceOps 2014 (2014), pp. 1–19 73. G. Codispoti, C. Cornacchini, S. Falzini, The Alphasat TDP # 5 Experiment and Challenges for Future Q / V-band Systems Exploitation. Technical challenges for the Q / V- band exploitation : The Alphasat TDP # 5 communications and propagation experiments : Description and rationale (2011) 74. E. Cianca, T. Rossi, A. Yahalom, Y. Pinhasi, J. Farserotu, C. Sacchi, Ehf for satellite communications: the new broadband frontier. Proc. IEEE 99(11), 1858–1881 (2011) 75. N. Deo, High power transmitters for q/v-band communications-beyond alphasat, in 2019 IEEE Aerospace Conference (2019), pp. 1–6 76. C.K. Chong, D.A. Layman, W.L. McGeary, W.L. Menninger, M.L. Ramay, X.Zhai, L3 technologies edd q/v-band helix twt for future high-data-rate communications uplink applications, in IVEC 2017 - 18th International Vacuum Electronics Conference, vol. 2018, no. 6 (2018), pp. 1–2 77. N. Deo, High performance transmitters for small satellites for data transmission and remote sensing, in IEEE Aerospace Conference Proceedings, vol. 2019 (2019), pp. 1–6 78. N.C. Deo, Solid-state transmitters and sources for remote sensing radars, instruments and communication, in Asia-Pacific Microwave Conference Proceedings, APMC, vol. 2018 (2019), pp. 977–979 79. A. Biswas, H. Hemmati, S. Piazzolla, B. Moision, K. Birnbaum, K. Quirk, Deep-space optical terminals (dot) systems engineering. IPN Progress Rep. 42, 183 (2010) 80. H. Hemmati, M.W. Wright, A. Biswas, C. Esproles, High-efficiency pulsed laser transmitters for deep-space communication. Free-Space Laser Communication Technologies XII, vol. 3932 (2000), pp. 188–195

Chapter 6

End-to-End Space System: Engineering Considerations Kar-Ming Cheung

6.1 Introduction Many research and development (R.&D) efforts in the fields of communications and networking of space systems have a tendency to concentrate on one specific capability, and do not consider its interactions with other functions in the signal processing and data flow chain, and with its environment. This focused view sometimes impairs the efficiency and the utility of the capability in operations. System engineering and architecture are after-thoughts in many cases, and the efforts are mostly qualitative, empirical, hands-on, and pragmatic in nature. This chapter attempts to treat the subject of end-to-end communication and network system design and optimization in a more methodical and analytical manner. The discussion is unique compared to other books on space communications and networks in the following ways: • It describes a holistic approach to design and to evolve a space communication network taking into account the relative geometry between the spacecraft, ground network, and the planetary bodies, as well as link layer and network layer behavior models that support high-performance and high-fidelity modeling and simulations. This includes the use of automated re-transmissions of lost data packets in a noise channel to ensure reliable communications. • It does not bog down on the design specifics of individual communication systems and network components. Instead, it investigates the dependencies and interactions between components and with the environment, with the goal of achieving an optimal system-level design.

K.-M. Cheung () Jet Propulsion Lab, California Institute of Technology, Pasadena, CA, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_6

139

140

K.-M. Cheung

Fig. 6.1 Data flow

• It helps to bridge the gap and to plot a reasonable evolving path between early conception of an architecture, which is studded with design uncertainties and the “end game”, which is characterized with rigid form factor and performance requirements and constraints. • This chapter introduces the use of simulation-based and analytical constrained optimization techniques that maximize the end-to-end data delivery efficiency of the communication systems and networks. To ensure a manageable scope and to illustrate the concepts, this chapter focuses on the communication downlink direction only. This includes onboard data generation, transmission and/or re-transmission through a noisy channel, and reception and confirmation by the ground antenna. Figure 6.1 shows the high-level downlink data flow. Space communication with ground network differs from other spacecraft functions in the following way: 1. Most spacecraft functions are localized onboard the spacecraft, for example, power, thermal, attitude control, etc. The communication and tracking functions depend on the resources offered by both sides of the link—the spacecraft and the network. For a given communication and tracking performance requirement, one can trade-off the resources of one with the other. 2. Onboard data generation is statistical. Instrument data, images, video, and spacecraft telemetry are digitized and are compressed, and the efficiency of the data reduction process is data content dependent, and is not known a priori. 3. The communication and tracking performance are dependent upon the spacecraft, the network, and the space environment. Some link parameters are statistical, e.g., space and ground antenna pointing’s. Link performance is affected by receiver’s thermal noise as well as noise emitted by celestial objects. For spacecraft communicating with an Earth-based ground station, the link is also affected by weather effects in terms of atmospheric attenuation and atmospheric noise temperature increase.

6.2 Statistical Nature of Onboard Data Generation Many spacecraft instruments employ data compression schemes to reduce the data generation bandwidth. Many data compression schemes consist of a lossless or

6 End-to-End Space System: Engineering Considerations

141

lossy transform (or predictive) stage1 that generates a de-correlated data stream of integers (indices), and a lossless entropy coding stage that converts the data stream into a smaller representation of the original data. The entropy coding process is a fixed-to-variable data conversion process whose efficiency is dependent on the data content and the compression algorithm used. The data generation rate at the output of the entropy coder is inherently statistical, as the instantaneous data rate is not known in advance. Moreover variable-rate compressed and fixed-rate uncompressed data from multiple instruments may be multiplexed into a single data stream before sending to the onboard data storage or to the flight communication system for downlink. The spacecraft bus bandwidth and the onboard data buffer have to accommodate the statistical fluctuations of the data generation rate and the onboard data volume. For larger spacecraft with multiple onboard instruments that generate data independently,2 the overall onboard data generation tends to have a Gaussian-like distribution due to the averaging effect.

6.3 Statistical Nature of Communication Link The purpose of communication is to send the information-bearing signals for one point to another through a channel. For RF communication, the signal is in the form of electromagnetic wave, and the channel is the pathway or medium between the two points. For space communication the channel might be just free-space. Or if signal passes through the atmosphere, weather effects would affect the link performance. Communication channel introduces two kinds of losses to the signals: path loss (signal attenuation), and additive noise. Path loss is the reduction of power of an electromagnetic wave. It may be due to many effects like free-space loss, refraction, diffraction, reflection, and absorption. Additive noise may include the equipment noise on the receiving end, and the galactic noise and the hot body noise detected by the receiving aperture. The additive noise is typically characterized as wideband white noise with s constant spectral density, and a Gaussian distributed amplitude. Upon reception of the signals on the receiving end, the signal processing chain attempts to detect and recover the information from the received signals that are corrupted by attenuation and noise. The effects of signal attenuation and additive noise are both statistical. Figure 6.2 shows the relationship between a discrete signal model and the link model.

1 For lossy compression, the transform stage includes a quantization stage that reduces the transform output values to a finite set of integer numbers. 2 No dominant instrument that generates disproportional amount of data.

142

K.-M. Cheung

Fig. 6.2 Relationship between discrete signal model and link model: .Rb is the data-rate, .Ls the free-space pathloss, .Lo other losses, M the margin, K the Boltzmann’s constant,   EI RP the Eb effective isotropic radiated power, . G the gain over noise temperature and . N0 T h the signalT to-noise ratio threshold

6.4 Link Optimization, Reliability, and Margin Policy In this section, we discuss the role of margin in link design and optimization for links between a spacecraft and a ground network. Link analysis is an indispensable system-engineering process used for sizing up the spacecraft communication system design and in planning for mission data return. The process is iterative in nature, and it is used in all phases of a mission lifecycle—from proposal phase through design, development, and operation phases. For space missions in the early mission phases, link analysis emphasizes finding the right spacecraft communication system components (e.g., antenna and power amplifier) that would meet the mission data return requirements. In later phases after the flight system design and mission operation concept are mature, link analysis is used to estimate the detailed data return profile of the mission. Link analysis consists of the calculation and tabulation of the useful signal power and the interfering noise power available at the receiver. The signal and noise terms in the link equation are mathematical abstractions of the performance behavior expressed in decibels (dB), and by summing up these terms, one can generate an overall signal-to-noise ratio (SNR) estimate that can be used to characterize communication system performance, to support system design tradeoff, and to manage the operational risks associate with the usage of a link. The goal of link analysis is to maximize the data throughput over a noisy channel, yet to maintain the integrity of the data. One important consideration in link analysis is the allocation of link margin, which is a balancing act between data return and communication reliability. There are inherent uncertainties associated with the signal power (e.g., atmospheric attenuations) and the noise power (e.g., equipment noise and hot body noise) at the receivers. Link margin is defined as the additional SNR (in dB) that imposes on top of a given SNR design point to guarantee the link design would meet a given data integrity requirement, which is typically expressed in bit-error-rate (BER) or frameerror-rate (FER). The current link analysis approaches used by telecommunication

6 End-to-End Space System: Engineering Considerations

143

engineers either adopt some rule-of-thumb link margin policy irrespective of the signal and noise statistics, or they compute a link margin quantity that addresses a different metrics, namely the probability of ‘not closing the link’, instead of computing the SNR that meets the given data integrity requirement. In this chapter, we quantify statistically the relationship between the BER/FER requirement, the operating SNR, and the coding performance curve. We compute the “minimum” SNR design point that would meet the BER/FER requirement by taking into account the fluctuation of signal power and noise power at the receiver, and the shape of the coding performance curve. We assume that no time diversity technique (e.g. interleaving) is used, and the channel coherence time of link fluctuation is comparable or larger than the time duration of a codeword length such that the timefluctuation of the link does not get randomized or averaged out within a codeword. This analysis yields a number of valuable insights on the design choice of coding scheme and link margin for the reliable data delivery of a communication system— space and ground.

6.4.1 Review of Link Analysis Techniques As the subject of link analysis, particularly the statistical link analysis, is not widely popular in the literature, we provide a brief overview on this topic to make this chapter more “self-contained.” Link analysis starts with the following link equation:   EI RP G Pt T . = N0 kLs Lo

(6.1)

where . NPt0 is the total power to noise power spectral density ratio, EI RP is the effective isotropic radiator power of the transmitter, . G T is the “Gain over System Noise Temperature” which is a measure of the receiver sensitivity, k is 2  the Boltzmann’s constant (.1.38 × 10−23 J /K), .Ls = 4πλ d is the space-loss where d is the distance between transmitter and receiver, .λ is the wavelength, and .Lo denotes all other losses and degradation factors not specifically addressed in Eq. (6.1). The EI RP term includes all of the gain and loss terms on the transmission side including pointing loss, the . G T term includes all of the gain and loss terms on the receiver side, and the .Lo term includes contributions of the intervening transmission media. Note that the link Eq. (6.1) is multiplicative in nature. By taking the base-10 logarithm and multiplying by 10 on both sides of (6.1), we convert the multiplicative relationship of the gain and loss terms to become an additive relationship. The additive terms are expressed in units of decibels (dB). Equation (6.1) can therefore be re-written as: .

Pt (in dB) = EI RP (in dB)) − k (indB) −) − Ls (in dB) − Lo (in dB) N0

(6.2)

144

K.-M. Cheung

Depending on the link environment, the system noise temperature T in the . G T term sometimes includes a number of components that are additive, namely, the equipment noise temperature, the atmospheric noise, and the cosmic background noise, etc. There are two major schools of thought on link analysis—the link budgeting approach and the statistical link analysis approach.

6.4.1.1

Link Budgeting Approach

The link budgeting approach assumes that the link parameters—the gain and loss terms of a link, are all single (deterministic values. The SNR at the receiver is computed by summing up the link parameters (in dB) along the signal processing chain as shown in Eq. (6.2). A rule-of-thumb margin policy, typically 3 dB, is then imposed to compute the supportable data rate for the given link design. This approach is simple, and has been popular in the analysis of communication links that are not power constrained. The problems with this approach are as follows: 1. Some link parameters are inherently statistical, for example, polarization loss, antenna pointing loss, and different weather effects. By restricting the link parameters to be single-value, the worst-case values are typically chosen for use. This introduces a systematic bias in link analysis towards the pessimistic direction. 2. There is no mathematical and statistical justification for the choice of 3-dB margin policy (a factor of 2). Why is a 3-dB margin enough, or is 3-dB too much?

6.4.1.2

Statistical Link Analysis

The concept of statistical link analysis relies on the additive nature of the link equation as given in Eq. (6.2). Instead of treating the gain and loss terms (with units of dB) as deterministic values, the link parameters are treated as random variables.3 Yuen formulated the analysis framework of statistical link analysis in the 1970’s [1]. Since then, the Jet Propulsion Laboratory (JPL) has adopted this approach as a flight principle to conduct link analysis for its deep space missions. The JPL Projects and the Deep Space Network (DSN) measure the performance statistics of hardware components, and they conduct experiments to characterize the statistics of weather effects on the link. These statistical data are folded into the statistical link analysis process.

3 Some of the link parameters, like transmission power and antenna gains, can still be treated as deterministic values. Their pdf’s are just Delta Dirac functions.

6 End-to-End Space System: Engineering Considerations

145

Fig. 6.3 Conversions between design, favorable tolerance, and advance tolerance, and mean and variance of some popular probability density functions (from [1])

JPL’s statistical link analysis methodology can be summarized as follows: without loss of generality, we express the link parameters as .xi ’s. Each statistical link parameter .xi can be described in terms of a design value .xdesign,i , a minimum value .xmin,i a maximum value .xmax,i , and a probability distribution function (pdf) .fi (xi ) such that that .fi (xi ) = 0 for .xmin,i < X < xmax,i , and .fi (xi ) = 0 for .xi < xmin,i and .xi > xmax,i . Some common forms of .f (x) are the rectangular (or uniform), triangular, and Gaussian distributions.4 From this setup, one can deduce the mean of x (denoted by m) and the variance of x (denoted by .σ 2 ). Let’s denote the design value .Dx = xdesig , and define the favorable tolerance .Fx = xmax − xdesign and the adverse tolerance .Ax = xmin − xdesign . The computations of the mean and variance of the uniform, triangular, and Gaussian distribution are given in Fig. 6.3. Assume that there are n link parameters .xi ’s that are independent. The ensemble   x has a mean .mz = of these link parameters .z = i i i mxi and a variance 2 2 . The pdf of z, which we denote as .f (z), can be computed by .σz = σ x i i convolving .fx1 (x1 ) , fx2 (x2 ) , · · · , fxn (xn ). This is in general a computationally intensive process as this involves .n − 1 levels of integration. To simplify the computation, when a large number of independent link parameters are added together (in dB), Yuen proposed to approximate received signal-to  the resulting noise ratio term with a Gaussian distribution .N mz , σz2 , where .mz is the mean and 2 .σz is the variance as defined above. From this, one can design a link and establish link margin policy based on statistical confidence level measured in terms of the .σ of a Gaussian distribution function (e.g., .2 − σ event, .3 − σ event etc.) Note that in

4 Strictly speaking the Gaussian distribution is unbounded. In link analysis it is typically used to model certain weather effects or to model the combined effect of a number of link parameters (derived parameter).

146

K.-M. Cheung

general link parameters have different means, variances, and pdf’s; thus the above Gaussian approximation approach does not conform to the sufficient conditions of the classical Central Limit Theorem, which requires that all the link parameters be independent and identically distributed. The procedure outlined by Yuen [1] did not justify this Gaussian approximation in a mathematically rigorous manner, and it did not explicitly state the conditions under which this Gaussian approximation is valid. However, decades of experience show that for links where there are many link parameters, this approach works well in most cases and closely approximates the Gaussian distribution. This is particularly true for S- and X-band links that are less susceptible to the weather effects. In [2], the author fills this gap by invoking a variant of the Central Limit Theorem known as the “Lyapunov’s condition” that provides the sufficient condition for the aforementioned Gaussian approximation to be valid under the condition that the link parameters do not include one or more “dominant terms”. A dominant term is one that has a variance that greatly exceeds the variances of the other terms. In [3], using sound statistical principles on Ka-band link analysis, the author describes a new technique that incorporates the dominant link terms in statistical link analysis by using a weighted sum of Gaussian distributions with the same variance and shifted means. The current statistical link analysis approach expresses link reliability as the likelihood of exceeding the SNR threshold that corresponds to a given BER/FER requirement. The method, however, does not provide the true BER or FER estimate of the link with margin, or the required SNR that would meet the BER or FER requirement in the statistical sense.

6.4.1.3

Minimum Margin for Link Design and Optimization

In standard telecommunication system engineering, requirements on link design and link reliability are typically expressed in two parts: (1) tolerable error rate of received data, and (2) link margin policy of the link. The tolerable error rate is typically expressed as BER or FER, and the choice is dependent upon the required quality of the received data. For example, uncompressed spacecraft telemetry data typically can tolerate a higher error rate, whereas compressed instrument data would require a lower error rate that minimizes error propagation [4]. Once the BER/FER requirement is established, a telecommunication system engineer would look up the SNR threshold value on the coding performance curve that delivers the required BER/FER. Now knowing that many link parameters are inherently statistical,5 the received SNR can take on random values within a certain range about the SNR design point. To ensure link reliability, a link margin is imposed onto the link design such that the SNR operates at a value

5 In link budgeting approach, the assumption that link parameters take on deterministic values (most likely worst-case values), and the need to adopt a link margin policy seem schizophrenic.

6 End-to-End Space System: Engineering Considerations

147

higher than the SNR threshold such that the likelihood that the SNR would dip below the SNR threshold is sufficiently small. The standard approach to generate a coding performance data point is by simulating and/or emulating the processes of encoding a known information bit stream, sending the encoded symbols through a noisy channel representative of a given SNR, and decoding the received symbols back into the information bit stream. The original bit stream is then compared to the decoded bit stream to collect undecodable errors that are used to estimate the code’s error rate for the given operating SNR. As the confidence level in estimating the code’s error rate depends on the number of error events, it takes more computation effort to generate data points at high SNR than at low SNR. Once a sufficient number of data points are generated, the data points are interpolated and/or sometimes extrapolated to form the coding performance curve. We denote this coding performance curve as .y = h (x) , where x is the SNR (in dB). However, during a communication session, signals can be attenuated by various unpredictable non-ideal operation effects and natural phenomena, and different random noises can be added to the receiver, the received SNR at the receiver is in fact a random variable. Without loss of generality, we denote the distribution of the received SNR to be .f (x | ·). In Sect. 6.4.1.2 and in [2] and [3], we show that this SNR fluctuation can be modeled as a Gaussian process when there is no dominant component in the link. In this case the SNR distribution can be expressed as: (x−m)2 1 f (x | m; σ ) = √ e 2σ 2 2π σ

.

(6.3)

where m is the mean of the link parameters (in dB) and .σ is the standard deviation. Also m is the maximum likelihood estimate of the SNR in the Gaussian case, so m is chosen to be the SNR design value that the link analysis is based on. When there are one or more dominant components, and when empirical measurements for each of the dominant components exist, we show in [3] that the SNR fluctuation can be modeled as a sum of Gaussian with shifted means. Using the same notations as in [3], for a given weather availability cumulative distribution CD, where .10 ≤ CD ≤ 99%, we define the discrete random variable .LD with a finite set of values .{l10 , l11 , · · · , lCD } with probability .{P10 , P11 , · · · , PCD }, where .li corresponds to the i-th percentile value of the total weather loss .LdB tot (θ ), where .θ is the elevation angle of the ground antenna, and m and .σ are the mean and standard deviation respectively of the link parameters excluding the weather effects. In this case .P10 = 1 10 CD , and .Pi = CD for .10 ≤ i ≤ CD%.  √1 The SNR distribution can be expressed as .f (x | m; σ ; LD ) = CD i=10 Pi 2πσ

(x−m−li )2 2σ 2

e . For this non-Gaussian case, there is no simple analytical way to find the value of x that maximize .f (x | m; σ ; LD ), or to show that if .f (x | m; σ ; LD ) is

.

148

K.-M. Cheung

¯ where uni-modal at all. A reasonable SNR design point would be .m ˆ = m + l,  P l . Re-writing . f | m; σ ; L the SNR distribution can be expressed l¯ = CD (x ) D i=10 i i as:

.

CD ¯ )2 ˆ l−l (x−m+ i    1 2σ 2 f x | m; ˆ σ ; LD = Pi √ e 2π σ i=10

.

(6.4)

In the subsequent sections, we will evaluate the mean error rates for link variations typical for S- and X-bands (.σ = 0.5 and .σ = 1.0) and for Ka-band (.σ = 1.0). To simplify the discussion for Ka-band, we will use the Gaussian approximation Equation (6.3) instead of the more complicated expression in Eq. (6.4) that requires statistics of local weather loss measurements. By averaging the error rate .h (x) over the distribution .f (x | m; σ ), the mean error rate .e¯ (x, σ ) for a given SNR design point x is given by:  e¯ (x, σ ) =

+∞

.

−∞

h (y) f (y | x; σ ) dy

(6.5)

We will demonstrate by examples in the next section that .e¯ (x, σ )  h (x) for all x. Or equivalently we can say that for a given error rate . and SNR’s .s1 and .s2 such that . = h (s1 ) = e¯ (s2 , σ ), .M = s2 − s1 is the additional SNR, or the minimum margin, required on top of the ideal SNR design point to guarantee that the link design would meet the given error rate requirement .. We will discuss this concept in more detail in Sect. 6.4.1.4, using the link-adjusted coding performance curves generated in the next section (Sect. 6.4.1.3) for some NASA codes.

6.4.1.4

Illustrating Examples

We use Eq. (6.5) in Sect. 6.4.1.3 to evaluate the link-adjusted error rate curve y = e¯ (x, σ ) for a number of popular error-correcting codes6 that are used for NASA missions, and investigate the minimum margin required to meet an error rate requirement typical of each code:   (a) . 7, 12 convolutional code, .BER = 10−5 .   (b) Concatenated code ((255, 223) Reed-Solomon Code and . 7, 12 convolutional

.

code, interleave depth = 5), .BER =10−5and .BER = 10−7 . (c) Low-Density Parity Check (LDPC) . 7, 12 code, .BER = 10−5 . For each of the code, we compute the .e¯ (x, σ ) for the following cases:

6 The

analytical expressions of the ideal link performance curves .h (x)’s are derived from an informal JPL student report by Adrienne Lam, a former intern student.

6 End-to-End Space System: Engineering Considerations

149

  Fig. 6.4 Comparison of error rate performances for the . 7, 12 convolutional code under different SNR variations

(a) .σ = 0.0—the ideal case of traditional error rate curve assuming no SNR variation. (b) .σ = 0.5—the typical SNR variation for DSN S- and X-band links. (c) .σ = 1.0—the typical SNR variation for non-DSN S- and X-band links. (d) .σ = 1.5—the typical SNR variation for Ka-band links. The results are shown in Figs. 6.4, 6.5 and 6.6.

6.4.1.5

Analysis Insight and Concept of Minimum Margin

The link-adjusted coding performance curves in Figs. 6.4, 6.5, and 6.6 reveal a number of interesting and important behaviors when compared to the ideal error rate curves that assumes no SNR variation: (a) The link-adjusted coding performance curves are always inferior compared to the ideal error rate curve for the same code. This can be explained as follows. An ideal “waterfall” error rate curve of a reasonable code has a shape that concaves downward in an error rate (in log scale) versus SNR (in dB) plot. As the error rate is expressed in log scale and the SNR variation is symmetric about the SNR design point for the Gaussian distribution, the link-adjusted mean error

150

K.-M. Cheung

Fig. 6.5 Comparison of error rate performances for the concatenated code under different SNR variations

rate .e¯ (x, σ ) is more biased by the higher error rate on the left side of the SNR design point. This results in a higher mean error rate for the same SNR design point as shown in Figs. 6.4, 6.5, and 6.6. (b) A powerful code typified by a steep slope in the code performance plot is more sensitive to SNR variation, thus losing more coding gain compared to an average code. This follows the same reasoning in (a) that the link-adjusted mean error rate .e¯ (x, σ ) can be a lot more biased by the higher error rate on the left side of the SNR design point for a coding performance curve with a steep slope. For example, to compensate for the SNR variation of .σ = 1.5 for  a BER requirement of .10−5 , Fig. 6.6 shows that the powerful LDPC . 1024, 12 code relinquishes a coding gain of 4.6 dB compared   to the constant SNR case. Conversely, Fig. 6.4 shows that the modest . 7, 12 convolutional code only requires 2.8    dB. Thus, for high SNR variation scenario likes Ka-band, the LDPC 1 1 . 1024, 2 code is only better than the . 7, 2 convolutional code by 0.6 dB in coding gain, and the coding gain of the concatenated code is the same as that of the convolutional code. (c) It takes more coding gain to compensate for the SNR variation in the link at lower error rate. As shown in Figs. 6.4, 6.5, and 6.6, the link-adjusted coding performance curves all fan outward in the direction of lower error rate. In the

6 End-to-End Space System: Engineering Considerations

151

  Fig. 6.6 Comparison of error rate performances for the LDPC . 1024, 12 code under different SNR variations

case of the concatenated code in Fig. 6.3, it takes 4.8 dB of coding gain to compensate for the link SNR variation .σ = 1.5 at .BER = 10−5 . Whereas in the case of .BER = 10−7 , the loss in coding gain is 5.4 dB. (d) The link budget’s rule-of-thumb margin policy of 3 dB or the statistical link analysis’ margin of .2σ may not be enough for links with large SNR variation in the link. In the case of concatenated code operating at .BER = 10−7 as shown in Fig. 6.5, the additional  SNR required to compensate for .σ = 1.5 is 5.5 dB. 1 For the LDPC . 1024, 2 code operating at .BER = 10−5 , the additional SNR required is 3.6 dB. The above examples illustrate the concept of using additional SNR (in dB) to compensate for the SNR variation in the link, so as to maintain the same link reliability as promised by the ideal coding performance curve. We can view this additional SNR to be the “minimum” margin required to offset the “known unknowns” of the link. This “minimum” margin is particularly profound in links with large SNR variation like the Ka-band links. An in-depth understanding on the relationship between coding gain, required “minimum” margin, and SNR variation of a link is particularly important in the optimal design and efficient operation of a reliable Ka-band link.

152

K.-M. Cheung

The above results also provide some guidance on the operation of a dynamic link: (a) During the Mars Reconnaissance Orbiter (MRO) Ka-band operation experiment between August 2005 and March 2006, it was observed that the measured symbol SNR .(Es/N o) variation is a lot higher during the rise and set of a pass [5].7 This is due to the longer signal path that traverses through the atmosphere at low elevation angle. Another observation in the Ka-band operation experiment is that the ratio between X-band link variation and Kaband variation is about 1:4,8 and this compares well with the analytical results of typical .σ value of 0.5 dB for DSN X-band links, and 1.8 dB for DSN Ka-band links. Thus, for Ka-band a higher link margin might be needed at the beginning and the end of a pass than the margin in the middle of a pass. (b) For communication system with Variable Coding and Modulation (VCM) capability that adapts to the dynamic link environment, one needs to take into account the SNR variation in addition to the mean SNR to determine the choice of the data rate, coding scheme, and modulation scheme.

6.4.2 ARQ Links for Reliable Communications and Its Statistical Characterization Current deep space communication links in S-band (2.3 GHz) and X-band (8.4 GHz) are characterized by the memoryless Additive White Gaussian Noise (AWGN) channel. The signal and noise statistics within a communication pass are relatively quiescent, and the channel is dominated by random errors. As such, the concept of operation usually favors the simple data-layer approach of sending the data only once through the channel. The communication system design including error correction coding schemes primarily focuses on the trade between power,9 spectrum,10 and complexity,11 and collectively meets a given data quality requirement expressed in terms of the bit-error-rate (BER) or the frame-error-rate (FER). The future deep space links are migrating towards higher frequency bands such as Ka band (32 GHz for deep space, 26 GHz for near-Earth) and optical. These links are susceptible to atmospheric attenuation and non linear effects such as turbulence, scintillation, antenna mis-pointing, jitter, etc. These dynamic links thus will experience various degrees of fast and slow fading losses, and some of

7 See

Figures 11 and 13 of [5]. communication with Shervin Shambayati. 9 Energy per information bit, usually expressed as Signal-to-Noise Ratio (SNR). 10 Channel bandwidth required for a given data rate, including bandwidth expansion due to coding. 11 This refers to the required hardware and processing for communications, primarily on the spacecraft side. 8 Private

6 End-to-End Space System: Engineering Considerations

153

these link disruptions cannot be sufficiently and practically mitigated by the “sentonce” approach using only forward error correction coding and/or interleaving. One effective way to ensure reliable communication is using a network-layer protocol called Automatic Repeat reQuest (ARQ), where the receiver acknowledges to the transmitter whether or not a data unit is successfully received. If a data unit is not successfully received (such as after a pre-set time-out), the transmitter would then re-transmit the lost data unit to the receiver. The ARQ approach therefore adds a new dimension of trade in the communication system design, namely data latency, to the standard considerations of power, spectrum, and complexity. This metric of data latency is an attractive quantity in the system design trade for long-haul links where light-time delay is already high, and/or for bulk science data that can tolerate high latency. The fundamental concept of ARQ protocol is that when data are corrupted during transmission, messages can be re-sent multiple times until they are received and acknowledged. We assume that all data transmissions use error-correction coding for channel error-correction and/or error-detection. Much work has been done in the performance analysis of ARQ protocols. Throughput and latency analyses can be found in early papers [6, 7] under the assumption that code-block errors occur independently. To analyze wireless communication channels that are characterized by fast fading and bursty errors, recent literature introduces channel models that assume an error process that is not random and is modeled as a Markovian process [8–11]. In this chapter we first consider the case when the ARQ system has no limit on the number of re-transmissions. This provides the upper bound of ARQ link performance. We then discuss the more practical case of truncated ARQ links that allows a maximum of K re-transmissions. With unlimited number of re-transmissions, an ARQ link is “error-free” in the sense that a data frame will eventually be successfully delivered (at the first transmission or a subsequent re-transmission). However, the penalties for ARQ link are (1) increased latency for re-transmission and (2) reduced link efficiency (measured in higher power or lower data rate) to accommodate the re-transmitted data frames. Thus, the key metrics to measure the quality of the “error-free” ARQ link (with unlimited number of re-transmissions) are: 1. Transmission latency in some statistical sense (e.g., maximum latency, mean latency, etc.) 2. Effective data rate .Rf in terms of the net data throughput, discounting the portion of the bandwidth that accommodates re-transmissions. The concept of effective data rate is also applicable to “send-once” links.12 Assuming the smallest data unit to be a frame, and denoting .Pblc as the Frame-

12 Links when messages are only sent once. When there are uncorrectable errors, the messages are lost.

154

K.-M. Cheung

Error-Rate (FER), the effective data rate .Reff in terms of the amount of reliable data available on the received side can be measured as: Reff = Rb (1 − Pblc )

.

(6.6)

where .Rb is the raw data rate. Note that in this interpretation of effective data rate for the “sent-once” link, .Reff includes the portion of the data frames that are successfully received. The corrupted data frames are lost and are thus nonrecoverable. In many prior ARQ studies and system designs, including the recent ones [12, 13] by the main author of this chapter, the assumption is that the SNR is fixed throughout the ARQ communication session. This assumption can be reasonable for the following cases: 1. When the ARQ turnaround time is much shorter than the channel coherency time. 2. When the links are relatively static, like the S-band and X-band links, when the SNR is not expected to change significantly during a communication session. In deep space communications, the ARQ turnaround time includes the roundtrip light time that can be tens of minutes long, plus the data processing time at the receiving end. Also, the future deep space links are migrating toward the higher frequency links like the Ka band and optical communication links that are susceptible to non Gaussian and non linear effects such as turbulence, scintillation, antenna pointing, jitter, etc. All these point to the fact that the assumption of a constant SNR in the analysis and design of an ARQ system might not be valid for deep space communications. In this section, using similar techniques developed in Sect. 6.4.1.3, we incorporate the effect of changing SNR, or link uncertainty, in the analysis of ARQ links. SNR is no longer considered as a fixed value, but a random variable whose long-term statistics can be characterized with a probability distribution function. We consider two limiting cases: 1. “Fast-varying” SNR: when SNRs in subsequent re-transmissions of a codeblock can assume different values, and they are independent. One example is the deep space link when the ARQ acknowledgement time is much larger than the coherency time of the channel. For communications between Earth’s ground stations and spacecraft at Mars, the round-trip light time is 20–40 min, and this is much more than the typical atmospheric coherency time of Ka-band. 2. “Slow-varying” SNR: when SNR values in subsequent re-transmissions of a code-block remain the same. One example is the proximity link between a low-Mars-orbit orbiter and a surface asset at Mars. In this case, the ARQ acknowledgement time is of the order of milliseconds and we can assume identical channel environment in subsequent re-transmissions. We show in later sections that in the case of “slow-varying” SNR, an ARQ link is always inferior than the “sent-once” link, in terms of error rate versus effective SNR. So it is not worthwhile to perform ARQ operations for “slow-varying” SNR scenarios. From this point on, we will only consider the “fast-varying” SNR case.

6 End-to-End Space System: Engineering Considerations

155

In the case of the “slow-varying” SNR scenario, if one can tolerate the latency a practical approach to improve link performance is to increase the gap time between re-transmissions, thus forcing the re-transmission SNR’s to be independent. We expect the ARQ behavior of real-world dynamic channels would fall in between the “fast-varying” and “sent-once” cases. The rest of 6.4 derives the statistical link performance and estimated latency of the truncated ARQ link, and provides some interesting insights.   We illustrate the aforementioned analysis using the NASA . 1024, 12 lowdensity-parity check (LDPC) code.

6.4.2.1

Summary of Prior Results in ARQ Link Analysis

In this section, we summarize key results in ARQ link analysis and latency estimation in [12] and [13]. These results form the basis for the derivations of the statistical ARQ link analysis and latency estimation techniques, which we will discuss in detail in later sections. We discussed automatic Repeat-reQuest (ARQ) link analysis and planning in terms of effective data rate, effective throughput, latency, and frame-error-rate (FER), under the assumption that the SNR remains the same throughout the ARQ communication session. The analytic expressions for ARQ link performance and latency are given below. ARQ Link Performance—For the Selective Repeat ARQ protocol13 and for a lossless acknowledgement channel, the effective data rate .Reff as derived in [12] is Reff = Rb (1 − Pbk )

.

(6.7)

Note that this expression is the same as the effective data rate .Reff of the “sendonce” link in Eq. (6.6), where .Rb and .Pbk are as previously defined. The difference is that the un-decodable data are lost in the case of “sent-once” link, whereas the erroneous data are eventually recovered in the case of ARQ link. Let’s consider the general case of a constant non-zero acknowledgement channel frame error rate .Pack , and the use of the Go-Back-N protocol;14 then, the effective data rate .Reff is given by:

N (1 − (1 − Pbk ) (1 − Pack )) −1 Reff = Rb 1 + (1 − Pbk ) (1 − Pack )

.

13 Selective

Repeat ARQ protocol sends one code-block per re-transmission.  2. Also, .N = 1 corresponds to the Selective Repeat protocol.

14 Typically, .N

(6.8)

156

K.-M. Cheung

 Now we express .Pbk in terms of .f

Eb N0

 , where .f (·) denotes the frame error rate

performance curve used in the return link, and Eb . N0

Eb N0

.

PD N0

is the information bit signal-

Also, = − 10log10 Rb , where . PND0 denotes the total data to-noise channel power signal-to-noise ratio (in dB), and can be computed from standard link analysis. Assuming .Rb is a tunable parameter, we can express .Reff as a function of raw data rate .Rb as follows: ratio.15

     ⎞−1 N 1 − 1 − f PND0 − 10log10 Rb (1 − Pack ) ⎠    = Rb ⎝1 + 1 − f PND0 − 10log10 Rb (1 − Pack ) ⎛

Reff

.

(6.9)

  Eb If we define effective signal-to-noise ratio . N 0

to be the energy per reliable   Eb information bit-to-noise spectral density, one can express . N as a function of 0 eff   Eb raw signal-to-noise ratio . N as follows: 0 eff

 ⎞−1     Eb N 1 − 1 − f − P ) (1 ack N0 Eb Eb ⎠ (dB)    = + 10log10 ⎝1 + . Eb N0 eff N0 1 − f N0 (1 − Pack ) (6.10)       Eb Eb Eb  N0 , and there exists a minimum . N0 that achieves Note that . N0 eff eff   Eb lossless communication within the range of . N . 0 ARQ Latency—For latency, we assume that when either or both of the codeblock and acknowledgement messages are in error, the transmitter would wait for a predetermined time .Tout before re-transmitting the code-block. For a welldesigned ARQ system, .Tout  2Tc + R , where .Tc denotes the one-way-light-time and .R denotes the receiver processing time to determine if the code-block is correctly decoded and to send an acknowledgement. There can be different ways to respond to missing acknowledgement messages and to those that are received and not decodable, resulting in different latency respond times to re-transmit. To simplify the problem, we assume that the transmitter always re-transmits after time .Tout if it does not receive an acknowledgement message, or if it receives an un-decodable acknowledgement message. The code-block transmission timeline, the acknowledgement message receiving timeline, and the processing latencies are shown in Fig. 6.7. Also assuming that .Pbk and .Pack do not change during the ARQ communication session, the latency of an ARQ link follows the discrete geometric distribution:

15 . Eb N0



⎛

is expressed in dB, not taking into account re-transmission energy required for reliable communications.

6 End-to-End Space System: Engineering Considerations

157

Fig. 6.7 ARQ transmission and receiving timeline

P rob [latency = Tc + iTout ] = θ (1 − θ )i f or i = 0, 1, 2, · · · ,

.

(6.11)

where .θ = (1 − Pbk ) (1 − Pack ) is the probability that the code-block is successfully sent and acknowledged. The mean latency as observed by the receiver is 16 Thus in an average sense, the additional latency computed to be .Tc + Tout 1−θ θ . cost of an ARQ link compared to a “send-once” link is .Tout 1−θ θ . 6.4.2.2

Statistical ARQ Link Analysis with Unlimited Number of Re-Transmissions

Due to link uncertainty,17 concepts like effective signal-to-noise ratio and latency in the ARQ link shown in Sect. 6.4.1.1 should be reevaluated statistically by incorporating the idea of “minimum margin” as defined in Sect. 6.4.1.3. In this section, statistical ARQ link analysis expressions for “fast-varying” case are derived. We will compare the performance of a coded ARQ link system using the low-density  1 parity check (LDPC) . 1024, 2 code with the “sent-once” link under a high linkuncertainty condition with .σ = 1.5, which is typical for a Ka-band link [3].18 The conventional coding performance curve (assuming SNR is constant), and the linkadjusted coding performance curve is shown in Fig. 6.8 and Table 6.1 below.

1−θ 16 Variance=.T out θ . 17 The results of this section are primarily excerpted from the authors’ paper “Statistical ARQ Link

Analysis and Planning for Dynamic Links,” IEEE Aerospace Conference 2016, Big Sky, Montana. 18 Though Ka-band links can have non-Gaussian components, we assume Gaussian distribution to simplify the discussion.

158

K.-M. Cheung

  Fig. 6.8 LDPC . 1024, 12 coding performance Table 6.1 Coding performance  for LDPC  1 . 1024, 2 code—FER versus E . b

N0

Coding performance type Constant SNR Constant SNR Link adjusted (.σ =1.5) Link adjusted (.σ =1.5)

FER 10.−3 10.−5 10.−3 10.−5

.Eb /N0

(dB)

1.62 1.94 5.54 7.33

“Fast-varying” alludes to the situation when ARQ acknowledgement time is much longer than the typical atmospheric coherency time of the channel, such as for a link between a ground station on Earth and a spacecraft at Mars. In this case, the round-trip-light-time between Earth and Mars is 20–40 min long, and the channel conditions between subsequent retransmission are assumed   to be independent. Eb As discussed in Sect. 6.4.1.1, the effective SNR . N assuming a constant 0 eff

Eb Eb is given by Eq. (6.5). If . N is fast-varying based on the definition given SNR . N 0 0  Eb in Sect. 6.4.1, the frame error rates .f N during transmission and re-transmission 0 Eb of a code-block are independent of each other. In this paper, we assume . N has 0   Eb a Gaussian distribution .h N0 | m, σ , where m is the designed SNR operation point. Using a similar argument on counting average   energy required for reliable Eb could be shown to be transmission of a code-block, the effective SNR . N 0 eff,f ast

6 End-to-End Space System: Engineering Considerations

159

Fig. 6.9 ARQ link performance for “Fast-varying” SNR case Table 6.2 Lowest effective SNR for ARQ protocols for “fast-varying” case Deterministi/varying Deterministic Deterministic Deterministic Deterministic Fast-varying Fast-varying Fast-varying Fast-varying

.

Eb N0

eff,f ast

Protocol type Selective repeat Go-back-2 Go-back-8 Go-back-32 Selective repeat Go-back-2 Go-back-8 Go-back-32

Effective SNR (dB) 1.428 1.5 1.628 1.741 3.099 3.669 4.626 5.402

Raw SNR (dB) 1.32 1.4 1.541 1.66 2.09 2.78 3.963 4.84

   ⎞ ⎛ Eb N 1 − e¯ N , σ − P (1 ) ac Eb 0    ⎠ in dB = + 10log10 ⎝1 +  Eb N0 1 − e¯ N0 , σ (1 − Pack ) (6.12)

We  applythe above analysis to the case of a coded ARQ system using the LDPC 1 . 1024, 2 code operating under a dynamic link environment typical of a Ka-band link, with .σ = 1.5. We  also  assume a lossless acknowledgement link with .Pack = 0. Eb The effective SNR . N0 for the ARQ protocols of Selective Repeat, Go-Backeff

2, and Go-Back-8, and Go-Back-32, in both deterministic SNR  and  in fast-varying Eb SNR cases are given in Fig. 6.9. The lowest effective SNR . N0 for the four eff

ARQ protocols considered in both deterministic SNR and “fast-varying” SNR cases are tabulated in Table 6.2. In conclusion, in this section, we incorporate the effect of changing SNR, or link uncertainty, in the analysis of ARQ links with no limit on number of re-transmissions. We derive analytical expressions of effective SNR and latency for the two cases, and apply the above analysis to the case of a coded ARQ system using the LDPC

160

K.-M. Cheung

 1024, 12 code operating under a dynamic link environment typical of a Kaband link, with .σ = 1.5. We observe the following interesting and insightful characteristics:

 .

1. Using the above example, we illustrate that an ARQ scheme achieves lossless performance with an effective SNR that is significantly lower than a “send-once” link with non-zero error rate. The gain comes in the expense of additional latency. In this case, the “send-once” link achieves FERs of .10−3 and .10−5 with SNR of 5.54 dB and 7.33 dB, respectively. Whereas an ARQ system using a Selected Repeat protocol achieves lossless communication with an effective SNR of 3.10 dB. 2. In the “fast-varying” SNR case that is typical of a deep space channel, the Selective Repeat re-transmission protocol exhibits significantly better SNR performance compared to Go-Back-N, particularly at low SNR regime. This shows that Selective Repeat protocol is suitable for a deep space link. 3. In the “fast-varying” SNR cases, the effective SNR changes slowly in the vicinity of the minimum as compared to the constant SNR case. This alleviates the need to estimate SNR accurately for the efficient operation of the ARQ system.

6.4.2.3

Statistical ARQ Link Analysis of Truncated ARQ with K Re-Transmissions

In this section,19 we consider the more practical case of a truncated ARQ scheme, where there is a limit on the number of retransmissions and therefore a non-zero error probability. We derive the error probability, the optimal SNR setting, and the latency statistics of the correctly received frames of the truncated ARQ schemes. As in the previous section we discuss the truncated ARQ link analysis principles   using the Gaussian assumption for SNR distribution, and using the . 1024, 12 Low Density Parity-Check (LDPC) code. We also consider the limiting cases of “fastvarying” SNR only, and show that the performance of ARQ link for “slow-varying” SNR is always inferior than the “sent-once” link. We then discuss the general case when there is statistical dependency of SNR’s between re-transmissions. Performance of truncated ARQ with K re-transmissions—As in the previous section, we consider the “fast-varying” case when SNRs in subsequent re-transmissions of a code-block are independent. For the truncated ARQ coded system there is a maximum of K retransmissions. We assume that the acknowledgement channel is error-free. An un-correctable error is made only when a codeword is failed to be delivered in all .K + 1 transmission and re-transmission attempts. Since the SNR’s K (x, σ ) as a function of raw SNR x (. Eb are independent, the frame error rate .Ffast N0 in dB) and variance (.σ in dB) is given by (the derivation is similar to Appendix in [13]): 19 The results of this section are primarily excerpted from the author’s paper “Truncated ARQ Statistical Link Analysis for Dynamic Links,” SpaceOps 2021, Cape Town, South Africa.

6 End-to-End Space System: Engineering Considerations

161

K Ffast (x, σ ) = e(x, ¯ σ )K+1

(6.13)

.

Next, we evaluate the effective SNR for the “fast-varying” SNR case, namely  K Eb . (x, σ ). As in [12] and [13], we define effective SNR to be the average No f ast

energy per information bit-to-noise spectral density, taking into account the additional energy dissipated during re-transmissions. Using a similar argument as  in the Eb derivations of (3) and (4), the effective SNR as a function of raw SNR x . N and 0 variance .σ 2 is given by: .

Eb No

K

(x, σ ) = x + 10 log10

fast

1 − e(x, ¯ σ )K+1 1 − e(x, ¯ σ)

(in dB)

(6.14)

We plot frame error rate as a function of effective SNR for the “fast-varying” SNR 20 and for K=3,7,and 15, and overlay the results with and “slow-varying” SNR   cases,   Eb Eb , σ as shown the ideal FER curve .f N0 and the link-adjusted FER curve .e¯ N 0 in Fig. 6.10. We observe the interesting trend that as SNR increases, the error rate performance of the truncated ARQ with “fast-varying” SNR (light blue curve) approaches that of the ideal “sent-once” link with constant SNR (dashed orange), and can be better or worse. Also, the light blue curve reaches an optimal point where the effective SNR is minimum, and starts to diverge from the dashed orange curve at some small FER, depending on K. The error rate performance curve of the truncated ARQ with “slow-varying” SNR (dark blue) is always worse than the worst-case SNR-adjusted “sent-once” link with varying SNR (light green). We do not have a good explanation on this observation, but it reveals the important implication that there is no reason to use ARQ protocol for channels with slow-varying SNR. If data latency can be tolerated, a smart strategy is to delay the re-transmission beyond the channel coherency time, thus artificially converting the channel to “fast-varying” and achieving close to ideal error rate performance. Data latency statistics of truncated ARQ with K re-transmissions—For latency, we assume that when either or both of the code-block and acknowledgement messages are in error, the transmitter would wait for a predetermined time .Tout before re-transmitting the code-block. We use the same notations and timeline as described in Sect. 6.4.2.1 (Fig. 6.7). Since the number of re-transmissions is limited to K, there is a non-zero probability that a codeword would be lost, that is, with latency equals to infinity (.∞). As such, latency statistics only make sense for the codewords that are successfully transmitted and received, and they take on discrete values of .Tc , .Tc + Tout , .Tc + 2Tout , · · · , Tc + KTout . A codeword that is successfully received at the .i t h re-transmission means that it has failed at the first transmission and the subsequent .i − 1 re-transmissions, and then will

20 The

derivations for “slow-varying” SNR case are not shown here.

162

K.-M. Cheung

Fig. 6.10 FER performance for truncated ARQ scheme (a) K = 3, (b) K = 7, (c) K = 15

achieve success at the .i t h re-transmission. The discrete conditional probability K .P fast ( latency = Tc + iTout | x; σ ) for the “fast-varying” SNR case is therefore K Pfast ( latency = Tc + iTout | x; σ ) =

.

1 − e(x, ¯ σ) e(x, ¯ σ )i 1 − e(x, ¯ σ )K+1

(6.15)

The mean latency .L¯ K fast (x, σ ) can be evaluated to be: K ¯ .Lfast (x, σ ) = Tc + K +

1+K 1 − 1 − e(x, ¯ σ ) 1 − e(x, ¯ σ )K+1

Tout

(6.16)

6 End-to-End Space System: Engineering Considerations

163

and the latency variance for the “fast-varying” SNR case can be shown to be:   δ + δ 2K+3 − δ K+1 (K + 1)2 − 2K(K + 2)δ + (K + 1)2 δ 2  2  2 1 − δ 2 1 − δ K+1

2 .Tout

(6.17)

where .δ = e¯ (x, σ ). For the ideal case of constant SNR, the ideal mean latency .L¯ ideal (x) is given by: L¯ ideal (x) = Tc + K +

.

 K Eb The effective SNR . N 0

ideal

.

Eb N0

1+K 1 − 1 − f (x) 1 − f (x)K+1

Tout

(6.18)

(in dB)

(6.19)

(x) can be expressed as:

K

(x) = x + 10 log10

ideal



1 − f (x)K+1 1 − f (x)



We plot the mean latency as a function of effective SNR for the “fast-varying” SNR case, and for K=3,7,and 15, and overlay with the ideal mean latency as shown in Fig. 6.11.21 Note that from (6.8) and (6.9) when the truncated ARQ link is operating at a reasonably high SNR, and/or with a large K, .L¯ K fast (x, σ ) approaches e(x,σ ¯ ) e(x,σ ¯ ) 2 T , and the latency variance approaches . T .Tc + 2 . Both out (1−e(x,σ 1−e(x,σ ¯ ) out ¯  K)) Eb are independent of K. Similarly, from (6.14), the effective SNR . N (x, σ ) 0 fast

Eb when operating at high SNR and/or for large K. approaches the raw SNR . N 0 This leads to the interesting conclusion that larger K does not increase the mean latency and its variance for the successfully transmitted and received codewords when operating in the high SNR regime. On the other hand, the improvement for error rate performance with larger K is exponential (see (6.13)). The general case of truncated ARQ with K re-transmissions—In this section, we consider the general case of a truncated ARQ system with a maximum of K retransmissions, in between the “fast-varying” SNR and “slow-varying” SNR cases where the SNR’s time-dependent statistics need to be taken into account. As in Sect. 6.2, we assume an SNR design value x and a large and constant variance .σ 2 , where x and .σ are in units of dB. For the general case of a time-varying fading channel we consider that the codeword is first sent at time .t0 with SNR .x0 , and re-sent at times .t1 , t2 , · · · , tK , with SNR’s .x1 , x2 , · · · , xK respectively, and that this satisfies the time order constraint .t1 < t2 < · · · < tK . The general form of joint SNR distribution for i re-transmissions is denoted by .hi (x0 , t0 ; x1 , t1 ; · · · , xi , ti | x; σ ), K (x, σ )22 is therefore given by: where .0 ≤ i ≤ K. The frame-error-rate .FXX

can be any value less than . 12 of .Tout . But for the ease of comparison using the Log scale, we set .Tc = 0. 22 The subscript XX denotes the random process .X(t) and its time-series correlations. 21 .T c

164

K.-M. Cheung

Fig. 6.11 Mean latency performance for truncated ARQ

 K FXX (x, σ ) =

+∞

.

−∞

f (x0 ) f (x1 ) , · · · f (xK ) hK (x0 , t0 ; · · · , xK , tK | x; σ ) dx0 dx1 · · · dxK

(6.20)

The probability of the initial transmission of a codeword is one. The discrete probability of the i re-transmissions is denoted by .GiXX (x, σ )) for .1 ≤ i ≤ K,  GiXX (x, σ ) =

+∞

.

−∞

f (x1 ) , · · · f (i)hi (x0 , to ; · · · , xi , ti | x; σ ) dx0 dx1 · · · dxi (6.21)

and the SNR after K re-transmissions is   K K  Eb i (x, σ ) = x + 10 log10 1 + GXX (x, σ ) ( in dB) . No XX

(6.22)

i=1

As in the discussion for truncated ARQ scheme, there is a non-zero probability that a codeword would fail in all .K + 1 transmission and re-transmission attempts. Thus, latency only makes sense in the case when the codeword is successfully transmitted and received. We use the re-transmission operation model as described in Fig. 6.7. Without loss of generality, we assume the transmission start is at time 0, and the codeword is first received at time .t0 = Tc . When there are .i t h re-transmissions, .1 ≤ i ≤ K, the codeword receiving time is .ti = Tc + iTout . The probability that the codeword is successfully transmitted and received in the first trial is  0 PXX ( latency = Tc | x; σ ) = 1 −

+∞

.

−∞

f (xo ) h0 (x0 , Tc | x; σ ) dx0

(6.23)

6 End-to-End Space System: Engineering Considerations

165

The probability that the codeword is successfully transmitted and received in the .i t h re-transmission, .1 ≤ i ≤ K, is i PXX ( latency = Tc + iTout | x; σ )    +∞ = f (x0 )f (x1 ), · · · f (xi−1 ) (1 − f (xi )) hi

.

−∞

(x0 , Tc ; · · · ; xi , Tc + iTout | x, σ ) dx0 dx1 · · · dxi

(6.24)

0 (·) and .P i (·)’s are independent events, the probability that a codeword is As .PXX XX successfully transmitted and received is therefore:

K PSuccess (x, σ ) =

K 

.

i PXX ( latency = Tc + iTout | x; σ )

(6.25)

i=0

So, given the codeword is successfully transmitted and received, the conditional probability that this event occurs at the .i t h trail, where .1 ≤ i ≤ K, is normalized as follows: i PXX latency = Tc + iTout | x; σ ; success

.

=

i PXX ( latency y = Tc + iTout | x; σ )

(6.26)

K PSuccess (x, σ )

The mean latency .L¯ K XX (x, σ ), given the codeword is successfully transmitted and received, is therefore: ¯K .L XX (x, σ )

= Tc +

K 

 i iPXX

( latency = Tc + iTout | x; σ ; success ) Tout

i=1

(6.27) The above general case of time-varying SNR is not easy to characterize as it is practically very difficult to estimate the time-dependent joint SNR distribution function in a channel. To simplify the analysis, we assume the SNR random process .X(t) to be Wide-Sense-Stationary (WSS), that is, its mean and its autocorrelation functions are time-invariant: E(X(t)) = x for all t

.

RXX (t1 , t2 ) = E (X (t1 ) X (t2 )) = RXX (τ ) where τ = t2 − t1

.

(6.28) (6.29)

From (6.28) and (6.29), it can be shown that the variance is also time-invariant,

166

K.-M. Cheung

  E (X(t) − x)2 = σ 2 for all t

.

(6.30)

and the autocovariance coefficient .ρXX (τ ) is defined to be: ρXX (τ ) =

.

RXX (τ ) − x 2 for all τ σ2

(6.31)

For the discrete time SNR random process .X(t0 ), X(t1 ), · · · , X(tK ) (similar to the analysis in the previous section), we assume that the random vector .x¯i = (x0 , x1 , · · · , xi )T , .1 ≤ i ≤ K, has a multivariate Gaussian distribution given by: hi (x0 , x1 , · · · , xi | x; σ ) = 

.

1 (2π )i+1 |i |

e

  − 12 (x¯i −x) ¯ T i−1 (x¯i −x) ¯

(6.32)

where .x¯ = (x0 , x1 , · · · , x)T is a .(i + 1)-tuple of x, and .i is a .(i + 1) × (i + 1) covariance matrix. Using the re-transmission operation model as described in Fig. 6.3, where re-transmissions occur at interval of .To ut, .i is generated as: i = σ 2 ⎡

.

⎤ ρXX (2Tout ) ... ρXX (iTout ) ρXX (Tout ) ρXX (Tout ) ρXX ((i − 1)Tout ) ⎥ ⎥ ⎥ .. .. ⎥ . ρXX (Tout ) ... . ⎥ ⎥ .. .. ⎦ . . 1 ρXX (Tout ) ρXX (iTout ) ρXX ((i − 1)Tout ) ... ρXX (Tout ) 1 (6.33)

1 ⎢ ρXX (Tout ) ⎢ ⎢ ⎢ ρXX (2Tout ) ⎢ ⎢ .. ⎣ .

ρXX (Tout ) 1

Substituting (6.32) and (6.33) into (6.20)–(6.27), the frame-error-rate, the effective SNR, and the latency statistics can be computed for a given .ρX X(τ ). .ρX X(τ ) can be estimated using time-series SNR measurements from the channel. For example, [14] described the experiment setup and the operation procedure to construct the autocovariance coefficient of SNR for a deep space Ka-band link (see Figure 9 of [14]). In this deep space Ka-band channel, the fading is caused by macro-scale weather effects. Figure 9 of [14] indicates that codewords separated by 30 s are practically uncorrelated, and this relates to the “fast-varying” SNR case. Autocorrelation functions of interferometric phase data for the Goldstone site consistently show time scales 20 s–25 s [15]. We consider a latency constrained scenario, where the codewords can only be re-transmitted at 5-s intervals (.To ut .= 5 s) up to 3 times, i.e. .K = 3. The initial codeword and the subsequent re-transmitted codewords are therefore correlated, with .ρXX (Tout = 5 s) = 0.4, .ρXX (2Tout = 10 s) = 0.2, and .ρXX (3Tout = 15 s) = 0.1 respectively.

6 End-to-End Space System: Engineering Considerations

167

Fig. 6.12 Comparison of FER versus effective SNR

Fig. 6.13 Comparison of mean latency versus effective SNR

The frame-error-rate as a function of effective SNR (light-green) in the correlated channel, when overlaid onto the .K = 3 case of Fig. 6.10, is given in Fig. 6.12. The penalty of re-transmitting the codewords at 5-s intervals (instead of 30-s intervals) is about 1.25 dB. The mean latency as a function of effective SNR for the Ka-band correlated channel case and .K = 3, is overlaid with the ideal mean latency and with the “fast-varying” SNR mean latency (from Fig. 6.11), and they are shown in Fig. 6.13. Note that the mean latency for the Ka-band correlated channel approaches that of the “fast-varying” SNR case when the SNR is high.

6.5 Concluding Remarks In this chapter, we discuss various statistical effects like data generation and changing channel SNR, and the system dependencies, in the end-to-end flight-

168

K.-M. Cheung

ground optimization of an ARQ reliable communication link. We illustrate the concept using deep space links, and assume block fading scenarios when the channel coherence time of link fluctuation is comparable or larger than the time duration of a codeword length such that the time-fluctuation of the link does not get randomized or averaged out within a codeword,23 and no interleaving schemes are used. For the cases with coded-interleaving systems, the readers are referred to [16]. The discussion in this chapter is high-level, and the intent is to use link analysis models which are mathematical abstractions of communication link and network functions, and statistical techniques to do flight-ground architecture trades without resorting to details of signal formats and data processing functions. Our aim is to provide readers with modern, practical, and systematic techniques to effectively iterate and to evolve the flight-ground architecture from its conception, through design and development, to operation and disposal. Among the many long-standing challenges of system design are as follows: • “Inverted” Problem—Many common engineering problems involve answering the question “given a design, what is the performance?” System architecture studies for early design, by contrast, sometimes require a solution to the much broader, reversed, and challenging problem: “given the high-level design requirements and operational constraints, is there a solution that is optimal among those that meet the requirements and constraints? Is it unique? Is it a local optimum or a global one?” • Interactions Between Design Components—In the early design phase of a complex system development it is difficult to detect and to characterize the subtle operation behaviors and idiosyncrasies due to interactions between components. The unexpected operation behaviors and idiosyncrasies can lead to major re-work in later phases. • Lack of Definitive Information—Architecture studies usually address system designs planned for far in the future, and are usually characterized by the lack of definitive information. For example, one might have to make assumptions on future market trends, technology readiness, and product availability, which can be difficult to predict and are based on educated guesses and speculations of the “experts”. As such, human bias and heuristic thinking often percolate into the design process, and it would be difficult to quantify the likelihood (design risk) that the final design would not meet the original design requirements and constraints. This problem was discussed in further details in [17] and [18], but this is beyond the scope of this chapter. • Evolving Design—Design is always evolving. One challenge of system architecting is to ensure that the system design progresses on the right path, and does not march into a solution that is either too conservative or too risky in the

23 The current Consultative Committee for Space Data Systems (CCSDS) standards do not provide the option of including interleaving in RF communication system. This is due to the fact that, historically, the effects of fading have been insignificant for lower frequency bands like S-band and X-band.

6 End-to-End Space System: Engineering Considerations

169

time-evolving design process, given that re-design in the late phase is always an expensive alternative. We hope that this chapter would provide the necessary tools to tackle the aforementioned challenges in the design and develop of modern-day communications and network architecture. Acknowledgments The author would like to thank Mark Sanchez-Net, Charles Lee, Thomas Choi, and David Morabito for their discussion and contributions to this chapter. The research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (80NM0018D0004).

References 1. J. Yuen, Deep Space Telecommunication Systems Engineering (Plenum Press, New York, 1983) 2. K.M. Cheung, Statistical link analysis – a risk analysis perspective. Interplanetary Netw. Progress Rep. 42, 183 (2010) 3. K.M. Cheung, Statistical link analysis for Ka-band links, in Proceedings of SpaceOps 2014 (Jet Propulsion Laboratory, Pasadena, 2014). 4. K.M. Cheung, M. Belongie, K. Tong, End-to-end system consideration of the galileo image compressio system. JPL TDA Progress Report 42–126 (1996) 5. S. Shambayati et al., MRO Ka-band demonstration: cruise phase lessons learned, in IEEE Aerospace Conference 2007 (Big Sky, Montana, 2007) 6. G. Fayolle, E. Gelenbe, G. Pujolle, An analytic evaluation of the ‘send and wait’. IEEE Trans. Commun. COM-26 (1978) 7. D. Towsley, J. Wolf, On the statistical analysis queue lengths and waiting times for statistical multiplexer with ARQ retransmission scheme. IEEE Trans. Commun. COM-27 (1979) 8. D. Towsley, A statistical analysis of ARQ protocols operating in a nonindependent error environment. IEEE Trans. Commun. COM-29 (1981) 9. M. Zorzi, R. Rao, Throughput analysis of go-back-N ARQ in Markov channels with unreliable feedback. Proc. IEEE (1995) 10. M. Zorzi, R. Rao, On the use of renewal theory in the analysis of ARQ protocols. IEEE Trans. Commun. COM-44 (1996) 11. M. Zorzi, R. Rao. Performance of ARQ go-back-N protocol in Markov channels with unreliable feedback. Mobile Netw. Appl. 2, 183–193 (1997) 12. K. Cheung, Problem formulation and analysis of the 1-Hop ARQ links. Interplanetary Netw. Progress Rep. 42–194, 1–15 (2013) 13. K. Cheung, C. Lau, C. Lee, Link analysis for space communication links using ARQ protocol, in IEEE Aerospace Conference 2014 (Big Sky, Montana, 2014) 14. D.D. Morabito, Deep-space Ka-band flight experience. Interplanetary Netw. Progress Rep. 42211 (2017). https://ipnpr.jpl.nasa.gov/progressreport/42-211/211B.pdf 15. D.D. Morabito, L. D’Addario, S. Finley, A comparison of atmospheric effects on differential phase for a two-element antenna array and nearby site test interferometer. Radio Sci. 51 (2016). https://doi.org/10.1002/2015RS005763 16. K. Cheung, D. Morabito, M. Sanchez-Net, Design and modeling of a coded-interleaving system in the presence of fading, in IEEE Aerospace Conference 2021 (Big Sky, Montana, 2021) 17. A. Babuscia, K.M. Cheung, Statistical risk estimation for communication system design. IEEE J. Syst. 7(1), 125–136 (2012) 18. A. Babuscia, K.M. Cheung, An approach to perform expert elicitation for engineering design risk analysis: methodology and experimental results. J. R. Stat. Soc. A, 177(2), 475–497 (2014)

Chapter 7

Intelligent Space Communication Networks Mario Marchese, Simone Morosi, and Fabio Patrone

7.1 Introduction Different fields are currently benefiting from the introduction of more intelligent solutions in space systems and devices. In most cases, this means the use of Artificial Intelligence (AI) based techniques to solve problems already addressed by previous solutions, but in a more efficient and effective way, and tackle issues not addressable yet due to the limitations of the available solutions. AI techniques are currently under study, development, and deployment in a huge plethora of scenarios and applications for improving a high number of services in terms of different performance indicators. In communication networks, including satellite communication networks, AIbased solutions may differ in different aspects: • Scope: AI techniques can be applied for different purposes related to the offered applications, i.e. to improve the quality of the user’s exploited applications, or to the offered connectivity service, i.e., to improve the quality of the user’s experienced connectivity and/or the number of users that can join the network. • Algorithm: AI solutions can be based on different Machine Learning algorithms that can be categorised in different subsets depending on how they perform the training phase, such as supervised, unsupervised, and semi-supervised learning,

M. Marchese () · F. Patrone Dipartimento di ingegneria navale, elettrica, elettronica e delle telecomunicazioni (DITEN), University of Genova, Genova, Italy e-mail: [email protected]; [email protected] S. Morosi University of Florence, Florence, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_7

171

172

M. Marchese et al.

or their basic principles, such as Deep Learning (DL) and Reinforcement Learning (RL). • Input information: AI techniques may use typically big datasets as input in the training phase before being able to perform well in the operational phase. These sets of information may refer to single or multiple variables analysed over time. If the analysed data are reported in terms of multiple and heterogeneous variables, correlating them is another task that AI solutions may perform differently. • Implementation location: AI algorithms can run on different nodes of the network depending on their scope, the needed input information, and the nodes’ available data processing and storing capabilities. For example, moving AI capabilities to the edge of the network, i.e., closer to the users, may bring some advantages, especially in terms of users’ perceived service, but it needs a higher control overhead and nodes with enough available resources. • Improved performance: different AI solutions can be the best solutions depending on the problem to solve and the performance we aim to improve. For example, some algorithms may offer better performance in terms of accuracy and reliability but suffer from high computational delays. This is suitable for applications with high-reliability requirements but may be intolerable for delay-sensitive ones, which instead prefer much faster decision times and tolerate higher error rates.

7.2 AI Improvements in Satellite Networks Space communication networks are benefiting from the employment of AI techniques in multiple ways [1]. Multiple issues can be addressed by AI-based solutions, leading to several improvements compared to the previously available techniques.

7.2.1 Communication Resource Allocation Communication resources are typically limited and have to be properly managed in order to, on one hand, satisfy each user’s QoS requirements and, on the other hand, avoid waste. This is even more prominent in satellite communication networks due to the stronger resource limitations compared to terrestrial networks. AI techniques can help address several sub-aspects: • Network traffic prediction: being able to predict network traffic evolution over time is an advantage that can improve multiple processes, such as congestion control, routing, and handover. Concerning routing, for example, currently designed and deployed satellite communication networks show an increasing trend in terms of the number of satellites to offer connectivity to a high number of possible interested users spread in wider, potentially worldwide, areas. Most

7 Intelligent Space Communication Networks

173

of them are composed of LEO satellites which are deployed on multiple orbital planes at different altitudes. Hierarchical solutions are also envisioned where LEO, MEO, and GEO satellites have different roles and cooperate to offer Internet connectivity and exchange data among them through RF or optical InterSatellite Links (ISL). The routing process and the consequent resource allocation are of primary importance due to the high availability and variability of end-toend paths and the numerous parameters that can be taken into account related to both the user performance requirements and network (satellites, ISLs, satelliteground links, . . . ). Traffic forecasting techniques not based on AI solutions suffer from two major difficulties: the limited onboard computational resources and the Long-RangeDependence (LRD) of satellite network traffic that makes lower complexity Short-Range-Dependence (SRD) models to fail achieving accurate forecasting. For these reasons, AI solutions have been proposed to further optimize this task. Some examples are a high-accuracy traffic forecasting method with lower training time which applies Principal Component Analysis (PCA) and then a generalized regression NN [2], an Extreme Learning Machine (ELM)-based technique employed for traffic load forecasting of satellite nodes before routing [3], and a method based on Fly Optimization Algorithm—Extreme Learning Machine (FOA-ELM) which uses the Empirical Mode Decomposition (EMD) to decompose the traffic of the satellite with LRD into a series with SRD to decrease the predicting complexity and improve the prediction speed [4]. • Channel model: the features of satellite channels may differ depending on multiple parameters, such as the satellite altitude, and may change over time. Considering a channel model as close as possible to the real scenario, it is useful to have a very clear idea about the transmission and interference conditions and so usefully allocate the available communication resources. However, the creation of precise satellite channel models and the consequent estimation of the channel parameters is a challenge due to the multiple factors to take into account. Even if several techniques are already consolidated with satisfactory performance, such as ray tracing, they all suffer from many limitations, such as the need for a huge amount of information that may not be available and the high required computational effort that are in contrast with the real-time optimization needs. AI-based solutions have proven to be effective in overcoming these limitations. Some examples are solutions based on more traditional ML techniques, such as NN [5], and more sophisticated DL-based solutions [6] aim to forecast packet losses, an aspect related to the channel modelling for optimal resource allocation. • Signal Detection: As each signal must be separated before classification, modulation, demodulation, decoding, and other signal processing, localization and detection of carrier signals in the frequency domain are crucial. Most of the traditional techniques are based on single or multiple threshold values, rely on tractable mathematical models under known noise process and/or deterministic interference, and required human intervention, making the process

174

M. Marchese et al.

Fig. 7.1 DL-based signal detection and demodulation strategy [7]

more complex and the needed effort more significant in an environment full of signals to identify and differentiate. AI approaches can be efficiently used under dynamic interference to effectively detect the target signals. AI detectors can be trained for detecting various modulation and coding techniques and be based on different algorithms [7] (see Fig. 7.1). A DL-based solution is proposed in [8] to morse signals blind detection in wideband spectrum data, while a FCN model is proposed in [9] to detect carrier signal in the broadband power spectrum. • Interference management: interference is a phenomenon that strongly affects communications, in particular through satellite links. It is a common event whose effects are worsening with the increasing congestion of the satellite frequency bands due to a higher number of deployed communication satellites, active satellite network users, and expected applications. As a consequence, interference management is essential to allow high-quality and reliable communications through detection, classification, and suppression of interference, as well as minimization of its occurrence. Interference detection is a well-studied subject that has been extensively addressed also for satellite communications. Most common solutions are based on theoretical models for signal characteristics and satellite channels used to estimate interference and techniques to properly counterbalance the transmitted signals optimizing interference cancellation [10]. To further minimize interference effects, examples of the proposed AI-based solutions include a framework combining Support Vector Machine (SVM) , unsupervised learning, and DRL-based approaches for satellite selection, antenna pointing, and tracking [11], an approach to forecast the signal spectrum to be received in absence of anomaly by using LSTM trained on historical anomaly-

7 Intelligent Space Communication Networks

175

free spectra [12] and a DNN and LSTM-based method to detect and classify interference [13]. • Beam hopping: conventional satellite systems uniformly allocated resources across beams, which may lead to lower resources than needed in some beams and the resource under-utilization in other beams due to the typical not uniform geographical distribution of the users underneath. Beam hopping has emerged as a promising technique to achieve great flexibility in managing non-uniform, time and spatial variant traffic requests. It is based on a dynamic allocation of the available resources to only a subset of the overall beams depending on the current users’ traffic demands. The problem is to optimally decide when to allocate resources to a new beam and for how long. Even if this problem has been already addressed by proposing solutions not based on AI techniques, the technological evolution is leading to complications that are difficult to properly take into account with traditional methods. For example, as the number of beams increases (reaching hundreds or thousands of beams per satellite), it is becoming more difficult and time consuming to find the optimal choice rather than one of many local optima. Some of the proposed AI solutions involve the use of DRL to reduce the transmission delay and increase the system throughput [14] (see Fig. 7.2), fully-connected NNs to predict non-optimal beam hopping patterns [15], and low-complexity Multi Objective-DRL to ensure the fairness of each cell and, at the same time, improve the achievable throughput [16]. • Energy management: communication satellites typically require a high amount of energy to transmit data due to the physical nature of the space environment

Fig. 7.2 DRL-based beam hopping algorithm [14]

176

M. Marchese et al.

and the high attenuation and interference factors than a typical terrestrial environment. Satellites suffer, at the same time, from severe energy limitations, considering the Sun as the only energy source, the generated power depending on the extension and orientation of the satellite’s solar panels, and the available energy depending on the satellite’s battery capacity. Besides, the increasing number of users and the decreasing satellite size is further stressing these limitations, imposing a careful management of the satellite energy consumption to avoid service disruptions. Resource scheduling schemes, even involving temporary complete or partial shutdown of the single identified satellite communications, have been designed to dynamically adjust the data overload of each satellite distributing the energy consumption throughout the satellite segment. Examples of AI solutions include using DNN compression before data transmission to improve latency and optimize the power allocation in satellite-toground communications [17], RL to share the workload of overworked satellites with near satellites with lower load [18], and DRL to allocate communication slots with high energy efficiency [19].

7.2.2 Security More satellite communication networks will be widespread with a higher number of users, the more they will be appealing for malicious attacks, especially cyberattacks, aiming to disrupt the offered service or even damage the network apparatus. Recent solutions focus more attention on security, following the principle of security-by-design, but there is still room for improvements, especially considering the problem of improving security of the already deployed satellite systems. Also in this case, AI techniques can help address several sub-aspects: • Anti-jamming: jamming is one of the simplest but most effective attacks that can be carried on against communication networks to interfere and, in the worst case, completely disrupt the offered service, isolating the users located in the attacked area from the network or making a base station incapable of offering connectivity. Traditional solutions have been proposed to alleviate the jammer effects, such as the FHSS and DSSS strategies. However, these solutions are not able to dynamically adjust their action depending on the jammer characteristics. AI principles have already been considered to develop more sophisticated attack methods. For example, a smart jammer able to automatically adjust the jamming channel and power in order to maximize the jamming effect is proposed in [20]. This makes of primary importance the use of AI-based antijamming solutions able to automatically protect the network nodes against jamming attacks, AI-based or not. Some examples include: a hierarchical learning approach proposed in [21] to improve the frequency selection process where both jamming and co-channel interference are present; a frequency-spatial

7 Intelligent Space Communication Networks

177

2-D anti-jamming scheme to resist jamming and interference and a fast DQN based 2-D mobile communication algorithm that applies DQN, macro-actions and hotbooting techniques to achieve the optimal frequency selection described in [22]; a spatial anti-jamming scheme based on DRL to take proper data routing choices to make the network more robust against jamming attacks that can disrupt a subset of the network links presented in [23]. • Monitoring telemetry data: telemetry is the group of information describing the status of the system, especially, in our case, the satellites. They are control packets that are sent in downlink to help ground operators to monitor the satellite status, such as satellite position, satellite attitude, and solar panel orientation, and operations to be sure that they are operating within the defined limits. All these data are recorded and can be analysed to detect abnormal events and predict possible upcoming abnormal situations in order to minimize failure risks. However, how to correlate heterogeneous data coming from multiple sources of information to find correlations, recognize patterns, and so detect anomalies, may be a challenge. Simple solutions involve setting operational ranges for the monitored parameters and periodically checking their values to detect single or systematic out-of-range events. Even if simplicity is their best advantage, they suffer from severe limitations as the systems are becoming more complex with a higher number of parameters to monitor. This in turn leads to a higher volume of data to send through the satellite links with a consequent higher delay needed to process all the data and close the loop with proper reactive actions. AI-based solutions help build more sophisticated and reactive health monitoring systems by using different techniques, such as probabilistic clustering [24]. Other examples involve using linear regression to forecast short-lifetime satellite behaviours (3–5 years) and NNs for long-lifetime satellite behaviours (15–20 years) [25] and a self-learning classification algorithm able to achieve onboard telemetry data classification with low computational complexity and low time latency [26].

7.2.3 Orbital Edge Computing (OEC) Orbital Edge Computing (OEC) is a recent vision in next-generation satellites, seen as powerful nodes equipped with additional data computational and storage capabilities that can be exploited by the offered services. Process data directly onboard satellites can help several applications reduce latency compared to centralized cloud processing platforms where raw data have to be forwarded through satellites from users to the platform and processed data on the backward path. Store a significant amount of data onboard satellites can help further reduce latency avoiding that data requests and responses traverse the satellite path to reach the data repositories and vice versa. Allocation of tasks to process and data batches to store among satellites is the main problem to address related to the OEC concept [27]. The dynamic and

178

M. Marchese et al.

Fig. 7.3 RL-based computing offloading approach [28]

time-varying nature of satellite networks, such as in terms of satellite-satellite and satellite-ground user links, require a careful task and data distribution strategy that should take into account different factors, such as the network topology changes over time and estimations of the user traffic flows and data processing requests. Some examples of AI-based solutions are a joint resource allocation and task scheduling approach that aims to allocate the computing resources to virtual machines and schedule the offloaded tasks for Unmanned Aerial Vechile (UAV) edge servers, whereas an RL-based computing offloading approach handles the multidimensional network resources and learns the dynamic network conditions [28] (see Fig. 7.3), and a joint user-satellite association and task offloading decision with optimal resource allocation methodology based on DRL to improve the longterm latency and reduce the energy consumption [29]. A novel AI-based architecture for Earth Observation satellites which embeds AI DNN algorithms for consuming data at source rather than on the ground aim to minimize the downlink bandwidth usage is presented in [30].

7.2.4 Remote Sensing Multiple applications and functionalities benefit from the use of AI-based solutions. Remote Sensing is the operation of collecting and processing information about the observed areas, objects, or phenomena from their reflected and emitted radiation. Its applicability regards numerous scenarios and applications with multiple advantages, such as the possibility to remotely monitor dangerous or unreachable areas. Traditional approaches are in use since the beginning of this discipline with considerable results. However, the need to monitor more complex phenomena and the development of more precise sensors able to collect a much wider set of different

7 Intelligent Space Communication Networks

179

information from the monitored subject to analyse and correlate to take useful conclusions recently emerged, with the consequent need to have more flexible and versatile solutions. The evolution of computer vision capabilities due to DL has led to the increased development of remote sensing solutions adopting state-of-the-art DL algorithms on satellite images. An example is a combined kNN and CNN-based solution to map coral reef environments by using remote sensing images [31]. Object detection and recognition are another set of applications whose capabilities have improved thanks to AI. CNN-based object detection algorithms have been developed to recognise different kinds of objects, such as clouds [32] and ships [33].

7.2.5 Space-Air-Ground Integrated Network Space-Air-Ground Integrated Network (SAGIN) is a recent evolution of satellite communication networks is not only leading to the deployment of Mega LEO satellite constellations made of thousands of satellites, but also to hierarchical networks composed of multiple layers of space (satellites), aerial (UAVs and/or HAPs), and ground communication nodes. This is also leading to a higher number of users interested in exploiting the connectivity services of this kind of networks which were previously limited to giving telephone, tv, or Internet coverage to unserved areas. Integration with terrestrial mobile communications, such as 5G, is another aspect deeply under investigation and standardization within the 3GPP. SAGINs aim to provide users with improved and flexible end-to-end services thanks to a hierarchical network where different kinds of nodes typically have different roles but they all collaborate and exchange users’ data to offer them the required QoS. AI-based solutions can help optimize the achievable performance improving multiple aspects. For example: a CNN-based solution is proposed in [34] to optimize the network overall performance by using traffic patterns and the remaining buffer size of GEO and MEO satellites as input information (see Fig. 7.4); a DRLbased solution that jointly optimises the satellite selection and the UAV location to maximise the end-to-end data rate of the source-satellite-UAV-destination communications is presented in [35]; a low-complexity technique for computing the capacity among satellites by using a time structure based augmenting path searching method and a long-term optimal capacity assignment RL-based model to maximize the long-term utility of the system is suggested in [36].

7.2.6 Satellite Operations Potential applications of AI are also being thoroughly investigated in satellite operations [37], in particular to support the operation of large satellite constellations,

180

M. Marchese et al.

Fig. 7.4 Typical SAGIN network topology (a); flow chart of the proposed AI-based routing solutions (b) [34]

including relative positioning, communication and end-of-life management. To this aim some of the experiments that have been planned on the OPS-SAT mission [38] included artificial intelligence: as a matter of facts, to develop autonomous

7 Intelligent Space Communication Networks

181

spacecraft that use artificial intelligence to take care of themselves would be very useful for exploring new parts of the Solar System and reducing mission costs. In addition, it is becoming more common to find ML systems analysing the huge amount of data that comes from each space mission, including spacecraft telemetry and product data; another application of AI would be the analysis of all this data. It is worth stressing that data coming from some Mars rovers is being transmitted using AI: particularly, intelligent data transmission software on board rovers removes human scheduling errors which can otherwise cause valuable data to be lost. The same technology could also be used in long-term missions that will explore the Solar System, meaning that they will require minimal oversight from human controllers on Earth. Nonetheless AI also currently lacks the reliability and adaptability required in new software; these qualities will need to be improved before it takes over the space industry.

References 1. F. Fourati, M.-S. Alouini, Artificial intelligence for satellite communication: a review. Intell. Converged Netw. 2(3), 213–243 (2021) 2. L. Ziluan, L. Xin, Short-term traffic forecasting based on principal component analysis and a generalized regression neural network for satellite networks. J. Chin. Univ. Posts Telecommun. 25(1), 15–28 (2018) 3. Z. Na, Z. Pan, X. Liu, Z. Deng, Z. Gao, Q. Guo, Distributed routing strategy based on machine learning for LEO satellite network. Hindawi Wirel. Commun. Mobile Comput. 2018 (2018) 4. Y. Bie, L. Wang, Y. Tian, Z. Hu, A combined forecasting model for satellite network selfsimilar traffic. IEEE Access 7, 152004–152013 (2019) 5. E. Ostlin, H.-J. Zepernick, H. Suzuki, Macrocell path-loss prediction using artificial neural networks. IEEE Trans. Veh. Technol. 59(6), 2735–2747 (2010) 6. J. Thrane, D. Zibar, H.L. Christiansen, Model-aided deep learning method for path loss prediction in mobile communication systems at 2.6 GHz. IEEE Access 8, 7925–7936 (2020) 7. B.A. Homssi, K. Dakic, K. Wang, T. Alpcan, B. Allen, S. Kandeepan, A. Al-Hourani, W. Saad, Artificial intelligence techniques for next-generation mega satellite networks (2022). Preprint. arXiv:2207.00414 8. Y. Yuan, Z. Sun, Z. Wei, K. Jia, DeepMorse: a deep convolutional learning method for blind morse signal detection in wideband wireless spectrum. IEEE Access 7, 80577–80587 (2019) 9. H. Huang, J.-Q. Li, J. Wang, H. Wang, FCN-based carrier signal detection in broadband power spectrum. IEEE Access 8, 113042–113051 (2020) 10. C. Politis, S. Maleki, C. Tsinos, S. Chatzinotas, B. Ottersten, On-board the satellite interference detection with imperfect signal cancellation. IEEE 17th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) (2016), pp. 1–5 11. Q. Liu, J. Yang, C. Zhuang, A. Barnawi, B.A. Alzahrani, Artificial intelligence based mobile tracking and antenna pointing in satellite-terrestrial network. IEEE Access 7, 177497–177503 (2019) 12. L. Pellaco, N. Singh, J. Jaldén, Spectrum prediction and interference detection for satellite communications, in IET International Communications Satellite Systems Conference (2019), pp. 1–8

182

M. Marchese et al.

13. P. Henarejos, M.A. Vázquez, A.I. Pérez-Neira, Deep learning for experimental hybrid terrestrial and satellite interference management. IEEE 20th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) (2019), pp. 1–5 14. X. Hu, S. Liu, Y. Wang, L. Xu, Y. Zhang, C. Wang, W. Wang, Deep reinforcement learningbased beam Hopping algorithm in multibeam satellite systems. Wiley IET Commun. 13(16), 2485–2491 (2019) 15. L. Lei, E. Lagunas, Y. Yuan, M.G. Kibria, S. Chatzinotas, B. Ottersten, Beam illumination pattern design in satellite networks: learning and optimization for efficient beam hopping. IEEE Acccess 8, 136655–136667 (2020) 16. X. Hu, L. Wang, Y. Wang, S. Xu, Z. Liu, W. Wang, Dynamic beam hopping for DVB-S2X GEO satellite: a DRL-powered GA approach. IEEE Commun. Lett. 26(4), 808–812 (2022) 17. V. Kothari, E. Liberis, N.D. Lane, The final frontier: Deep learning in space, in 21st International Workshop on Mobile Computing Systems and Applications (2020), pp. 45–49 18. H. Tsuchida, Y. Kawamoto, N. Kato, K. Kaneko, S. Tani, S. Uchida, H. Aruga, Efficient power control for satellite-borne batteries using Q-learning in low-earth-orbit satellite constellations. IEEE Wirel. Commun. Lett. 9(6), 809–812 (2020) 19. B. Zhao, J. Liu, Z. Wei, I. You, A deep reinforcement learning based approach for energyefficient channel allocation in satellite internet of things. IEEE Access 8, 6219–62206 (2020) 20. C. Han, Y. Niu, Cross-layer anti-jamming scheme: a hierarchical learning approach. IEEE Access 6, 34874–34883 (2018) 21. F. Yao, L. Jia, Y. Sun, Y. Xu, S. Feng, Y. Zhu, A hierarchical learning approach to anti-jamming channel selection strategies. Springer Wirel. Netw. 25(1), 201–213 (2019) 22. L. Xiao, D. Jiang, D. Xu, H. Zhu, Y. Zhang, H.V. Poor, Two-dimensional antijamming mobile communication based on reinforcement learning. IEEE Trans. Veh. Technol. 67(10), 9499– 9512 (2018) 23. C. Han, L. Huo, X. Tong, H. Wang, X. Liu, Spatial anti-jamming scheme for internet of satellites based on the deep reinforcement learning and Stackelberg game. IEEE Trans. Veh. Technol. 69(5), 5331–5342 (2020) 24. T. Yairi, N. Takeishi, T. Oda, Y. Nakajima, N. Nishimura, N. Takata, A data-driven health monitoring method for satellite housekeeping data based on probabilistic clustering and dimensionality reduction. IEEE Trans. Aerosp. Electron. Syst. 53(3), 1384–1401 (2017) 25. S.K. Ibrahim, A. Ahmed, M.A. Zeidan, I.E. Ziedan, Machine learning methods for spacecraft telemetry mining. IEEE Trans. Aerosp. Electron. Syst. 55(4), 1816–1827 (2018) 26. P. Wan, Y. Zhan, W. Jiang, Study on the satellite telemetry data classification based on selflearning. IEEE Access 8, 2656–2669 (2019) 27. B. Zhao, J. Liu, Z. Wei, I. You, Orbital edge offloading on mega-LEO satellite constellations for equal access to computing. IEEE Commun. Mag. 60(4), 32–36 (2022) 28. N. Cheng, F. Lyu, W. Quan, C. Zhou, H. He, W. Shi, X. Shen, Space/aerial-assisted computing offloading for IoT applications: a learning-based approach. IEEE J. Sel. Areas Commun. 37(5), 1117–1129 (2019) 29. G. Cui, X. Li, L. Xu, W. Wang, Latency and energy optimization for MEC enhanced SAT-IoT networks. IEEE Access 8, 55915–55926 (2020) 30. G. Furano, G. Meoni, A. Dunne, D. Moloney, V. Ferlet-Cavrois, A. Tavoularis, J. Byrne, L. Buckley, M. Psarakis, K.-O. Voss, Towards the use of artificial intelligence on the edge in space systems: Challenges and opportunities. IEEE Aerosp. Electron. Syst. Mag. 35(12), 44– 56 (2020) 31. A.S. Li, V. Chirayath, M. Segal-Rozenhaimer, J.L. Torres-Perez, J. van den Bergh, NASA NeMO-net’s convolutional neural network: mapping marine habitats with spectrally heterogeneous remote sensing imagery. IEEE J. Sel. Top. Appl. Earth Observations Remote Sens. 13, 5115–5133 (2020) 32. G. Mateo-García, V. Laparra, D. López-Puigdollers, L. Gómez-Chova, Cross-sensor adversarial domain adaptation of Landsat-8 and Proba-V images for cloud detection. IEEE J. Sel. Top. Appl. Earth Observations Remote Sens. 14, 747–761 (2020)

7 Intelligent Space Communication Networks

183

33. F. Wang, F. Liao, H. Zhu, FPA-DNN: a forward propagation acceleration based deep neural network for ship detection, in IEEE International Joint Conference on Neural Networks (IJCNN) (2020), pp. 1–8 34. N. Kato, Z.M. Fadlullah, F. Tang, B. Mao, S. Tani, A. Okamura, J. Liu, Optimizing space-airground integrated networks by artificial intelligence. IEEE Wirel. Commun. 26(4), 140–147 (2019) 35. J.-H. Lee, J. Park, M. Bennis, Y.-C. Ko, Integrating LEO satellite and UAV relaying via reinforcement learning for non-terrestrial networks, in IEEE Global Communications Conference (GLOBECOM) (2020), pp. 1–6 36. C. Jiang, X. Zhu, Reinforcement learning based capacity management in multi-layer satellite networks. IEEE Trans. Wirel. Commun. 19(7), 4685–4699 (2020) 37. A. Russo, G. Lax, Using artificial intelligence for space challenges: A survey. Appl. Sci. 12.10, 5106, (2022) 38. G. Labrèche, D. Evans, D. Marszk, T. Mladenov, V. Shiradhonkar, T. Soto, V, Zelenevskiy, OPS-SAT spacecraft autonomy with TensorFlow lite, unsupervised learning, and online machine learning. IEEE Aerospace Conference (AERO) (2022), pp. 1–17

Chapter 8

Technologies and Infrastructures for a Sustainable Space Ernestina Cianca, Simone Morosi, and Marina Ruggieri

8.1 Space Sustainability: The Problem The room the sustainability concept is gaining in the design of space systems, missions and infrastructures is encouraging [1–9]. Until a few years ago, the progress in activities related to the space domain was mainly measured in terms of conquering destinations (planets, stars), creating larger infrastructures (for example, mega-constellations), assuring longer manned stay in the International Space Station (ISS), conceiving more and more innovative services to be provided from space, etc. All the above matter is certainly very important and it represents in different ways a progress for Humanity. However, in the meantime, the priorities for assuring a decent future to Humanity were rapidly changing or, to be more precise, priorities that should have been dealt with efficiently and effectively long time ago were knocking hard at Humanity’s door to be considered with the proper attention. In the above frame, space-related activities seemed for some years out of the loop of the revolutionary change in Humanity’s priorities: the knowledge to be rapidly acquired is clearly becoming how to survive to the effects of ignoring sustainability for too long.

E. Cianca () University of Rome “Tor Vergata”, Rome, Italy e-mail: [email protected] S. Morosi University of Florence, Florence, Italy e-mail: [email protected] M. Ruggieri Center for Teleinfrastructures (CTIF), University of Roma “Tor Vergata”, Roma, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_8

185

186

E. Cianca et al.

The recent recognition from some of the space key-players that sustainability is also a matter to be considered in the development of infrastructures and services is an important step forward [4, 6, 8]. In fact, sustainability has an intrinsic holistic nature that needs attention not only on Earth but also in the whole space Earth is a small part of [3]. To give some numbers of the problem: together with active satellites, there are currently an estimated 330 million pieces of space debris, including 36,500 objects bigger than 10 cm, such as old satellites, spent rocket bodies and even tools dropped by astronauts orbiting around Earth. This crowded situation poses several challenges such as: • interference to astronomical observations; • radio frequency interference to other communication systems and challenging spectrum management; • challenges in space operations by shrinking the margin of error for maintaining separation between satellites; • high probability of collisions, further increasing the debris. Space soon will become an unsafe place to operate. Caring for sustainability of the space environment implies the following actions: • cleaning space from the junk produced by past (and most of the current) activities; • stop polluting through a common sustainability-prone strategy for future activities. In the following sections of the Chapter both the cleaning and the stop-polluting actions will be discussed, highlighting the current and envisaged status of their implementation and the related challenges. Authors’ aim is also to stimulate thoughts, new ideas and innovative solutions from the readers, because sustainability is not a matter of a few, but the biggest challenges ever in the history of Mankind.

8.2 Space Debris Mitigation/Removal With more than 300 million fragments populating the orbits around Earth and about 5000 defunct satellites and large abandoned objects, space debris has become one of the most severe threats to a sustainable access to space for humanity in the next future [10]. Moreover, the density of debris in space is growing with an exponential trend as depicted by Fig. 8.1 [11]. All these “space bullets” are travelling at relative speeds of several kilo-meters per second and they are wondering uncontrolled, risking to collide with other operational satellites. This gives an idea on how unsustainable risky space activities are: if the satellites launched in orbit are not quickly disposed at the end of their mission, the possibility of chain collisions first predicted in the 1970s by NASA

8 Technologies and Infrastructures for a Sustainable Space

187

Fig. 8.1 Space Debris growing trend, courtesy of ESA, Space Environment Statistics updated to Aug 2022

scientists that could possibly jeopardize the satellite classes around the Earth, could become real. Current disposal practices have shown to be insufficient: many studies prove that a removal efficiency of at least 90%, in cooperation with dedicated Active Debris Removal missions is the minimum viable to keep the debris population at a steady value but the current success rate is still very far from that value, being around 50%. The main strategies for Space Debris Mitigation/Removal are: • SSA and Collision Avoidance; • Space Debris Removal Techniques. Space Situational Awareness (SSA) refers to the knowledge of the space environment, including location and function of space objects and space weather phenomena. SSA is generally understood as covering three main areas: • Space Surveillance and Tracking (SST) of man-made objects; • Space Weather (SWE) monitoring and forecast; • Near-Earth Objects (NEO) monitoring (only natural space objects). Particularly, an SST system is a network of ground-based and space-based sensors capable of surveying and tracking space objects, together with processing capabilities aiming to provide data, information and services on space objects that orbit around the Earth. As a result the SST Systems are the basis for the implementation of suitable Spacecraft collision avoidance strategies, namely to provide risk assessment of collision between spacecraft or between spacecraft and space debris by minimizing the chance of orbiting spacecraft inadvertently colliding with other orbiting objects. The most common subject of spacecraft collision avoidance research and development is for human-made satellites in geocentric orbits. The subject includes procedures designed to prevent the accumulation of space debris in orbit, analytical

188

E. Cianca et al.

Fig. 8.2 Adherence to PMD regulations in terms of PMD manoeuvre. (a) For GEO satellite. (b) For LEO satellite

methods for predicting likely collisions, and avoidance procedures to maneuver offending spacecraft away from danger. The removal of space debris from highly crowded orbits can be done according to the following techniques: • de-orbiting, i.e. the forced reentry of a space object into the Earth’s atmosphere usually via a propulsion system at the End of Life (EOL); • reducing the orbital lifetime by accelerating the natural decay of spacecraft; • moving the space object is less populated “disposal” orbits at the EOL; • active removal of space debris. It must be outlined that the implementation of dedicated Post-Mission Disposal (PMD) technologies is still seen by many operators and officials as a burden for space industry’s competitiveness [12]–[13]. In LEO, where the removal manoeuvre is often more complicated than in GEO and the commercial exploitation of the orbits has just begun, the average level of adherence to PMD regulations and guidelines in terms of PMD manoeuvre has been about 45% over the past 10 years. In GEO, where there is a commercial interest in removing the satellites from its operational slots, in order to replace them with the new and more performing satellite the average level of adherence to PMD regulations and guidelines in terms of PMD manoeuvre has been of about 65% over the last 10 years. Above statistics are shown in Fig. 8.2. Current solutions for implementing SDM requirements rely mostly on the propulsive system already on-board the satellites for performing station-keeping manoeuvres. However this solution has shown a lower rate of success in implementing decommissioning manoeuvres. Active in-orbit debris removal would require some kind of space vehicle dedicated to this purpose. As a matter of fact, the cost of such a vehicle could be very high. Some studies have estimated a cost of 15 million for each piece of debris in LEO removed, not counting the cost of developing an orbital maneuvering vehicle.

8 Technologies and Infrastructures for a Sustainable Space

189

Also the use of tethers have been considered for deorbit large objects, but the cost would be high in any case. Other active removal schemes for small debris are: • “debris sweepers” such as large foam balls or braking foils; • ground- or space-based laser evaporation of debris surface material; All of the proposed techniques are expensive and technically daunting. Yet it is becoming increasingly clear that this will be a necessary component of space sustainability.

8.3 Sustainable-by-Design Approach: Enabling Technology 8.3.1 Concept of Sustainability by Design Awareness about the space sustainability matter should translate into a focused set of actions during the conceivement, design, deployment and management of any new system, infrastructure, mission or service. If sustainability becomes a goal only in an advanced stage of development effectiveness of any action will be much lower. Some of the readers might remember the dawn of security requirements in information systems. Caring about security needs in an advanced stage of the development was often bringing unsatisfactory results and exposing the system to risks. The criticality was increased when systems or infrastructures were integrating nonhomogeneous components (for instance, terrestrial, aerial and satellite portions). If we compare the integrated system or infrastructure to a patchwork blanket, interpatch stitching is often the most critical for the blanket lifetime. A successful approach in the design of a system or infrastructure, particularly if integrated, should then take security into account from the very beginning of the conceivement to guarantee a lasting resilience. Sustainability has a similar impact as security in both the behaviour of the system or infrastructure and its resilience capability in time, particularly if the system or infrastructure is integrated (e.g. [14]). The sooner the sustainability requirements are in the loop, the better the system or infrastructure will perform over time and under both predictable and unpredictable circumstances. Sustainability-by-Design (SubyD) is the approach that moves from sustainability requirements at the very beginning of the design phase and, even better, during the conceivement stage. A sustainable space system takes into account what is already available in space in order to both reduce/simplify hardware and upcycle existing infrastructures. This capability allows the system or infrastructure under design a backward compatibility (BW-Comp), that is feasible when a SubyD approach is used, while it becomes quite complicate and very costly when the sustainability requirements are not in an early phase of the design chain.

190

E. Cianca et al.

Furthermore, the system under design has also to look strategically at the future and, thus, it has to be prone to be used by systems that will be in space later. This capability allows the system under design a forward compatibility (FW-Comp), that makes sense if a SubyD approach is used not only by the system under design but also by most or, in a long term vision, all systems to be deployed in space. Both BW-Comp and FW-Comp imply that each system is a block of a multicomponent integrated infrastructure that evolves in the time domain. As highlighted in the comparison with security requirements, sustainability requirements become more critical in an integrated infrastructure. On the other hand, the integrated infrastructure is the means to implement both BW-Comp and FW-Comp. Therefore, the SubyD approach admits no excuse because only if it is followed by most or, in a long term perspective, all space players the achievable results will be effective. It looks pretty complicate to start the virtuous cycle of developing a sustainable space. The neutrality of technology is a good starting point. In fact, technology is neutral and only its use can be for good or bad purposes: in the space sustainability framework, “good” or “bad” measures the capability of a given technology to easy the realization of both BK-Comp and FW-Comp. In the next sections some technologies that are prone to the deployment of the SubyD approach and its pillar architectures based on backward and forward compatibility will be highlighted. The design based on sustainability requirements measures its effectiveness in terms of a quality parameter, the Sustainability design Efficiency (SdE) that can be expressed as: SdE = ηBW ηF W ηT K

.

(8.1)

where .ηBW , .ηF W and .ηT K are the design efficiencies related, respectively, to the use of BW-Comp, FW-Comp and ally technologies in the design. Each of the three factors of Eq. (8.1) can be increased by an extensive use of the existing space infrastructures, an informed and strategic vision on future missions and services and a brave and effective use of those technologies that support the SubyD approach. The unitary values of the three .ηBW , .ηF W and .ηT K efficiencies, that would imply a unitary value of SdE, are very unlikely to be reached in the short, medium and perhaps also long term. In fact, .ηBW = 1 would imply that the mission or service of the system under design can be performed without the launch of any additional hardware, .ηF W = 1 would indicate that all future missions or services could take advantage from the system under design and .ηT K = 1 would mean that all technologies adopted in the system under design be SubyD prone. Present values of the efficiencies in Eq. (8.1) are almost close to zero. To be optimistic, let’s say that there is a wide margin of improvement that can be spent to increase the three factors and, through them, the SdE value. A coordinated effort of the various space players and focused standardization activities would easy the capability of each system under design to increase both .ηBW and .ηF W . The conceptual flow of the SubyD approach is depicted in Fig. 8.3. The design is based not only on the conventional set of system requirements but also on

8 Technologies and Infrastructures for a Sustainable Space

191

Fig. 8.3 Conceptual flow of the SubyD approach

sustainability requirements that can be translated into three thresholds (A, B and C) of the three efficiencies composing the SdE quality parameter. The three SubyD pillars are all important to meet the goal of a sustainable space, but there is a logical flux of actions that envisages a sequential check first on .ηBW , then on .ηF W and, after system architecture is finalized, on .ηT K , as highlighted in Fig. 8.3. When the design meets the three thresholds, cost and time-to-market./time-tooperations might bring further trade-offs to be considered before the development phase, with eventual feedback on the choice of both system architecture and usable technologies. The definition and the adoption of a sustainability-aware approach to the design of space networks encompass the formalization of novel and targeted Key Performance Indicators (KPIs) to effectively assess the sustainability of the integrated infrastructure in its broader sense. In 2019 the World Economic Forum has launched an initiative to develop a Space Sustainability Rating (SSR) tool, [15]. ESA and the MIT are developing the SSR tool to score the sustainability of manufacturers and operators on the basis of factors such as plans to de-orbit systems upon completion of missions; choice of orbital altitude; ability of systems to be detected and identified from the ground; collision-avoidance tools; size and number of objects left in space

192

E. Cianca et al.

from the launch vehicle; and sharing of data. However, most of the considered KPIs of space sustainability are related to choices made in production phase (choice of the materials, lifetime, etc.) or they are related to the space mitigation/removal techniques that are implemented. The approach proposed in this Chapter calls for the definition of new KPIs that will have to take into account the following features: • the level and the pervasiveness of the softwarization and virtualization of the specific technologies which are adopted in the considered systems and networks; • the capabilities of inter-operation with previous and future technologies by means of BW-Comp and FW-Comp.

8.3.2 BW/FW Compatibility: Federated Satellite Systems A concept that is strongly related to the need of BW/FW compatibility is the concept of Federated Satellite Systems(FSS), which is an evolution of Distributed Satellite Systems (DSS) [16–18]. A satellite federation consists of a group of satellites that during their mission may decide to establish opportunistic collaborations with other groups of satellites to share resources that are underutilized such as commodities, data storage, data processing, downlink capacity, power supply, or instrument time. Such collaboration should result in a benefit for the satellite operator that decide to establish it and should not lead to a degradation of performance for the main mission of satellites. Therefore, the concept of FSSs is strictly related to the capability to interoperate with other spacecrafts/constellations already deployed and or that will be deployed in the future which would allow the “reuse” of the same infrastructure to provide other services. The concept of FSS was first introduced by Golkar [16] and mainly for Earth Observation constellations. Most of the previous works on FSS has focused on the business cases and opportunity to establish such a collaboration but not much effort has been posed to solve the challenging technical issues that are related to its implementation. In the short term, the concept of FSS could be implemented by using a negotiator node, a kind of gateway that adapts the communication protocols to enable the communication between satellites belonging to different communication systems and eventually operators. Newly designed missions should be flexible enough to intrinsically enable the establishment of opportunistic links between different satellite systems. The feasibility of an effective FSS requires high level of flexibility both at paylaod level and at network level. On one hand, it should be possible to establish communication links between heterogeneous communication nodes, and hence the transceiver must be highly adaptable and flexible. Moreover, opportunistic links between different constellations will make the network topology highly variable and hence, high level of adaptability is also required at network level. Finally,

8 Technologies and Infrastructures for a Sustainable Space

193

the feasibility of FSSs is strictly related to the feasibility of stable ISLs among heterogeneous spacecrafts, characterized by different sizes, characteristics and dynamic behaviour.

8.3.3 BW/FW Compatibility: Joint Communication and Sensing The design of Information and Communication Technology (ICT) systems that are able to jointly perform communication and sensing (and localization as a specific type of sensing), is a hot research area [19]. Such systems are referred as Joint Communication and Sensing Systems (JCS). On one hand, there is a strong interest in using LEO mega-constellations born to provide broadband communication services, to provide Positioning Navigation and Timing (PNT) services mainly in the events in which GLobal Navigation Satellite Systems (GNSS) signals become unavailable (deep urban canyons, under dense foliage, during unintentional or intentional interference). Such a solution is a key enabler of the backward compatibility and hence, of the space sustainability, as already deployed infrastructure is reused to provide novel services. Research in this field is now focused on facing the following challenges: • satellites do transmit satellite ephemerides and using the information that can be found in the two-line-element files introduces an error of kilometers due to several sources of perturbations. • LEO are not equipped with atomic clocks so they are not tightly synchronized. • LEO satellites are owned and operated by private entities which use proprietary protocols and hence novel specialized receivers must be developed that are capable for extracting navigation observables. An emerging area of research is the network/satellite-based geolocalization of Internet of Remote Things [20] devices via satellites that are used to provide them communication services [21]–[22]. The need to geolocalize IoRT terminals is not only related to the possibility to provide location-based services but also to improve the communication performance. In the release 17 of [23], when proposing adaptation of 5G NR or Narrow Band (NB)-IoT standards to the use with satellites links, the assumption is the IoT terminals are equipped with GNSS receiver. This is not always feasible if the IoT terminals are low cost, batterypowered devices. Therefore, it would be crucial to localize them from the satellite by using the communication signals. On the other hand, much effort is nowadays also focused on the design of future space systems that can, by design, provide jointly communication/navigation and sensing. Therefore, such approach is a key enabler of the FW compatibility. In this framework, lot of research activity has addressed the issue of novel waveforms. In particular, the novel waveform OTFS, which has attracted interest for high date rate communications from LEO satellites

194

E. Cianca et al.

(see Chap. 1), has also interesting characteristics when joint communication and localization services must be provided [24].

8.3.4 Ally Technologies As shown in Fig. 8.4, the concept of FSS and JCS are key ingredients to make satellite infrastructure BW and FW compatible thus reducing the need to launch new nodes and infrastructure elements. The use of a negotiator node would enable the establishment of FSSs with currently deployed satellite systems and thus contributing to the BW compatibility. On the other hand, future missions should be designed with communication nodes already able to establish opportunistic links with other elements of the deployed infrastructure to provide novel services, thus enabling the FW compatibility. At the same time, the reuse of already deployed infrastructure (BW compatibility) will be enabled by the research in the field of JCS towards the use of signals transmitted by current systems (e.g., the use of communication signals for navigation purposes). On the other hand, the research on JCS aims to make future system more flexible and able to natively provide different services using the same signals. Moreover, in Fig. 8.4 The key enabling technologies to the implementation of FSS and JCS, shown in Fig. 8.4, are presented in the rest of this Section.

Fig. 8.4 Key technologies enagling BW/FW compatibility

8 Technologies and Infrastructures for a Sustainable Space

195

Fig. 8.5 Full space SDN approach

8.3.4.1

Softwarization

The sustainability-driven design can take an enormous advantage from the choice of technologies that easy the development and management of system architectures being both backward and forward compatible over a long period of time. A pervasive use of the Software Defined (SD) paradigm into the space system or infrastructure is a major ally to the easy, massive and lasting application of the backward and forward compatibility [3]. The SD approach can benefit both networking and data storage through the powerful decoupling between physical and control/service components [25]. In the last years the SD paradigm is receiving attention and focused efforts, in particular for integrated space-terrestrial frameworks (e.g. [26, 27]). In the medium and long term, architectures when even the Software Defined Network (SDN) controller is in space are envisaged (Fig. 8.5). Considering, then, that the FW compatibility moves the overall architecture ahead over a time sliding window, also the SDN controller could be moved from a current system to a future one, due to the flexibility of a full space-based SD approach and an effective coordinated effort among space players. Besides the obvious issues related to the pervasive implementation in space of new paradigms, like the SD approach, there is a further important aspect that relates to all sustainability ally technologies.

196

E. Cianca et al.

Any technology that supports sustainability-prone design and architectures has to be sustainable itself. This, perhaps, is the most challenging aspect. To understand the problem, let’s focus for a moment on what is happening on the effort of terrestrial connectivity infrastructures to become “green”. On one hand, the pervasiveness of connectivity is the key to render “green” most of the vertical domains, from energy to heath to industry, just to mention some very popular application realms. On the other hand, the pervasiveness and its consequences, like for instance the spreading of edge computing, the amount of small and spread around data centres, the increased transport and computational capacity related to the cloud operations need to become truly energy efficient so that pervasive connectivity be indeed a relief for the vertical sectors and, thus, for the Planet from the sustainability viewpoint [28–30]. Similarly, when dealing with space sustainability ally technologies, in particular with SDN, focused efforts are needed to render them energy efficient (e.g. [31, 32]).

8.3.4.2

Autonomy and AI Tools

Automated systems are systems where the system knows exactly how to react for any situation that is predicted. When unpredicted situations occur the system gets stuck. On the other hand, an autonomous system is able to to react at its best in any possible situations. Historically, the work on spacecraft autonomy has been focused on deep-space exploration missions. In the framework of space sustainability, the introduction of some level of autonomy could be crucial for: (1) prevention of collisions due to space debris; (2) spectrum and interference management [33]. Deep learning tools have been proposed to detect external threats (space debris) and react to avoid collisions by replanning the route [34–36] to maximise mission efficiency and minimise the risk of collision with resident space objects. Moreover, autonomy could be used also for establishing more quickly, only based on local information, opportunistic connections with heterogenous spacecrafts to support the introduction of new missions [36]–[37]. In [37], a predictive algorithm was developed to estimate future satellite contacts and predict routes overtime in which federations can be established.

8.3.5 Very High-Speed Inter-Satellite/Inter-Layer Links In such a highly softwarized space infrastructure, with many decentralized functions and high level of autonomy, higher volumes of signalling and control data will have to be exchanged by network nodes, besides the user data. It must be outlined, that space systems are part of a multi-layered architecture whose non terrestrials nodes are not only the satellites in different orbits, but also High Altitude Platforms (HAPs) and Unmanned Aerial Vehicles (UAVs). Therefore, high throughput links between space nodes and more in general space and aerial nodes are needed and

8 Technologies and Infrastructures for a Sustainable Space

197

the use of millimiter waves (mmWave)/optical links and more recently also of Terahertz links [38], become a key enabler of such a vision. As presented in Chap. 1, the use of Q/V bands in the feeder link of HTS systems for broadband services is a consolidated concept. Their use for ISLs but also inter-layer links between satellites (also in different orbits) and HAP/UAVS poses many challenges such as lack of channel models considering platform vibrations, related mispointing and tracking losses, high Doppler shifts besides the atmospheric attenuation. Novel waveforms and error correcting mechanisms should be investigated [39]. Another key issue that have an impact of the feasibility of mm-waves/THz links and on the trade-off between HW and SW implementation, between flexibility and number of nodes that are needed to cover a given service area, is the antenna design and beamforming architectures [40]. Some recent papers have proposed hybrid analogdigital implementation for beamforming in mm-waves UAV-ground links as a good trade-off between the flexibility offered by the full digital implementation and the lower power consumption associated to an analog implementation [41]. An important challenge for THz links is the need to fine alignment of pencil-beams in presence of high Doppler and relative speeds. In [42], the use of Reconfigurable Intelligent Superfaces (RISs) is proposed as a highly energy efficient fashion to facilitate the beam alignment.

8.4 Conclusions The way for a truly sustainable space is paved with challenges and brave decisions as well as with a high degree of coordination among key players. It looks very hard, but the result would be very rewarding: an unprecedented growth in the ability of deploying and managing systems and infrastructures able to last much more than usual, due to the sliding time window of the forward compatibility that, in turns, moves from a convincing backward compatibility. It should be also noted that the efforts for energy efficiency reported in the literature are named “green” even if they are referred to the space realm. Perhaps, space sustainability could bring to a new two-dimensional colour definition: “green” and “blue”. In fact, “green” is the goal for the impact on Earth of producing a given technology, while “blue” is the goal of the impact that technology should have in terms of space sustainability. The holistic nature of sustainability suggests that a sustainable space needs only “green-blue” technologies. What a fascinating and coloured goal!

References 1. M. Ruggieri, T. Rossi, New fascinating challenges for space systems: Softwarization, ai-based robotization and sustainability. Which role for cubesats? in Advances in the Astronautical Sciences (Univelt, Escondido, 2020), pp. 609–615

198

E. Cianca et al.

2. A. Murtaza, S.J.H. Pirzada, T. Xu, L. Jianwei, Orbital debris threat for space sustainability and way forward (review article). IEEE Access 8, 61000–61019 (2020). https://doi.org/10.1109/ ACCESS.2020.2979505 3. E. Cianca, M Ruggieri, Space sustainability: towards the future of connectivity, in Chapter 14, Women in Telecommunications Springer Cham, (2023). ISSN 2509-6435 4. ESA, Applications/telecommunications & integrated applications. Sustainable Connectivity in Space (2022). https://www.esa.int 5. BSR, Sustainability in space: the next frontier (2022). https://www.bsr.org/en/emerging-issues/ 6. E. Howell, SpaceX promises sustainability and safety for Starlink constellation. Space (Future US, New York, 2022). https://www.space.com/spacex-sustainability-safety-starlink-satellitemegaconstellation 7. R. Shields, Space sustainability as a national priority in the United States. J. Space Saf. Eng. 1–5 (2022). https://doi.org/10.1016/j.jsse.2022.08.002 8. OneWeb, Responsible Space (2022). https://oneweb.net/about-us/responsible-space 9. Secure World Foundation, Space sustainability: a practical guide (2018). www.swfound.org 10. A. Fanfani, Communication techniques, architecture and services for satellite application in critical scenarios. Ph.D. Thesis Dissertation, 2017 11. S. Frey, S. Lemmens, B. Bastida Virgili, T. Flohrer, Level of aherence to sdm guidelines. Technical Report Issue 1.0, ESA/ESOC Space Debris Office, 2016 12. IADC. space debris mitigation guidelineas, UNCOPUOS 40-th session, Vienna 17-28 February 2003, Space Debris, A/AC.105/C.1/L-260 13. F. Alby, D. Alwes, L. Anselmo, European code of conduct for space debris mitigation. Technical Report Issue 1.0, 2004 14. M. Hosseinian, J.P. Choi, S.H. Chang, J. Lee, Review of 5G NTN standards development and technical challenges for satellite integration with the 5G network. IEEE Aerosp. Electron. Syst. Mag. 36(8), 22–31 (2021). https://doi.org/10.1109/MAES.2021.3072690 15. M. Rathnasabapathy et al., Implementing the space sustainability rating: an innovative tool to foster long-term sustainability in orbit, in 72nd International Astronautical Congress, Dubai, United Arab Emirates, 25–29 October (2021) 16. A. Golkar, Federated satellite systems: a case study on sustain ability enhancement of space exploration systems architectures, in Proceedings of the International Astronautical Congress, IAC, vol. 11 (2013), pp. 9063–9076 17. J.A. Ruiz-de Azua, L. Fernandez, J.F. Munoz, M. Badia, R. Castella, C. Diez, A. Aguilella, S. Briatore, N. Garzaniti, A. Calveras, A. Golkar, A. Camps, Proof-of-concept of a federated satellite system between two 6-unit cubesats for distributed earth observation satellite systems, in Proceedings of IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium (2019), pp. 8871–8874 18. J.A. Ruiz-de Azua, N. Garzaniti, A. Golkar, A. Calveras, A. Camps, Towards federated satellite systems and internet of satellites: the federation deployment control protocol. Remote Sens. 13(5) (2021) 19. R. Giuliano, The next generation network in 2030: applications, services, and enabling technologies, in Proceedings of the 8th International Conference on Electrical Engineering. Computer Science and Informatics (EECSI) (2021), pp. 294–298 20. M. De Sanctis, E. Cianca, G. Araniti, I. Bisio, R. Prasad, Satellite communications supporting internet of remote things. IEEE Int. Things J. 3(1), 113–123 (2016). https://doi.org/10.1109/ JIOT.2015.2487046 21. C.A. Hofmann, A. Knopp, Tracking of remote IoT devices by satellite assisted geolocation, in Proceedings of IEEE International Conference on Communications (ICC) (2020), pp. 1–6 22. I.S. Mohamad Hashim, A. Al-Hourani, B. Ristic, Satellite localization of IoT devices using signal strength and doppler measurements. IEEE Wirel. Commun. Lett. 11(9), 1910–1914 (2022) 23. 3GPP TR 38.821, Solutions for NR to support non-terrestrial networks (NTN). Technical Specification Group Radio Access Network, 16.0.0. 3rd Generation Partnership Project (3GPP). Technical Report (2019–12)

8 Technologies and Infrastructures for a Sustainable Space

199

24. S. Li, W. Yuan, C. Liu, Z. Wei, J. Yuan, B. Bai, D.W.K. Ng, A novel ISAC transmission framework based on spatially-spread orthogonal time frequency space modulation. IEEE J. Sel. Areas Commun. 40(6), 1854–1872 (2022) 25. T. Rossi, C. Fragale, M. De Sanctis, E. Cianca, M. Ruggieri, H. Fenech, Software defined networking and high throughput satellite: the best matching for space-based communications infrastructures, in Proceedings of the 21st Ka and Broadband Communications Conference, Bologna (Italy) (2015). On line proceedings ISSN -2573-6124 26. Y. Bi, G. Han. S. Xu, X. Wang, C. Lin, Z. Yu, P. Sun, Software defined space-terrestrial integrated networks: architecture, challenges, and solutions. IEEE Netw. 22–28 (2019). https:// doi.org/10.1109/MNET.2018.1800193 27. T. Li, H. Zhou, H. Luo, S. Yu, SERvICE: a software defined framework for integrated spaceterrestrial satellite communication. IEEE Trans. Mobile Comput. 17(3), 703–716 (2018) 28. T Huang, W. Yang, J. Wu, J. Ma, X. Zhang, D. Zhang, A survey on green 6G network: architecture and technologies. IEEE Access 175758–175768 (2019). https://doi.org/10.1109/ ACCESS.2019.2957648 29. A. Montazerolghaem, M.H. Yaghmaee, A. Leon-Garcia, Green cloud multimedia networking: NFV/SDN based energy-efficient resource allocation. IEEE Trans. Green Commun. Netw. 4(3), 873–888 (2020). https://doi.org/10.1109/TGCN.2020.2982821 30. A.A.Z. Ibrahim, F. Hashim, A. Sali, N.K. Noordin, S.M.E. Fadul, A multi-objective routing mechanism for energy management optimization in SDN multi-control architecture. IEEE Access 20312–20327 (2022). https://doi.org/10.1109/ACCESS.2022.3149795 31. J. Zhang, X. Zhang, M.A. Imran et al., Energy efficient hybrid satellite terrestrial 5G networks with software defined features. KICS J. Commun. Netw. 19(2), 147–162 (2017). https://doi. org/10.1109/JCN.2017.000024 32. Z. Tu, H. Zhou, K. Li, M. Li, A. Tian, An energy-efficient topology design and DDoS attacks mitigation for green software-defined satellite network. IEEE Access 211434–211450 (2020). https://doi.org/10.1109/ACCESS.2020.3039975 33. A. Moubayed, T. Ahmed, A. Haque, A. Shami, Machine learning towards enabling spectrumas-a-service dynamic sharing, in Proceedings of IEEE Canadian Conference on Electrical and Computer Engineering (CCECE) (2020), pp. 1–6. https://doi.org/10.1109/CCECE47787.2020. 9255817 34. European Space Agency, AIKO: Autonomous operations thanks to artificial intelligence. Online available. (2018) 35. V. Kothari, E. Liberis, N.D. Lane, The final frontier: deep learning in space, in in Proceedings of the 21st International Workshop on Mobile Computing Systems and Applications (HotMobile ’20). Association for Computing Machinery, New York, NY, USA (2020), pp. 45–49. https://doi. org/10.1145/3376897.3377864 36. E. Lagona, S. Hilton, A. Afful, A. Gardi, R. Sabatini, Autonomous trajectory optimisation for intelligent satellite systems and space traffic management. Acta Astronautica 194, 185–201 (2022). ISSN 0094-5765. https://doi.org/10.1016/j.actaastro.2022.01.027 37. J.A. Ruiz-De-Azua, V. Ramírez, H. Park, A.C. Augè, A. Camps, Assessment of satellite contacts using predictive algorithms for autonomous satellite networks. IEEE Access 8, 100732–100748 (2020). https://doi.org/10.1109/ACCESS.2020.2998049 38. Y. Li, Y. Chen, Propagation modeling and analysis for terahertz inter-satellite communications using FDTD methods, in Proceedings of IEEE International Conference on Communications Workshops (ICC Workshops) (2021), pp. 1–6. https://doi.org/10.1109/ICCWorkshops50388. 2021.9473712 39. M. De Sanctis, E. Cianca, T. Rossi, C. Sacchi, L. Mucchi, R. Prasad, Waveform design solutions for EHF broadband satellite communications. IEEE Commun. Mag. 53(3), 18–23 (2015). https://doi.org/10.1109/MCOM.2015.7060477 40. W. Jiang, H.D. Schotten, Initial access for millimeter-wave and terahertz communications with hybrid beamforming, in Proceedings ICC 2022 - IEEE International Conference on Communications (2022), pp. 3960–3965. https://doi.org/10.1109/ICC45855.2022.9838386

200

E. Cianca et al.

41. Z. Xiao et al., A survey on millimeter-wave beamforming enabled UAV communications and networking. IEEE Commun. Surv. Tutorials 24(1), 557–610 (2022). https://doi.org/10.1109/ COMST.2021.3124512 42. K. Tekbıyık et al., Reconfigurable intelligent surface empowered terahertz communication for LEO satellite networks (2020). ArXiv abs/2007.04281

Part III

Interplanetary Networking

Chapter 9

Softwarization in Satellite and Interplanetary Networks Sisay Tadesse Arzo, Riccardo Bassoli, Michael Devetsikiotis, Fabrizio Granelli, and Frank H. P. Fitzek

9.1 Introduction The first human who traveled to the moon is Niels Armstrong when he landed on the surface of the moon under the Apollo 11 program. Other programs have targeted other plants, including Jupiter, Mars, Venus, and Mercury along with the available moons the planets, etc [1]. Voyagers 1 and 2 are the furthest a spaceflight traveled. They are expected to have sufficient electric energy and thruster fuel that could enable them to continue their current suite of science instruments on until at least 2025. When Voyager 2 was 18.4 billion km from the Sun and then Voyager 1 was about 22.1 billion km [2]. Most recently, the Artemis was launched by NASA the leading institute in space exploration [3] program. The Artemis is aimed at landing humans on the Moon by 2024. This is expected to embark new pioneering technologies in investigating much of the lunar surface than ever. The program is aimed at learning also to learn what is on and around the Moon in taking the next massive step, which is to send humans to Mars. At the end of the current decade, it is expected that the collaboration

S. T. Arzo () · M. Devetsikiotis University of New Mexico, Albuquerque, NM, USA e-mail: [email protected]; [email protected] R. Bassoli · F. H. P. Fitzek Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany Centre for Tactile Internet with Human-in-the-Loop (CeTI), Dresden, Germany e-mail: [email protected]; [email protected] F. Granelli Department of Information Engineering and Computer Science (DISI), University of Trento, Trento, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_9

203

204

S. T. Arzo et al.

between international and emerging commercial partners could create sustainable exploration and discoveries on both martian and lunar surfaces [3]. For a long time in human history, the exploration of the moon, Mars, and other deep space elements has been desired. So far, we have also made significant progress on various technological fronts. However, we have never been able to travel further than the moon. Since our first landing, we are not even able to go back to the moon itself. All space travel has only happened toward orbiting the earth, through ISS. In the next decade, given the current attention and activities by commercial and governmental organizations and other entities, exploration of the Moon and Mars is expected to explode. This is already been witnessed with great signs in the launching of various programs, in particular, the commercial launching of space programs such as SpaceX is leading the way as a business venture. SpaceX was founded by Elon Musk in 2002 with aim of reducing the costs of space transportation enabling the colonization of Mars. Since its inception, the company manufactured different space vehicles such as the Falcon 9, Falcon Heavy, and Starship launch vehicles. It also produced Starlink communications satellites, several rocket engines, and Cargo Dragon and Crew Dragon spacecraft. As a governmental organization, the recent Chines Chang’e 5 spacecraft which had successfully landed on the moon and brought soil. The program was designed as an experiment for the return to lunar spacecraft. On 1 December 2020, Chang’e 5 landed in the vicinity of Mons Rümker on the moon, which was launched in Nov. 2020. It comes back with 2 kg of lunar soil on 16 December 2020 [4]. Moreover, as a part of the Tianwen-1 mission, China’s first interplanetary venture enabled a successful landing of the Mars rover called ‘Zhurong’ Red Planet [5]. Furthermore, ESA also planned to send spacecraft to Mars in 2020. The program is called N.◦ 6–2020: ExoMars, which is postponed to take off for the Red Planet in 2022 [6]. More recently, the successful landing of the Perseverance rover excited the research and space community as well as the public. The main mission of the perseverance rover is to search for traces of ancient life and gather what could be the first rocky samples from Mars that will be sent back to Earth [7]. The most promising samples will be packed for return to Earth with the later missions. Controlled flight was performed on another planet on Ingenuity Mars Helicopter which was carried by Perseverance. Figure 9.1 depicted the images of curiosity and perseverance Mars rover and ingenuity helicopter drone. From a communication and network coverage perspective, the most notable advancement is the recent mission by Nokia and NASA to deploy LTE access on the moon. 4G is expected to transform lunar surface communication access. It is aimed at providing reliable, high-data rates while optimizing power consumption, cost, and size. Wireless communications is a vital part of NASA’s Artemis program as it will create a sustainable existence on the Moon, which is expected to be achieved by 2030 [8]. Moreover, the commercial launching of Starlink communications satellites by SpaceX is also significant progress. The recent advancement in the communication industry is dramatically changing space communication. In particular network softwarization, flexible splitting of monolithic-based network functions, application of artificial intelligence for

9 Softwarization in Satellite and Interplanetary Networks

205

Fig. 9.1 Curiosity and perseverance Mars rover and ingenuity helicopter drone. Images are furnished by NASA

network management, and drone-based communication are driving the space communication industry to the next level. This is through 3D networking with the deployment of terrestrial connectivity using access points or cellular networks along with drone, balloon, underwater, and satellite connection. In this regard, network softwarization creates flexible system design approaches which pave the way for network intelligence. Network intelligence and automation are very necessary for remote-based network design, development, deployment, and monitoring. A softwarized network can be deployed in a vastly distributed environment with dynamic deployment. With the introduction of Machine Learning (ML) and AI the deployed network can be managed autonomously without the physical presence of a network manager. This is very relevant in the case of space applications where the presence of network design and operate extremely challenging.

9.2 Computational and Communication Technologies for Massive Space Exploration The great advancement in computational and network technologies in the last decade has transformed the way we design, deploy and operate communication systems. In this regard, space exploration could utilize these technologies to facilitate rapid and massive exploration of outer space. In particular cloud, edge, and fogbased computing along with network softwarization and in-network intelligence, communication in space exploration would be facilitated.

206

S. T. Arzo et al.

9.2.1 Cloud, Edge, and Fog Computing Computing is required to perform various analyses on Mars or any remote mission. This could be the weather condition soil content, chemical composition of rocks, analysis, and before sending them back. Even autonomic control of rovers, drones and networks that provides coverage requires huge storage and computational power. As demonstrated by the recent perseverance rover landing on Mars, it is possible to reprogram the onboard device for a different mission. For the perseverance, after the rover landed, the controller is re-programmed by NASA engineers using commands sent from Earth to potentially perform mobility based on visual processing. This demonstrates the possibility of complex task execution by a single rover or more collaborative rovers in the future. However, for massive and complex missions it may need various rovers, drones, autonomous equipment, or other IoT devices. The collaboration of such a mission requires both network coverage and standard computing. It is possible to fully equip the collaborating devices with internally embedded computing. However, it will be inefficient in a distributed and collaborating mission. Therefore, a cloud-based computing provisioning to a remote mission will demonstrate significant efficiency in availing storage and computing power to the exploration missions. Computing, networking, and control cannot be alienated in a space mission. Moreover, computing, control, and networking are complementary technologies that could facilitate the Martian massive exploration. Principles of control help for network control, edge computing for networking, networking for clustering, and interconnecting of separate computing units. Computing provides a resource for sophisticated controlling algorithm computation. There are some works on the use of edge for control and management algorithm deployment demonstration. An interesting work on Mission-critical service control using edge computing and 5G network is presented in [9].

9.2.2 Network Coverage, Network Softwarization, and Network Automation In this section, we will focus on network softwarization for network connectivity on remote sites such as Mars and the moon and communication with earth. There is a great advancement in the networking industry that is implemented and under development to be deployed in the near future. These technologies are providing a tremendous benefit in various sectors of human development. These are from simple voice communication to the video call services, from simple computer interconnection up to the worldwide web from simple on-demand video access to the critical for remote surgery. The advancement has enabled various types of devices to interconnect providing worldwide coverage with various types of

9 Softwarization in Satellite and Interplanetary Networks

207

technologies wireless and wired. To enable this various mechanisms are developed such as twisted pair cable, coaxial cable, and fiber access as a physical transmission. The same is true in the wireless domain. Several (de)encoding, channel access protocols, (de)encryption techniques, security tools have been developed to enable communication through both wired and Wireless. Numerous architectural models, theoretical concepts, implementation mechanisms, have also been developed and continuously improved. In this subsection, we review important advancements in networking that could have the potential to be adopted in space exploration.

9.2.2.1

Virtualization and Softwarization

Virtualization and softwarization of the network will help tremendously for two main reasons: reduction of the need for physical equipment and generalization of the hardware required (general purpose CPU, Memory, and storage). Software-defined networking (SDN) is a technique for centralized programming of networks through a centralized controller. This provides flexibility and dynamic control of networks. In remote exploration like that of Mars, SDN is an ideal approach for network operation and management. It provides the possibility of developing a dynamically adaptive network. Moreover, it paves the way for the autonomic controlling of a network through AI. The architecture of an SDN is shown in Fig. 9.2. Network function virtualization (NFV) is also an important network technology that provides a software version of network functions that provide the controlling, management, and operation of the network. The architecture of an NFV is presented in Fig. 9.3. A Software Defined Radio (SDR) is a softwarized radio communication system that process various signals such as coding, decoding, modulation, demodulation, etc. It uses software for the modulation and demodulation of radio signals in a communication system. Since it is a softwarized technique. SDN will have a huge part to play in the future communication network for space exploration missions that can potential benefit a lot from SDRs. It can be developed to have an adaptive algorithms such as machine learning and artificial intelligence. Currently, there are huge development in the military and terrestrial application of SDR [12]. However, a lot has to be investigated in this area for space mission. In this regard, the work in [13] A Reconfigurable Multi-Modal SDR Transceiver for CubeSats aiming at reducing the payload of CubeSats. Using SDR it could be possible to reprogram the payload based on new transmission standard and protocols.

9.2.2.2

Backbone Network Technologies for Space Applications

The existing communication between moon/mars and earth is through wireless links. For example, the curiosity rover, which had touched down Mars, sends radio waves through its ultra-high frequency (UHF) antenna with 400 Mhz to communicating with station on Earth relaying on NASA’s Mars Odyssey and Mars Reconnaissance Orbiters. To serve as both its “voice” and its “ears.”, curiosity has

208

S. T. Arzo et al.

Fig. 9.2 SDN management architecture [10]

three antennas. They are installed at the back of rover equipment deck. To increase reliability, a back up communication option, the rover is equipped with multiple antennas. There are networks of antennas deployed in three strategic locations of the earth. They are called Deep Space Network (DNS) which are located in the United States (California), Spain (Madrid), and Australia (Canberra). They support NASA’s interplanetary spacecraft missions [14]. Each DSN site has one huge, 70 m diameter antenna. The antennas are designed with the largest and most sensitive capability. They are able to track spacecrafts traveling a distance of billions of km from Earth [14, 15]. There is some advancement in satellite-based networks that could be extended to encompass deep space communication. This technological advancement and convergence of satellite communications would provide a converged network of networks such as a worldwide web in Mars, moon, earth, etc. In [16], the authors suggested a potential architecture of Space-Terrestrial Integrated Network (STIN) that integrates the existing Internet, mobile wireless networks, and the extended space network. The architecture is aimed at providing comprehensive services globally that can be accessed anytime and anywhere. Moreover, very recently, 3D network is gaining application for the integration of terrestrial network with the Marian based connectivity. The work in [17] presented a investigation of 3D network based on 6G flexible network targeting provisioning of mobile connectivity on martian surface. Future network such as beyond 5G and 6G are expected to

9 Softwarization in Satellite and Interplanetary Networks

209

Fig. 9.3 Network function virtualization [11]

deliver guaranteed Quality-of-Service (QoS) for heterogeneous applications with optimized network resource usage. In this regard, the backbone plays a crucial role in interconnecting geographically distributed and vast distant networks. On earth, the transitional backbone network extends 100–1000 km distances. However, when it comes to space the backbone network ranges in millions of km. Thus it is significantly affected by the distance in terms of electromagnetic wave propagation and physical deployment possibility. The difficulty of erecting a wired technology for space communication hinders the use of traditional backbone technologies such as coaxial cable and optical networks. The most viable technology that could be used as a backbone network is wireless communication. This could be through radio links as in the case of DSN, microwave links, and free-space optical networks.

9.2.2.3

Network Coverage Technologies on Remote Environment

Network coverage in the remote site could take some inspiration from the existing technology that is implemented on earth [18]. The most convenient connectivity technology that could have an important contribution to space exploration is

210

S. T. Arzo et al.

wireless technologies. For example, a wireless cellular network could be used to provide a wireless access network on Mars [19, 20]. The authors in [20] discuss the possible use of IEEE 802.11 a and b wireless local area network for wireless networks on the Martian surface. They presented modeling of the physical layer. Moreover, in [19] discussed the communication aspects of Martian missions. They used the deployment of a Martian wireless network infrastructure considering LTE on Mars (LTE-M). Other existing works in the area of access coverage through cellular, drone and balloon-based network coverage could be considered to adopt in Mars [21]. The physical layer modeling of the Martian surface is dependent on the Maritain atmosphere and terrain. Depending on the geography of Mars, it also varies from place to place that should be considered in the design and modeling of the physical layer signaling propagation. Depending on the mission plane, which could be long term or short-term plane, the technological adaptation in providing the coverage could also be considered. Dynamic changes as the exploration mission being executed the technology needs changes over time. E.g., first the mission could be to evaluate the composition of a given place and weather conditions of the same place and time. In that scenario what kind of device should be used, and what kind of rover should perform the task should be defined. Based on the required exploration task the network could be dynamically provided. Moreover, when the mission changes, which could be to check on the other part of the martian surface such as Eberswalde, Holden Crater a different network dynamics could be configured that could be based on drones or balloons. The work in [22], presents early results for the modeling the RF considering Martian environment to determine the characteristics of possible wireless, rovers, and sensor networks. The work used commercial available RF propagation modeling software, which are designed for traditional cellular telephone system planning, along with the topographic data of Martian environment to determine and construct Mars’ propagation path-loss models. A code division multiple access communication system for Mars based on geostationary relay satellite is presented in [23]. The paper defines CDMA based communication network for various assets including rovers and landers on Martian surface, low Mars orbiters, and CubeSats. They are in the vicinity of Mars, and they use a geostationary relay satellite at 17,000 km above Martian surface. Using 8.40 GHz frequency, the assumed data rates are between 50 Kbps and 1 Mbps. In [24] proposed an adaptive feedback-supported communication technique that can minimize the energy consumption of the communication with a spacecraft or wireless sensor nodes. There are various IoT connectivity technologies with the potential to revolutionize space exploration. The first IoT connectivity technology to be adopted in space is Wi-Fi. Wi-Fi has enabled a networked space exploration. NASA has provided Wi-Fi access by installing the first access points (APs) on the International Space Station in 2008 [25]. Lora could also be the next to provide connectivity in Mars or moon exploration.

9 Softwarization in Satellite and Interplanetary Networks

9.2.2.4

211

Network Automation

Automated network management is necessary for deep space exploration due to the difficulty of human presence in space. Network automation is the ability of a network to autonomously manage itself. Autonomic networking is scales up the management capability in addressing the expected dynamic growth networks. Due to the obvious reasons for the unavailability of humans to install and manage the network, we require the following capability of a network that should be deployed on Mars. A general cyclic network management for automated processing is presented in Fig. 9.4. • Self-Installation:installing hardware equipment is required to provide coverage. Once the required equipment is delivered in the appropriate places, the network equipment has to install itself. The delivery could be through a martian rover or drone. The installation could require digging holes on the Marian surface to fix the antenna or other required hardware equipment. The digging, placement, and fixing of the hardware may need to consider the Marian surface for dust and rocky areas. • Self-Configuration is a feature of a network to configure itself using predefined policies to achieve a particular control and management performance. That is performed autonomously. • Self-optimizing is the ability of the network to utilizes the available computational and communication resources to achieve the best performance dynamically adjusting itself to meet the dynamic demands. The network mostly follows a set

Fig. 9.4 Network management cycle [10]

212

S. T. Arzo et al.

of pre-defined policies and measure its performance to make sure that it satisfies the expectations. • Self-protecting ensures that the network can protect itself against any potential security breaches or attacks such as Denial-of-Service or Distributed-Denial-ofService attacks. • Self-healing is a capability of the network to discover and resolve the failures automatically in the shortest amount of time possible. This is necessary to protect or re-establish the service in the network whenever failure of any network element happens. • Self-drone based areal coverage of network considering the demands that could be performed using self by a combination of driving drone, autonomous network control and management and drone coordinating controller in unknown environments. E.g using swarm of drones for network coverage. As in-network intelligence introduction for autonomic network management system is presented in Fig. 9.5 as an example for cloud radio access network(C-RAN).

Fig. 9.5 Autonomic network management system [10]

9 Softwarization in Satellite and Interplanetary Networks

9.2.2.5

213

Service Oriented Architecture for Intelligent Network Design

Microservice and multi agent system have been competing as a replacement for monolithic service design for service oriented architecture. In this regard, micorservice has been studied and adapted widely both in research community. However, recently, multi-agent has been proposed as the preferred paradigm to design a decoupled service integrating in-network intelligence [10]. A multi agent based network management automation architecture has been proposed. The architecture showed how to design intelligent and autonomous atomic agent incorporating appropriate machine/deep learning algorithms as a cognitive component. The agents can be used as a building block to design a complete autonomous system. This is very useful for remote network design, deployment and operation, in particular for massive network infrastructure deployment in lunar or Martian surface to facilitate the massive exploration process.

9.2.3 Internet of Things Internet of Things (IoT) has various applications in our daily life such as health monitoring, green energy, environment monitoring, smart home, and smart city [26– 29]. IoT has provided so many applications using computing and wireless networks advancement. Due to this, IoT has caught the attention of researchers and private industries. The rate of interconnected IoT devices is overwhelmingly increasing and continuously growing with time. As more IoT objects are connected, there is an increase of information in the form of data in the interconnected system [30]. IoT for space application is starting to take shape. Currently, IoT in space is at the conceptual development stage than actual applications. It is because of many obstacles to overcome before organizations can start to deploy and use IoT in space for practical applications. However, a different and alternative approach may need to be explored. Spacial on-site manufacturing and utilization of IoT devices are more viable than transporting them from the earth over a long distance. Both mechanisms have huge challenges before being realized and are sometimes complementary in that what can not be manufactured or important to initial materials should be transported. What can be manufactured in remote sites could help the exploration paving the way for human transportation and presence preparing for human arrival. This article reviews the most recent research activities on the application of IoT technology for space applications. This challenging issue is difficult to resolve with the existing infrastructure. It means that there should be a solution with a new concept and approach that takes data rate, performance, and physical environment into consideration in trying to come up with interplanetary communication. When communication and controlling technology are advanced enough, IoT is expected to have immense potential to facilitate and revolutionize space exploration. The peculiarity of space exploration which comes due to the vast distance to the target environment to be studied has

214

S. T. Arzo et al.

tremendous challenges. The challenge is to effectively deploy, configure, control, and manage remotely which requires extremely expensive operations. NASA is putting an incredible effort into the adoption of IoT for space application. It has already setup an IoT lab [31] at Johnson space center and other virtual labs in anther places such as at Ames research center, Kennedy space center, and jet propulsion laboratory. It was setup in the federal government, which has completed the first phase and documentation, searching for an IoT platform and collecting data on the twenty selected devices. More recently, in another effort, NASA and Stanford collaborated to launch a tiny IoT satellite into Earth’s Orbit [32]. NASA named the centimeter-scale satellites sprites or ChipSats. The main purpose of the IoT satellites is to perform research activities. More than hundred of them are already in orbit by the spring 2019. First confirmation signals had been received the back by next day. By enabling communication between the satellites, they would like to demonstrate how the satellites can work together. This is necessary if they eventually operate in a swarm. The launching of TechEdSat-5 nanosatellite, which is a Technical Education Satellite-5, is a specific example of the application of IoT in space [33]. The TechEdSat-5 nano-satellite is a 3U CubeSat which is sometimes also called as TES5. The satellite is developed students from San Jose State University, the University of Idaho, and NASA’s Ames research center. It is developed by students of. The main objectives of the TES-5 are to establish a better uncertainty analysis for eventually controlled flight in earth thermosphere. It performs an in-depth comparison of the TES-3 and TES-4 concerning important uncertainty variable of the thermosphere. It also improves the prediction of location re-entry while providing model for return technology from orbital platforms. Moreover, it provides the experimental investigation of independent TDRV-based missions planetary travel. Furthermore, it provides important data to an on-orbit tracking device, which possibly enhance the prediction of discharged debris from the ISS [34]. Lander to Mars-rover communication may require better connectivity in terms of QoS, latency, reliability, and range. Whereas, for environmental monitoring requirements, it can be satisfied with unlicensed LoRa-based IoT equipment for environmental parameters including temperature, humidity, soil content, wind direction and etc. Moreover, the connectivity between the two technologies could further increase the possibility of more types of device interconnection at various locations of the planet and times of the mission. This gives the mission further possibility of exploring more information about the target planet in a single mission. Moreover, the same mission may have single backbone connectivity from landers to the geostationary satellite station or directly to earth stations. These with the interoperability of IoT networks, the collected information using various technologies could be forwarded through a single interconnection point. Moreover, this enables the processing of each data collected from each type of IoT device at some aggregation point such as edge computing. The processing of the collected data reduces the amount of integrated data for efficient transmission and interoperability of IoT technologies could enables this possibility [35].

9 Softwarization in Satellite and Interplanetary Networks

215

Edge computing has also seen its way in space demanding a new way of designing and transporting satellites. The challenges and functions of edge in space application demand are presented in [36]. A lot of IoT applications have stringent requirements that are impossible to meet with the traditional cloud computing techniques [37]. Space is not free from adverse competitive animosity which could result in security concerns between major space players in the race for space exploration. IoT devices implanted in space for measurement and other exploration activity could be attacked or hijacked by the adversary. Unattended access or hijacking of a single IoT device or robot or wireless sensor network may result in unintended consequences, in which the sensors could potentially be placed on territory accessible with competitors or adversaries. Therefore, the security mechanism for IoT devices and connectivity networks is of the essence. In [24], considering IoT in space applications, the authors discuss adaptive feedback-supported communication. The suggested technique is to minimize the amount of data in the transmission from the wireless sensors making the task more difficult or impossible for the adversarial observer to intercept. In this sense, it is to take advantage of hiding the sources of wireless communication. In addition, this technique allows the energy savings if a decrease occurs in the transmitted signals from the source node. This allows for longer operational time from a spacecraft or wireless sensor node. In [38], a security framework is provided which is intended to provide support to IoT device producers. The author proposed a framework called IoT-HarPSecA (A Framework and Roadmap for Secure Design and Development of Devices and Applications in the IoT Space) [38]. IoT-HarPSecA has three main functionalities such as elicitation of security requirements, guidelines for secure development best practice, and a feature that supports peculiar lightWeight cryptographic algorithms for software and hardware implementations.

9.2.4 Artificial Intelligence for Space Applications Artificial intelligence would play a significant role in the massive Martian exploration in a range of areas. This includes the automatic controlling of the navigation of rovers in the Martian surface; areal maneuvering of Martian helicopters for various missions; Controlling of networking management system; performing analysis of the collected scientific experimental data; automated manufacturing of equipment, tools, chemical products (e.g CO.2 ), etc. Few works explore the application of AI for space exploration. For example, the authors in [39], presented a a technique based on reinforcement learning with a multi-objective approach for cognitive space communications. They presented a hybrid radio resource allocation, control, and management algorithm that leverages deep reinforcement learning neural networks with multi-objective. Communication management between system resources can be improved by observing the dependant variables resulting in conflicting goals which leads to a better performance. Another interesting work

216

S. T. Arzo et al.

is the data mining application of AI [40]. The authors presented an ML-based telemetry data mining of space missions. An application of AI in aerospace is presented in [41].

9.3 Network Softwarization and In-Network Intelligence for Teleoperation, Telerobotic and Telepresence in Massive Space Exploration Massive space exploration requires remote operation at least at the first stages of the exploration. Remote operation could be on the ground, air, under the sea, or in space. Operating in such an environment is difficult for humans through physical presence. Teleoperation, telepresence, and telerobotics are the main mechanize to solve these problems. Teleoperation is the operation of a system or machine at a distance. Telepresence on the other hand allows a person to feel as if he is present at a remote location than the actual location, to give the appearance of being present, or to have an effect, using telerobotics. Telerobotics is a branch of robotics that deals with the control of semi-autonomous robots from a distance mainly using wireless networks such as Wi-Fi, Bluetooth, the DSN, or tethered connections. In this subsection, we review the three aspects providing the most important work and progress in literature in the context of advanced computing and communication networking from a space application perspective.

9.3.1 Teleoperation Teleoperation has various applications such as remote surgery/telesurgery, military and defensive applications, security applications, underwater vehicles navigation, forestry, mining applications, and space applications. Here our main focus is on space applications. To enable humans to travel a vast distance to space and operate a device or a vehicle in space demands several resources, and suitable conditions in the device’s vicinity, specifically for the currently targeted moon and Mars exploration. In the case of the sun and other plants, it could be impossible to physically visit with humans or robots with ordinary equipment due to the extreme heat. Hence, it is more appropriate and efficient to use teleoperated devices such as specially designed rovers and extraterrestrial unmanned aerial vehicles (UAVs). The recent successful landing of NASA’s perseverance is a great example that has also transported a UAV called ingenuity. Perseverance rover carried ingenuity to Mars and successfully released it to make a flight test. The aim of the Mars helicopter is to test the possibility of flying vehicles in the Martian atmosphere. Teleoperation is a long-sought subject in the research community. Considering the environment, operator, and task adaptive controllers systems for teleoperation, the authors in [42]

9 Softwarization in Satellite and Interplanetary Networks

217

presented an interesting survey. The authors classified the existing approaches that focus on the environment, operator, or task specific (EOT) information within the controller-structure called EOT-adapted controllers. For each method, they have also provided a study of the improvements and requirements. Based on their analysis, they indicated that several mechanisms need either the use of more sensors or require accurate model assumptions. The most challenging aspect of teleoperation in space application is the delay due to the vast distance between the operator and the target environment. It is because control data must be sent to and returned from the remote environment which poses a significant delay for real-time control. Therefore flight control is not observable from the controlling site in real-time. The authors in [43] presented a control technique for bilateral teleoperation of a pair of multi degree of freedom nonlinear robotic systems, in the case of persistent delays in the communication network. The presented technique uses a simple proportional derivative control. Master and slave robots are directly connected through spring and damper on the delayed imposed channels. By combining the controller passivity’s principle, Parseval’s identity, and the Lyapunov-Krasovskii method, they pacify the sum of the communication network and control part robustly delayed altogether. The idea relay on the assumption that delays are finite constants. Moreover, an upper bound for the round-trip delay is known. Despite the development in teleoperation technology, the old-fashioned method of teleoperation works based on the the human operator which the human operator does the exercise more all the times or does a less direct control. Developments in teleoperation has given rise to complex telepresence models in which the operator can observe its presence on the teleoperation part. It is worth mentioning that most researchers invented better teleoperation methods for complex tasks. Those complex tasks can be done by using the stereo vision and anthropomorphic manipulators using force feedback. S. Lichiardopol presented an advanced teleoperation systems with their various applications and the control problems that deals with the system control community [44]. Teleoperation is performed using a communication network. For space application, the DSN plays a central role in connecting the remote environment with the earth. An early study on the use of the Internet is presented in [45, 46]. In [45] the authors extend outcome on stable force reflecting teleoperation by having the timedelays and the transmission delays changes with time in unforeseeable trend. They showed that stability is maintained as a result of the systematic use of wave filters. In [46] paper attempted to address the challenges of time-dependent network time delay in force reflecting bilateral teleoperation. The idea is to address the Internet communication problems in terms of time delay due to bandwidth and physical distance. Moreover, the web-based teleoperation of a humanoid robot is presented in [47]. The paper incorporates an entire web and controller teleoperation to allow for various applications which control a robot. A fairly recent study based on the drone is conducted in [48]. The authors presented a technique called DronePick which is a collection of items and teleoperation of delivery. It is controlled by drone

218

S. T. Arzo et al.

by a portable tactile sight. An interesting book on teleoperation and human-robot interactions is presented in [41]. Due to the advance in computing, artificial intelligence, and communication technologies, the recent trend in teleoperation is to use such technologies in the operation process [49]. It is based on the idea that, instead of a direct human operation, efficiently transfer to distant locations, without being present. This represents a new challenge in this interconnected digital environment. In this approach, humans may experiment and execute actions in remote locations by a agent carrying immersive interfaces for physical sensation. Nevertheless, compromising skill-based performances, technological contingencies could impact human perception. Recommendations to the making of immersive teleoperation systems are provided taking into the account the findings of human factor studies. It is also followed by a sample assessment method. The authors expand a testbed to investigate intuitive problems that might influence job achievement while users works with the environment through immersive interfaces. The investigation of its impact on manipulation , navigation, and perception depend on achievement measurements and individual response. The objective is to reduce the impact of factors like system time delay, a reference frame, viewing field, or frame-rate to obtain the feel-of-telepresence. By dividing the flows of an immersive teleoperation system, they aimed at uncovering how human vision and interaction fidelity affects spatial cognition. To use teleoperation for massive space application and exploration, it requires a highly reliable, low latency, and high throughput communications system. The existing networks in 4G support limited functionality. However, current research in 5G, Beyond 5G, and 6G are considering the adaptation of networks to space. In particular, 3-dimensional (3D) networks have emerged as the future internet interconnecting existing network with space communications [10, 17]. For instance, in [10] discussed a possible Martian deployment of Cloud Radio Access Network (C- RAN) as a 3D network. Moreover, the End-to-end performance assessment of the 3D network considering Martian surface is an optimization of the advancement in the area that could propel the adaption of the 6G network for teleoperation.

9.3.2 Telerobotic As defined above teleoperation mean human control of remote sensors and actuators [50]. While telerobotic means human monitoring of semi-automatic systems at a distance. Moreover, it is assumed that surveillance control are equivalent terms to those that apply to teleoperation, or to a distance like as detection, manipulation, [51]. Supervisory control considers that the human operator, that could be acting remotely or at the vicinity of the equipment in the space environment, supervises a lower-level intelligence equipped in the teleoperator itself. The supervision is through sporadic monitoring and reprogramming the embedded intelligence whenever necessary for routine or emergencies operations. Telerobotics focus on

9 Softwarization in Satellite and Interplanetary Networks

219

the fact that the teleoperator transports enough effectors, sensors, and computer intelligence onboard to do basic duties intuitively. It could be by updated control programs over a telecommunications link. For example, in the case of ingenuity helicopter, the flight is conducted based on a flight plan uploaded from the earth ground which may consist of a series of waypoints that were telerobotically planned and scripted by operators at the jet propulsion laboratory. It is easy to explain why a supervisor-controlled robot is preferable to an autonomous robot or an astronaut able to perform the necessary tasks. At the moment, autonomous robots are neither smart enough nor reliable enough to accomplish the simplest and most routine tasks. Astronauts with a necessary radiation closing have been proven to be able perform tasks. However, the costs are extremely high, as are the risks associated with long-term and non-executed tasks. The first remotely controlled and operated robotic system in space environment is presented in [52]. There are numerous basic technologies designed by the space robot that they use ROTEX. Their features are technologies with multisensory gripper, shared autonomy using a local sensory feedback control concepts, and the simulation telerobotic ground station that is equipped with an advanced delay compensating SD-graphics. The article focuses on the method of programming the telesensor programming and the prediction simulation used to control the ground remotely. Very early work on human supervisory and control of the robotic system is discussed in [53]. An earlier study on the modern in space telerobots is presented in [54]. Including common requirements,design elements, and operational constraints, the authors examined the design issues for space telerobotics. They also identified the peculiar challenges for space telerobotics for terrestrial systems. Furthermore, they presented case studies of a number of various space telerobots while exploring the design of key side systems design and human-robotic interaction. They also outline telerobots and operational designs for future space exploration tasks. A review of space robotics for highly level science with space exploration is presented in [55]. Similar survey of space robotics is presented in [56]. The article outline a NASA survey to determine the current activities in space robotics while predicting future robotic possibilities in a nominal and intensive development strive. The space robotics analysis explored both planetary surface operations and space operations. Planetary surface operations includes mobility and most commonly associated with robotics and mobile robots exploration. The space operations focus on assembly, inspection, and maintenance. An older report on the development of automation and robotics in space exploration is presented in [57]. Similar to the remotelyoperated vehicles which humans explore the depths of oceans from the top, NASA is considering how a similar approach could help astronauts explore other worlds [51]. On June 17 and July 26, 2013, the Surface Telerobotics exploration concept is tested by NASA. In the test, an astronaut in an orbiting spacecraft remotely operated a robot on another planetary environment. For the future, astronauts orbiting other planetary bodies could use this approach to perform work on the surface using robotic avatars. This could be on Mars, asteroids, or the moon.

220

S. T. Arzo et al.

9.3.3 Telepresence Telepresence is the sensing and display technology that displays remote locations to the users such that they feel they are physically there. It is to make the operator feels as if he is actually present at the remote working site. If a enough information such as vision, sound, force are collected from the teleoperator site to the operator, then using the reconstruction techniques it is possible form the operator to feel physically present on the site. A simple camera based monitoring could creates some level of physical presence. However, typically, a more advanced and sophisticated system is used to recreate telepresence. The usual mechanism to produce telepresence are to enable the cameras to follow the operator’s head movements along with other input sensing equipment such as stereo vision, sound feedback, force feedback, and tactile sensing. In providing a more accurate telepresence, all human senses should be communicated from the remote teleoperator location to the operator site. Caldwell presented an interesting example of multi-sense telepresence. The proposed system supports multiple sensing input feedbacks such as stereo-vision and stereo-hearing, head tracking, force, tactile, temperature, and pain. The vision, hearing, and touch senses are comparatively simpler to transmit. However, smell and taste are very complicated. However, these two senses are not that much necessary for machine teleoperation.

9.3.4 Augmented Telerobotic Augmented presence/Augmented reality: It is a combination of real-world sensor information and virtual reality. An interesting example of this is an actual camera image with added computer-generated virtual information. In augmented teleoperation , The operator interface is in charge of generating virtual fixtures to improve the teleoperation accuracy. It is similar to virtual presence or virtual reality. Augmented teleoperation is similar to telepresence, except the environment where the operator feels to be present. A computer generates the sensor information is artificially. In tele-autonomy, the robot’s autonomous behaviors along with human commands make remote operation efficient. A survey of augmented reality is presented in [58]. An early work on telerobotic control using augmented reality is discussed in [59]. the use of technology that stimulates the senses of touch and motion is called haptics in telerobotics. Especially in preforming remote operation or computer based simulation of the sensations that could be felt by an operator interacting with physical objects. This requires research in the following fields: robotic hardware, hand controller, teleoperators with considering time delay. In [60], focusing on the control research, the aspects of haptics in telerobotics are discussed. In [61], an application of augmented reality for human-robot communication is presented. A relatively recent study on the design of an augmented telerobotic showcase system and its potential security concerns [62]. The aim of augmented

9 Softwarization in Satellite and Interplanetary Networks

221

reality (AR) is to solve the critical problems of network delay which may lead to teleoperation instability. Such problem can be mitigated by utilizing the concepts of superimposing virtual objects onto the real video image of the workspace which enables to reconstruct a simulation plan in the local machine. An interesting recent work to improve the collocated robot teleoperation with augmented reality is discussed in [63]. Despite significant progress in human-robot interaction techniques, there are issues such as natural and intuitive interaction and communication costs. To mitigate these limitations, the authors proposed RoSTAR(ROS-based Telerobotic Control via Augmented Reality) [64]. RoSTAR is an open-source human-robot interaction system using robot operating system and AR. In the article, a comprehensive model to augment a stereo-vision system along with the AR is presented.

9.4 Teleoperation Using Edge Computing Remote controlling of a networked system has been studied as discussed above. However, recent advancements in networking through network softwarization and automation, cloud computing, edge computing, machine learning, IoT, UAV, and automation have instigated the need for a new approach considering the current advancement in these cross multidisciplinary domains [65]. Moreover, the target of this literature server is space exploration martian and moon exploration in particular. Edge computing is a new computing technology aiding the responsiveness, scalability, and reliability of terrestrial computing and IoT based sensing networks in space exploration. Edge computing can mitigate the long distance problem between the processing servers and the end users by bringing the resources closer to the end users specifically in space applications. An interesting work on orbital edge computing is proposed in [66], presenting conceptual definition and characterization. They described power and software optimizations for the orbital edge. They also discussed the use of formation flying to parallelize computation in the case of space application. Since the concept of edge computing in space is relatively new, there are few works on the deployment for deep space exploration. However, an application based on space edge computing is also discussed in [67]. The authors presented a real time based motion control techniques utilizing measured latency value on edge computing. Similarly, the work on optimized control design for connected cruise control using edge computing, caching, and control [68] is used for remote operation on earth such as Arctic, and marine environment exploration. The paper describes an optimal control design for the system that use edge controllers with respect to communication latency with computing, caching, and control capabilities. It models the motion dynamics of every vehicle in the platoon. Then it formulates a linear quadratic optimization problem with regard to the network delay and the sampling period. In minimizing the deviations of the vehicle’s

222

S. T. Arzo et al.

motion direction and speed, the control strategy is to use backward recursion in solving iteratively. A recent interesting work using satellite for edge computing for IoT in aerospace is presented by Arzo et al. [10]. They propose converting the legacy satellite into a space based edge computing site. This enables to automatically upload and download software in orbit, to flexibly and efficiently share on-board computing resources while providing services coordinating with the legacy cloud computing [69]. They also provided the hardware structure along with the software architecture of the satellite. The work in [70] discussed the application of edge intelligent computing in satellite IoT. Similar work with a focus on latency and energy consumption optimization for mobile edge computing on improved SAT-IoT networks is presented in [71]. The authors in [72] presented a survey on the application of edge computing considering IoT. An interesting recent work for industrial remote control application is presented in [73]. The paper explored the use of edge computing for multi-tier industrial control system.

References 1. Solar System Exploration, Available Online At: https://solarsystem.nasa.gov/missions/? order=launch_date+desc&per_page=50&page=0&search=&fs=&fc=&ft=&dp=&category=, Accessed 26 Feb 2021 2. Chang’e-5: China’s Moon Sample Return Mission, Available Online At: https://www.jpl.nasa. gov/edu/news/2017/8/29/the-farthest-operating-spacecraft-voyagers-1-and-2-still-exploring40-years-later/, Accessed 20 Feb 2021 3. Artemis:Human’s Return to the Moon, Available Online At: https://www.nasa.gov/specials/ artemis/, Accessed 20 Feb 2021 4. Chang’e-5: China’s Moon Sample Return Mission, Available Online At: https://www. planetary.org/space-missions/change-5, Accessed 20 Feb 2021 5. China’s 1st Mars rover ’Zhurong’ lands on the Red Planet, Available Online At: https://www. space.com/china-Mars-rover-landing-success-tianwen-1-zhurong, Accessed 28 Feb 2021 6. ExoMars to Take off for the Red Planet in 2022, Available Online At: https://www.esa.int/ Newsroom/Press_Releases/ExoMars_to_take_off_for_the_Red_Planet_in_2022, Accessed 26 Feb 2021 7. Touchdown NASAs Mars Perseverance Rover Safely Lands on Red Planet, Available Online At: https://mars.nasa.gov/news/8865/touchdown-nasas-Mars-perseverance-roversafely-lands-on-red-planet/, Accessed 28 Feb 2021 8. Nokia Selected by NASA to Build First Ever Cellular Network on the Moon, Available Online At: https://www.nokia.com/about-us/news/releases/2020/10/19/nokia-selectedby-nasa-to-build-first-ever-cellular-network-on-the-moon/, Accessed 20 Feb 2021 9. P. Skarin, W. Tärneberg, K. Årzen, M. Kihl, Towards mission-critical control at the edge and over 5G, in 2018 IEEE International Conference on Edge Computing (EDGE) (2018), pp. 50–57. https://doi.org/10.1109/EDGE.2018.00014 10. S.T. Arzo, R. Bassoli, F. Granelli, F.H.P. Fitzek, Multi-agent based autonomic network management architecture. IEEE Trans. Netw. Serv. Manag. 18(3), 3595–3618 (2021). https:// doi.org/10.1109/TNSM.2021.3059752 11. S.T. Arzo, D. Scotece, R. Bassoli, F. Granelli, L. Foschini, F.H.P. Fitzek, A new agent-based intelligent network architecture. IEEE Commun. Stand. Mag. 6(4), 74–79 (2022). https://doi. org/10.1109/MCOMSTD.0001.2100053

9 Softwarization in Satellite and Interplanetary Networks

223

12. C. Bergstrom, S. Chuprun, S. Gifford, G. Maalouli, Software defined radio (SDR) special military applications, in MILCOM 2002. Proceedings, vol. 1 (2002) pp. 383–388. https://doi. org/10.1109/MILCOM.2002.1180472 13. S. Bonafini, C. Bianchi, F. Granelli, C. Sacchi, A reconfigurable multi-modal SDR transceiver for cubeSats, in 2021 IEEE Aerospace Conference (50100) (2021), pp. 1–12. https://doi.org/ 10.1109/AERO50100.2021.9438235 14. What is the Deep Space Network? Online Available At: https://www.nasa.gov/directorates/heo/ scan/services/networks/deep_space_network/about. Accessed: 26 Feb 2021 15. , Deep Space Network, Online Available At https://eyes.nasa.gov/dsn/dsn.html. Accessed: 26 Feb 2021 16. H. Yao, L. Wang, X. Wang, Z. Lu, Y. Liu, The space-terrestrial integrated network: an overview. IEEE Commun. Mag. 56(9), 178–185. https://doi.org/10.1109/MCOM.2018.1700038 17. B. Stefano, S. Claudio, B. Riccardo, K. Koteswararao, G. Fabrizio, H.P.F. Frank, End-toend performance assessment of a 3D network for 6G connectivity on Mars surface. Comput. Netw. 213 (2022). ISSN: 1389-1286. https://doi.org/10.1016/j.comnet.2022.109079 https:// www.sciencedirect.com/science/article/pii/S1389128622002171) 18. D. Sikeridis, E.E. Tsiropoulou, M. Devetsikiotis, S. Papavassiliou, Context-aware wirelessprotocol selection in heterogeneous public safety networks. IEEE Trans. Veh. Technol. 68(2), 2009–2013 (2018) 19. C. Sacchi, S. Bonafini, From LTE-A to LTE-M: a futuristic convergence between terrestrial and martian mobile communications, in 2019 IEEE International Black Sea Conference on Communications and Networking (BlackSeaCom) (2019), pp. 1–5. https://doi.org/10.1109/ BlackSeaCom.2019.8812825 20. A. Daga, G.R. Lovelace, D.K. Borah, P.L. De Leon, Terrain-based simulation of IEEE 802.11a and b physical layers on the martian surface. IEEE Trans. Aerosp. Electron. Syst. 43(4), 1617– 1624 (2007). https://doi.org/10.1109/TAES.2007.4441762 21. D. Sikeridis, E. EleniTsiropoulou, M. Devetsikiotis, S. Papavassiliou, Self-adaptive energy efficient operation in UAV-assisted public safety networks, in 2018 IEEE 19th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), (2018), pp. 1–5. https://doi.org/10.1109/SPAWC.2018.8446007 22. V. Chukkala, P. De Leon, S. Horan, V. Velusamy, Modeling the radio frequency environment of Mars for future wireless, networked rovers and sensor Webs, in 2004 IEEE Aerospace Conference Proceedings (IEEE Cat. No.04TH8720), vol. 2 (2004), pp. 1329–1336. https:// doi.org/10.1109/AERO.2004.1367731 23. A. Babuscia, D. Divsalar, K. Cheung, CDMA communication system for Mars Areostationary relay satellite, in 2017 IEEE Aerospace Conference (2017), pp. 1–10 https://doi.org/10.1109/ AERO.2017.7943941 24. P. Farkaš, Adaptive feedback supported communication for IoT and space applications, in 2019 NASA/ESA Conference on Adaptive Hardware and Systems (AHS) (2019), pp 61–64. https:// doi.org/10.1109/AHS.2019.00005 25. Wi-Fi Enables Next Generation Space Exploration, Available Online At: https://www.wi-fi. org/download.php?file=/sites/default/files/private/Wi-Fi_in_Space.pdf, Accessed 25 Feb 2021 26. K. Routh, T. Pal, A survey on technological, business and societal aspects of Internet of Things by Q3, 2017, in 2018 3rd International Conference On Internet of Things: Smart Innovation and Usages (IoT-SIU) (2018), pp. 1–4. https://doi.org/10.1109/IoT-SIU.2018.8519898 27. P. Yadav, S. Vishwakarma, Application of internet of things and big data towards a smart city, in 2018 3rd International Conference On Internet of Things: Smart Innovation and Usages (IoT-SIU) (2018), pp. 1–5. https://doi.org/10.1109/IoT-SIU.2018.8519920 28. A. Zeinab, E. Mona, S. Dimitrios, D. Michael, Internet of things-enabled passive contact tracing in smart cities. Int. Things 100397 (2021) 29. D. Sikeridis, B.P. Rimal, I. Papapanagiotou, M. Devetsikiotis, Unsupervised crowd-assisted learning enabling location-aware facilities. IEEE Int. Things J. 5(6), 4699–4713 (2018) 30. D. Sikeridis, E.E. Tsiropoulou, M. Devetsikiotis, S. Papavassiliou, Energy-efficient orchestration in wireless powered internet of things infrastructures. IEEE Trans. Green Commun. Netw. 3(2), 317–328 (2018)

224

S. T. Arzo et al.

31. J. Sprague, NASA’s Internet of Things Lab. Online Available At: https://internet-of-things. cioreview.com/cxoinsight/nasa-s-internet-of-things-lab--nid-13174-cid-133.html 32. Stanford and NASA Launch Tiny IoT Satellites Into Earth’s Orbit, Online Available At: https://theiotmagazine.com/stanford-and-nasa-launch-tiny-iot-satellites-into-earths-orbit9e5f92487500, Accessed 20 Feb 2021 33. TechEdSat-5 (Technical Education Satellite-5), Online Available At: “https://directory. eoportal.org/web/eoportal/satellite-missions/t/techedsat-5. Accessed 25 Feb 2021 34. Is the IoT in Space About to Take Off? Online Available At: https://www.networkworld.com/ article/3315736/is-the-iot-in-space-about-to-take-off.html. Accessed 20 Feb 2021 35. S.T. Arzo, F. Zambotto, F. Granelli, R. Bassoli, M. Devetsikiotis, F.H.P. Fitzek, A translator as virtual network function for network level interoperability of different IoT Technologies, in The 2021 IEEE 7th International Conference on Network Softwarization (NetSoft) (2021), pp. 416–422. https://doi.org/10.1109/NetSoft51509.2021.9492677 36. H. Jungha et al., IoT Edge Challenges and Functions (2022). Last Updated 11 Jan, 2022, Accessed: 09 May 2022, Online Available At: https://datatracker.ietf.org/doc/draft-irtf-t2trgiot-edge/ 37. D. Sikeridis, I. Papapanagiotou, B.P. Rimal, M. Devetsikiotis, A comparative taxonomy and survey of public cloud infrastructure vendors (2017). https://doi.org/10.48550/ARXIV.1710. 01476, Online Available At: https://arxiv.org/abs/1710.01476 38. M.G. Samaila, J.B.F. Sequeiros, T. Simões, M.M. Freire, P.R.M. Inácio, IoT-HarPSecA: a framework and roadmap for secure design and development of devices and applications in the IoT space. IEEE Access 8, 16462–16494. https://doi.org/10.1109/ACCESS.2020.2965925 39. P.V.R. Ferreira, R. Paffenroth, A.M. Wyglinski, T.M. Hackett, S.G. Bilén, R.C. Reinhart, D.J. Mortensen, Multi-objective reinforcement learning-based deep neural networks for cognitive space communications, in 2017 Cognitive Communications for Aerospace Applications Workshop (CCAA) (2017), pp. 1–8. https://doi.org/10.1109/CCAAW.2017.8001880 40. A.E. Hassanien, A. Darwish, S. Abdelghafar, Machine learning in telemetry data mining of space mission: basics, challenging and future direction. Artif. Intell. Rev. 53, 1573–7462 (2020) https://doi.org/10.1007/s10462-019-09760-1 41. D. Lary, Artif. Intell. Aerosp. (2010). ISBN:978-953-7619-96-1, https://doi.org/10.5772/6941, https://www.intechopen.com/chapters/6832 42. P. Carolina, P. Angelika, B. Martin, A survey of environment-, operator-, and taskadapted controllers for teleoperation systems. Mechatronics 20(7), 787–801 (2010). Special Issue on Design and Control Methodologies in Telerobotics. ISSN: 0957-4158. https://doi. org/10.1016/j.mechatronics.2010.04.005, url: https://www.sciencedirect.com/science/article/ pii/S0957415810000735 43. D. Lee, M.W. Spong, Passive bilateral teleoperation with constant time delay. IEEE Trans. Robot. 22(2), 269–281 (2006). https://doi.org/10.1109/TRO.2005.862037 44. S. Lichiardopol, A Survey on Teleoperation (Technische UniversiteitEindhoven, 2007). eprint: 2007.155 https://pure.tue.nl/ws/files/4419568/656592.pdf 45. G. Niemeyer, J.-J.E. Slotine,Towards force-reflecting teleoperation over the internet, in Proceedings 1998 IEEE International Conference on Robotics and Automation (Cat. No.98CH36146), vol. 3 (1998), pp. 1909–1915. https://doi.org/10.1109/ROBOT.1998.680592 46. N. Chopra, M.W. Spong, S. Hirche, M. Buss, Bilateral teleoperation over the internet: the time varying delay problem, in Proceedings of the 2003 American Control Conference, vol. 1 (2003), pp. 155–160. https://doi.org/10.1109/ACC.2003.1238930 47. C.L. Fok, F. Sun, M. Mangum, A. Mok, B. He, L. Sentis, Web Based Teleoperation of a Humanoid Robot (2016). https://doi.org/10.48550/ARXIV.1607.05402 https://arxiv.org/abs/ 1607.05402, arXiv.org perpetual, non-exclusive license 48. R. Ibrahimov, E. Tsykunov, V. Shirokun, A. Somov, D. Tsetserukou, DronePick: object picking and delivery teleoperation with the drone controlled by a wearable tactile display, in 2019 28th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) (2019), pp.1–6. https://doi.org/10.1109/RO-MAN46459.2019.8956344

9 Softwarization in Satellite and Interplanetary Networks

225

49. L. Almeida, P. Menezes, J. Dias, Interface transparency issues in teleoperation. Appl. Sci. 10(18) (2020). https://www.mdpi.com/2076-3417/10/18/6232, ISSN: 2076-3417, https://doi. org/10.3390/app10186232 50. T.B. Sheridan, Teleoperation, telerobotics and telepresence: a progress report. Control Eng. Pract. 3(2), 205–214 (1995). ISSN: 0967-0661, https://doi.org/10.1016/0967-0661(94)00078U, http://www.sciencedirect.com/science/article/pii/096706619400078U 51. Station Astronauts Remotely Control Planetary Rover From Space, Available Online At: https://www.nasa.gov/mission_pages/station/research/news/rover_from_space, Accessed 20 Feb 2021 52. G. Hirzinge, B. Brunner, J. Dietrich, J. Heindl, ROTEX-the first remotely controlled robot in space, in Proceedings of the 1994 IEEE International Conference on Robotics and Automation, vol. 3 (1994), pp. 2604–2611. https://doi.org/10.1109/ROBOT.1994.351121 53. T. Sheridan, Human supervisory control of robot systems, Proceedings 1986 IEEE International Conference on Robotics and Automation, vol. 3 (1986), pp. 808–812. https://doi.org/10. 1109/ROBOT.1986.1087506 54. F. Terrence, R.Z. Jenniferand, C. Nancy, M. Andrew, L.A. David, Space telerobotics: unique challenges to human–robot collaboration in space. Rev. Human Factors Ergon. 9(1), 6–56 (2013). https://doi.org/10.1177/1557234X13510679, https://doi.org/10. 1177/1557234X13510679, https://doi.org/10.1177/1557234X13510679 55. Y. Gao, S. Chien, Review on space robotics: toward top-level science through space exploration. Sci. Robot. 2(7), eaan5074 (2017) . https://doi.org/10.1126/scirobotics. aan5074. https://robotics.sciencemag.org/content/2/7/eaan5074, https://robotics.sciencemag. org/content/2/7/eaan5074.full.pdf 56. L. Pedersen, D. Kortenkamp, D. Wettergreen, I. Nourbakhsh, D. Korsmeyer, A Survey of Space Robotics (2003). https://www.researchgate.net/publication/2874965 57. M. Montemerlo, NASA’s automation and robotics technology development program, in Proceedings. 1986 IEEE International Conference on Robotics and Automation, (1986), pp. 977–986. https://doi.org/10.1109/ROBOT.1986.1087562 58. R.T. Azuma, A survey of augmented reality, vol. 6, no. 4 (MIT Press, Cambridge, 1997), pp. 355–385. ISSN: 1054–7460, https://doi.org/10.1162/pres.1997.6.4.355, Presence: Teleoper Virtual Environ 59. P. Milgram, A. Rastogi, J.J. Grodski, Telerobotic control using augmented reality, in Proceedings 4th IEEE International Workshop on Robot and Human Communication, pp. 21–29 (1995). https://doi.org/10.1109/ROMAN.1995.531930 60. C. Preusche, G. Hirzinger, Haptics in telerobotics. Visual Comput. 23 (2007). https://doi.org/ 10.1007/s00371-007-0101-3 61. P. Milgram, S. Zhai, D. Drascic, J. Grodski, Applications of augmented reality for human-robot communication, in Proceedings of 1993 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS ’93), vol. 3 (1993), pp. 1467–1472. https://doi.org/10.1109/IROS. 1993.583833 62. , S. Islam, M. Al-Mohammed, R. Islam, M. Bhattacharya, T.M. Alkharobi, S.M. Buhari, Design of an augmented telerobotic stereo vision system and associated security concerns, in 2015 IEEE 10th Conference on Industrial Electronics and Applications (ICIEA) (2015), pp.1809– 1814. https://doi.org/10.1109/ICIEA.2015.7334405 63. H. Hedayati, M. Walker, D. Szafir, Improving collocated robot teleoperation with augmented reality, in Proceedings of the 2018 ACM/IEEE International Conference on Human-Robot Interaction. HRI ’18 (Association for Computing Machinery, New York, 2018), pp. 78–86. ISBN:9781450349536, https://doi.org/10.1145/3171221.3171251 64. C. Xue Er Shamaine, Y. Qiao, J. Henry, K. McNevin, N. Murray, RoSTAR: ROS-based telerobotic control via augmented reality, in 2020 IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP) (2020), pp. 1–6. https://doi.org/10.1109/MMSP48831. 2020.9287100

226

S. T. Arzo et al.

65. D. Sikeridis, E.E. Tsiropoulou, M. Devetsikiotis, S. Papavassiliou, Wireless powered public safety IoT: a UAV-assisted adaptive-learning approach towards energy efficiency. J. Netw. Comput. Appl. 123, 69–79 (2018) 66. B. Denby, B. Lucia, Orbital edge computing: machine inference in space. IEEE Comput. Archit. Lett. 18(1), 59–62 (2019). https://doi.org/10.1109/LCA.2019.2907539 67. Y. Koyasako, T. Suzuki, S. Kim, J. Kani, J. Terada, Real-time motion control method using measured delay information on access edge computing, in 2020 IEEE 17th Annual Consumer Communications Networking Conference (CCNC) (2020), pp. 1–4. https://doi.org/10.1109/ CCNC46108.2020.9045470 68. Z. Wang, Y. Gao, C. Fang, Y. Sun, P. Si, Optimal control design for connected cruise control with edge computing, caching, and control, in IEEE INFOCOM 2019 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS) (2019), pp. 1–6. https://doi.org/ 10.1109/INFOCOMWKSHPS47286.2019.9093766 69. D. Sikeridis, I. Papapanagiotou, B.P. Rimal, M. Devetsikiotis, A comparative taxonomy and survey of public cloud infrastructure vendors (2017). Preprint. arXiv:1710.01476 70. J. Wei, S. Cao, Application of edge intelligent computing in satellite internet of things, in 2019 IEEE International Conference on Smart Internet of Things (SmartIoT) (2019), pp. 85–91. https://doi.org/10.1109/SmartIoT.2019.00022 71. G. Cui, X. Li, L. Xu, W. Wang, Latency and energy optimization for MEC enhanced SATIoT networks. IEEE Access 8, 55915–55926 (2020). https://doi.org/10.1109/ACCESS.2020. 2982356 72. W. Yu, F. Liang, X. He, W.G. Hatcher, C. Lu, J. Lin, X. Yang, A survey on the edge computing for the internet of things. IEEE Access 6, 6900–6919 (2018). https://doi.org/10.1109/ACCESS. 2017.2778504 73. Y. Ma, C. Lu, B. Sinopoli, S. Zeng, Exploring edge computing for multitier industrial control. IEEE Trans. Comput. Aided Des. Integr. Circuits Syst. 39(11), 3506–3518 (2020). https://doi. org/10.1109/TCAD.2020.3012648

Chapter 10

Extraterrestrial Radio Access Network: The Road to Broadband Connectivity on Mars Stefano Bonafini, Claudio Sacchi, Fabrizio Granelli, Koteswararao Kondepu, Riccardo Bassoli, and Frank H. P. Fitzek

10.1 Introduction The exploration of Mars is a key topic in Space Science since early ’60s. The early missions, e.g. NASA Mariner series (1965–1971), were mostly based on orbiters passing over the planet. Such missions collected and transmitted precious information about both the chemical composition of the Martian atmosphere and the planet orography [1]. However, the scientific community expected more concrete and tangible data, directly acquired from the Mars surface. Indeed, since early ‘50s, there were a lot of theories and conjectures about the existence of residual iced water in some regions of the planet. It is well known that the presence of water is an indicator of the presence of some life forms. The turning point of Mars missions was in July 1976, when the first lander (Viking 1) successfully reached the Mars surface. This mission paved the road to the true exploration of Mars, upon the meaning of the verb “to explore” reported in [2]: “to look something in a careful way to learn more about it: to study or analyze (something)”. In the recent years, the

S. Bonafini · C. Sacchi () · F. Granelli Department of Information Engineering and Computer Science (DISI), University of Trento, Trento, Italy e-mail: [email protected]; [email protected]; [email protected] K. Kondepu Department of Computer Science and Engineering, Indian Institute of Technology Dharwad, WALMI Campus, Dharwad, India e-mail: [email protected] R. Bassoli · F. H. P. Fitzek Deutsche Telekom Chair of Communication Networks, Technische Universität Dresden, Dresden, Germany Centre for Tactile Internet with Human-in-the-Loop (CeTI), Dresden, Germany e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_10

227

228

S. Bonafini et al.

Mars exploration has been based on autonomous vehicles and robotic technologies. Among the vehicles currently operating on the Mars surface, a prominent position is taken by the rovers. Mars rovers are small vehicles that collect different type of information, mostly of geological nature [2]. The rovers can be regarded as an extension of the human, rather than autonomous agents [2]. We can mention, as example, the “Perseverance” rover, launched by NASA in 2020. As claimed in [3], the main job of “Perseverance” is to seek signs of ancient life and collect samples of rock and regolith (broken rock and soil) for possible return to Earth. Despite the outstanding results of unmanned missions, NASA and other national and international Space agencies are producing intensive efforts to launch a manned mission on Mars within the next two decades. By signing the Space Policy Directive-1, the President of the United States of the America directed the NASA Administrator: “to lead an innovative and sustainable program of exploration with commercial and international partners to enable human expansion across the solar system and to bring back to Earth new knowledge and opportunities. Beginning with missions beyond low-Earth orbit (LEO), the United States will lead the return of humans to the Moon for long-term exploration and utilization, followed by human missions to Mars and other destinations.” [4]. In particular, NASA has been committed to land American astronauts in the Lunar South Pole in 2024 with the claimed aim to develop the necessary technologies and capabilities for future missions on Mars, regarded as the true long-term objective of the campaign. The landing of astronauts on Moon and Mars should pave the road to the final ambitious goal to establish a sustainable and permanent human presence in the Deep Space. The requirements to sustain a crew for a continuous visit for up to 60 days at a time have been listed in [5]: • • • • •

Availability of adequate habitable volume and stowage. Deployment of environmental control and life support systems. Power pass-through for other elements. Thermal control. Availability of efficient surface-to-surface, Gateway-to-surface, and direct to Earth communications. • Presence on-site of galley, crew quarters, exercise equipment, medical workstations and a personal hygiene compartment. Fulfilling each of the above listed “things to do” is, by itself, a very challenging task. In this chapter, we partially deal with the communication needs, focusing our attention on the surface Martian network deployment. This aspect is of paramount importance in future manned missions on Mars. Indeed, the presence of an efficient surface network infrastructure will allow the exchange of real-time information, emergency messages, post-processing of data, navigation and remote control of rovers, landers and unmanned aerial vehicles (UAVs). Some experiments conducted by NASA demonstrated that delays and limited bandwidth in extra-terrestrial missions would produce frustration and uncertainty in the crew [6]. In this chapter, the authors will present an overview of the current status of Martian surface connectivity and the solutions proposed by the most recent literature

10 Extraterrestrial Radio Access Network: The Road to Broadband. . .

229

aimed at bridging the gap toward future broadband mobile networking on the Red Planet. The chapter is structured as follows: Sect. 10.2 will analyze the current situation of wireless connections on Mars (with a look to the near future), Sect. 10.3 will discuss some proposed solutions for future Martian networking, based on the support of local satellite constellations, Sect. 10.4 tries to analyze the feasibility of porting terrestrial standards like LTE and 5G on Mars. In Sect. 10.5, cuttingedge solutions based on 3D networks and C-RAN virtualizazion will be considered. Finally, chapter conclusion will be drawn in Sect. 10.6.

10.2 The Current (and Near Future) Picture of Martian Surface Connectivity A pictorial description of the current situation of Martian surface connectivity is shown in Fig. 10.1. The wireless landscape is populated by some rovers, landers and orbiters. Actually, a wireless sensor network (WSN) is deployed on ground that communicates with the rovers working like hubs and also provided by their own sensors [7]. The communication between sensors and rovers is managed by using the IEEE 802.15.4 Zigbee standard [8] or an adapted version of the IEEE 802.11 WLAN standard [7]. Even though the direct communication between rovers and landers with Earth is enabled, however it has been convened that the best way to save energy and augment efficiency is to exploit a relay connection with an orbiter. In [9], rover-to-orbiter connections have been studied in terms of channel modelling and performance. This communication landscape, quite similar to terrestrial Internet-ofThings (IoT), does not consider real-time transmission because only scientific data are acquired by surface sensors and one-way transmitted to Earth. In such a scenario, real-time is not a requirement. If a human crew landed today on Mars, it would find such a degree of connectivity that is clearly not adequate for a manned mission. However, the situation might improve in the near future. Indeed, some promising flying platforms, potentially capable of supporting payloads for readily-available surface communications, have been experimented on the Mars proximity. In 2018, two small satellites have been launched in the Martian atmosphere in the framework of MarsCubeOne (MarCO) mission [10]. The purpose of MarCO was to demonstrate the viability of CubeSat technology on Mars, in particular to perform a test about pushing the limits of miniaturized technology and seeing just how far it could have taken. From this viewpoint, the mission was accomplished, because the two satellites, namely EVE and WALL-E, served as communication relays during the InSight rover landing, transmitting data in near real-time at each stage of InSight descent. Besides this primary task, EVE sent some wonderful pictures from Mars, while WALL-E performed some radio science experiments. It is stated in [10] that future Martian CubeSat missions will be targeted at further and more advanced radio system experimentation.

230

S. Bonafini et al.

Fig. 10.1 Picture of the current situation of Martian communication scenario

Martian CubeSats will be complemented in the Sky segment by the Martian Reconnaisance Orbiter (MRO) [11], which is a multi-purpose spacecraft for planet exploration that can be also reused as low-orbit satellite. However, as claimed in [12], it is expected that future Martian communications will be supported also by a stationary-orbit relay satellite. Stationary orbits around Mars, known as aerostationary orbits, have similar characteristics as Earths geostationary orbits. The aerostationary satellite is located at 17,000 km above Mars surface and is always in the same place in the sky on Mars. Thus, it can receive data from the rovers/landers, low Mars orbiters, and CubeSats in vicinity of Mars, providing relay to an Earth station or working as backbone for in-situ surface networks. Another interesting experiment has launched an UAV (namely: the Mars helicopter Ingenuity [13]) just to verify the viability of such a technology in the Martian environment. Further experiments, foreseen in the near future, consider the real possibility of embarking on Martian drones light but vital payloads, i.e.: to survive the cruise to Mars, to autonomously charging themselves with their solar panels and to communicate to and from the helicopter via the Mars helicopter base station subsystem [13]. In Fig. 10.2, the picture of a possible “near future” scenario of Martian communications is drawn. Aerostationary satellites, cubesats, orbiters forms kind of “Sky layer” of the Martian network that can work as relay and backbone for the rovers and the landers staying on the planet surface. UAVs will communicate as well with sky nodes and surface nodes. In the scenario, two astronauts are stepping over the Martian soil. The fundamental question is: “Who

10 Extraterrestrial Radio Access Network: The Road to Broadband. . .

231

Fig. 10.2 Picture of the “near future” situation of Martian communication scenario

will give them the necessary bandwidth to stay connected?” We shall try to answer to such a question in the following chapter section.

10.3 Martian Communications Supported by Sky Connections As mentioned in Sect. 10.2, it is expected that some communication payloads will orbit around Mars at different altitudes. These payloads are obvious candidate to support future Martian connectivity with autonomy from Earth. A conceptual design of satellite-based Martian networking has been presented by Bell, Cesarone et al. in [14]. The in-flight element of the network are an aerostationary satellite and a constellation of low-altitude microsatellites, these last regarded as the proximity connectivity providers for integrated navigation and communication services to Martian vehicles and human crews (the aerostationary satellite is used for backhual of Martian nodes and long-haul to Earth). The microsatellites of [14] were targeted to 800 km altitude, near equatorial and high-inclination orbits. The constellation was designed to return 1 Gbit per Sol (Sol .= Martian day .= 24 terrestrial hours and 37 min), using 1 W power and omni-directional antennas. The reference bandwidths for transmission are UHF and X-band. The actual bit rate of such satellite connections is about 11.3 Kb/s, which is good to support voice, sensor data and localization data transmission.

232

S. Bonafini et al.

A more recent work dealing with satellite-based Martian connectivity [15] considers the use of Mars Reconnaisance Orbiter (MRO) as the space node providing both proximity link and relay to Earth. The related architecture looks simpler as compared to that of [14] and easier to be managed. Data-rates up to 2 Mb/s can be achieved by the MRO-rover link. The presence of astronauts on the Martian soil has not been considered in [15]. A more futuristic project for future Martian connectivity has been presented in [16]. In this paper, a physical network topology is described consisting of a high power ground station to communicate with orbiters. The ground station is then connected through a local wireless network to the manned installations on Mars. The units or habitats have Internet-like connections. Each of them has its own IPv6 router or switch, connected by Ethernet to multiple neighbors behind wall panels. An additional wireless network connects wireless devices (usually personal crew devices, but may also include sensors) to the network. This 802.11 service is provided by access points from three routers opposite of each other [16]. Such a network configuration is clearly inspired by similar terrestrial satellitebased networks bringing connectivity to small buildings for indoor interactive TV or Internet services. Indeed, such an arrangement has been proposed for a future Mars science station, where the mobility is essentially of a nomadic kind.

10.4 LTE Connections Operating on the Mars Soil: Would They Be Practicable? The solutions proposed in Sect. 10.3 and based on Sky-to-surface connections are not targeted at providing an effective mobile coverage on the Red planet. They are more customized to offer backhaul to rovers and landers rather than to allow to astronauts present on the Martian surface to readily communicate together and/or with Earth. Moreover, these approaches do not consider the possibility of proactive cooperation between landers, rovers, and orbiters, enabling on-site data exchange and distributed information processing. For these reasons, a couple of recent papers preliminarily investigated the feasibility of porting LTE connectivity on Mars by installing somewhere enodeBs and switching spots. In [17], a possible architecture for a “Martian LTE” mobile communication system has been proposed and pictorially described in Fig. 10.3, in particular focusing on a “Martian cell”. The eNodeB is installed in a lander, while operations of backhaul, long-haul and handover to other cells are managed by an aerial satellite. Then, the LTE uplink and downlink transmission have been tested in the Gusev and Hematite craters, by referring to the Martian propagation analysis of [18] and [19]. The main outcome of the analysis of [17] was that the LTE solution on Mars would be practicable, given the use of terrestrial equipment, in a cell range of 50 m. The analysis of [17] has been further extended in [20]. Considering the propagation impairments inherent to the rocky nature of Martian soil, a preliminary coverage

10 Extraterrestrial Radio Access Network: The Road to Broadband. . .

233

Fig. 10.3 Pictorial scheme of an hypotetical “Martian LTE” mobile communication cell: blue arrows and orange arrows depict uplink and downlink respectively (courtesy by Sacchi and Bonafini [17])

analysis of in-situ LTE Martian connectivity has been performed. The achieved results, related to Gusev1 and Gusev3 crater areas, demonstrated that high capacities of tenths of Mb/s are achievable at distances of 100 m (so, the hypothetical Martian cells should be conveniently small), while for distances of the order of 1000 m, the available capacity decreases a lot.

10.5 Advanced Solutions Based on 3D NTN and C-RAN The Martian connectivity solutions proposed in Sects. 10.3 and 10.4 lacks of efficiency. If satellites will be launched on the Martian atmosphere, Sky connections would be feasible, but the throughput provided by full-satellite connectivity is limited by distance and latency. The integration between aerial satellite with the orbiters and the Cubesats would be a solution to definitely improve QoS. The alternative solution considering the installation of LTE infrastructures on Mars might be critical from the energy consumption viewpoint. Terrestrial mobile networks rely on fixed infrastructures, like eNodeBs and mobile switching centers, continuously operated by the electrical power supply network. Unfortunately, on the Mars planet, only renewable power supply sources are available. Thus, the installation of an eNodeB on a rover or a lander would involve an unacceptable power consumption that might even compromise the vehicle functionalities. A completely different approach proposed in the most recent literature considers the deployment of 3D networks on Mars that are based on functional splitting among

234

S. Bonafini et al.

Fig. 10.4 The overall Martian 3D network and the related splitting options (courtesy by Bonafini et al. [23])

the different network nodes, some of them of aerial and even satellite kind. This is a typical arrangement belonging to the “beyond 5G” networking framework that, on the Earth, takes the name of NTNs. In [21], the potential of multilayered hierarchical NTNs in a 6G perspective are investigated. NTNs, made of integrated terrestrial nodes, UAVs, HAPS and small satellites, play a leading role in future networks by covering different verticals, including healthcare, intelligent transportation, public safety, and many others [21]. They can offer highly desirable features like: service continuity, service ubiquity and service scalability. When combined with functional splitting, NTNs can allow to deploy ad-hoc network configurations based on ondemand connectivity requirements, thus providing a very high degree of flexibility and reconfigurability. The combination of NTNs and functional splitting have been studied for terrestrial applications, like border monitoring, where this arrangement targets to flexibility, coverage and energy efficiency in scenarios where the installation of fixed terrestrial networking infrastructures is prevented by unsolvable logistic reasons [22]. The feasibility study of the different splitting options in 3D Martian networks has been presented in [23] by following an analytical approach. The Martian network proposed in [23] is made of four layers (see Fig. 10.4): • The surface layer, essentially made by the UEs. • The aerial layer that should take in charge remote radio unit (RU) functions of the base station. • A Very-Low Mars Orbit (VLMO) layer, based on constellations of CubeSats that will take in charge to run the distributed units (DUs) and centralized units (CUs) network functions. • a Low Mars Orbit (LMO) layer, with larger orbiters such as the Mars Reconaissance Orbiter (MRO), which will host the evolved packet core (EPC) of the network and will provide backhaul links with aerostationary satellites. The outcomes of the analysis of [23] fix some crucial parameters about the CubeSat orbit heights as a function of the allowed latency inherent to each

10 Extraterrestrial Radio Access Network: The Road to Broadband. . .

235

Fig. 10.5 Drag force imposed by Martian atmosphere to 1U, 6U and 12U CubeSats on VLMO with respect that of terrestrial low Earth orbit (LEO). The dotted horizontal lines mark the propulsion force applied by off-the-shelf, Vacco’s MiPS and JPL hybrid thrusters (HT).(courtesy by Bonafini et al. [23])

splitting option and considering the density of the Martian atmosphere that is considerably lower than that measured on the Earth. The maximum altitude of the CubeSat to cope with the most stringent latency constraints issued by 3GPP for 5G splitting option is 75 km, under ideal delay conditions (0.25 ms). The approximate atmospheric drag force and, therefore, the required propulsion force required to maintain the CubeSat in orbit had been computed in [23] and shown in Fig. 10.5. In such a figure, it is evident that existing propulsion systems cannot support so low altitudes on Earth, but things are different on Mars, where 12U Cubsats with powerful thrusters might orbit even at orbit altitudes of 20 km, while 6U MarCO satellites with their thrusters could support an altitude of about 110 km, which would involve a slight delay increase up to 0.37 ms. In Fig. 10.6, the available session time for the connection CubeSat-UAV vs. orbit altitude is shown along with the variation of CubeSat velocity. Lowering the orbit altitude, we can noticeably increase the session time. The price to be paid is in terms of resources spent to correct the trajectory and of reduced coverage.

236

S. Bonafini et al.

Fig. 10.6 Variation of CubeSat velocity and available communication window from UAV to CubeSat for a CubeSat altitude ranging from 35 km to 75 km (courtesy by Bonafini et al. [23])

In another very recent work [24], requirements and solutions for the fronthaul link of the Martian 3D network have been discussed. In our case, the fronthaul will be implemented by a point-to-point connection from the UAV to the CubeSat. In [24], available front-end technologies have been matched with the fronthaul requirements in terms of transmission bitrate related to each 5G splitting options. Assuming the 5G sampling rate of 153.6 MHz, considering the use of commercial analogic front-end components capable at achieving a net spectral efficiency of 4.61 bit/symbol, and a roll-off factor of pulse shaping of 0.3, the fronthaul bandwidth requirements of the various splitting options are shown in Table 10.1. As mentioned in [23], the needed bandwidth resources can be found in X-band, Ku-band and Kaband that are negligibly impaired by Martian atmospheric attenuation. The end-to-end performance evaluation of the Martian 3D network infrastructure based on functional splitting has been thoroughly assessed in [25]. The evaluation has been obtained by network emulations performed on the Open Air Interface (OAI) open-source platform (https://openairinterface.org). The splitting option 7.1 has been fully emulated, considering different values of front-haul delay and, therefore, of CubeSat orbit altitude. The numerical results that better highlight the achievements of [25] are shown in Fig. 10.7. In such a graphic, the percentual

10 Extraterrestrial Radio Access Network: The Road to Broadband. . . Table 10.1 Fronthaul link requirements for the different considered 5G splitting options

Splitting option 8 7.1 7.2 7.3 6

Fronthaul bitrate 6.14 Gb/s 2.6 Gb/s 447 Mb/s 396 Mb/s 104 Mb/s

237 Required bandwidth 1.73 GHz 733.2 MHz 126 MHz 111.6 MHz 29.6 MHz

Fig. 10.7 E2E packet loss and delay versus fronthaul latency, fixing the backhaul latency (courtesy by Bonafini et al. [25])

packet loss and the packet delay have been plotted versus the fronthaul delay by fixing the backhaul delay (i.e. the delay between CubeSat and Orbiter). It is easy to be noticed an inflection point toward 0.65 ms, above which both packet loss and delay dramatically ramp up, while staying behind such a point, the network system yields to very good performance. This leads involves the possibility of increasing the CubeSat orbital altitude up to about 200 km, thus providing a higher degree of freedom to the whole architectural design.

10.6 Conclusion In this chapter, some solutions proposed to bring connectivity on the Martian surface in the perspective of the landing of human crews and the settlement of future Martian stations have been surveyed. The current picture of Martian networking is very far to cope with the connectivity requirements imposed by the presence of manned personnel on the planet surface. In the future, aerial communication payloads will be launched at different orbits and this will provide useful bandwidth for broadband connections. The installation of terrestrial mobile network infrastructure on Mars is

238

S. Bonafini et al.

prevented by logistic and energy consumption reasons. Therefore, 3D network and functional splitting would allow to exploit in efficient manner the available payload and bandwidth resources in a renewed “5G and beyond” vision transferred from Earth to Mars.

References 1. Missions to Mars, The Planetary Society (2018). http://www.planetary.org/explore/spacetopics/space-missions/missions-to-mars.html 2. K. Zawieska, B.R. Duffy, Social exploration: Mars rovers, in Proceedings of the 2014 ACM/IEEE international conference on Human-robot interaction (HRI ’14), New York, NY, USA (2014), pp. 324–325 3. Mars Perseverance Press Kit (2021). https://mars.nasa.gov/resources/25529/mars-2020perseverance-landing-press-kit/ 4. K.G. Boggs, K. Goodliff, D. Elburn, Capabilities development: from international space station and the Moon to Mars, in 2020 IEEE Aerospace Conference (2020), pp. 1–10 5. G. Flores, D. Harris, R. McCauley, S. Canerday, L. Ingram, N. Herrmann, Deep space habitation: establishing a sustainable human presence on the moon and beyond, in 2021 IEEE Aerospace Conference (50100) (2021), pp. 1–7 6. NASA is Laser-focused on Deep Space Communication (2015). https://www.nasa.gov/ 7. X. Hong, M. Gerla, H. Wang, L. Clare, Load balanced, energy-aware communications for Mars sensor networks, in Proceedings, IEEE Aerospace Conference (2002), pp. 3–3 8. R. Pucci, L.S. Ronga, E. Del Re, D. Boschetti, Performance evaluation of an IEEE802.15.4 standard based wireless sensor network in Mars exploration scenario, in 2009 1st International Conference on Wireless Communication, Vehicular Technology, Information Theory and Aerospace & Electronic Systems Technology (2009), pp. 161–165 9. S. Li, D.T.H. Kao, A.S. Avestimehr, Rover-to-orbiter communication in Mars: taking advantage of the varying topology. IEEE Trans. Commun. 64(2), 572–585 (2016) 10. In Depth: MarCO (Mars Cube One) (2021) https://solarsystem.nasa.gov/missions/mars-cubeone/in-depth/. Date last accessed on 11 Oct 2022 11. D. Edwards Jr., R. DePaula, Key telecommunications technologies for increasing data return for future Mars exploration. Acta Astronautica 61(1), 131–138 (2007) 12. A. Babuscia, D. Divsalar, K.-M. Cheung, CDMA communication system for mars areostationary relay satellite, in 2017 IEEE Aerospace Conference (2017), pp. 1–10 13. Jet Propulsion Laboratory, Ingenuity Mars Helicopter: Landing Press Kit (2021). https://www. jpl.nasa.gov/news/presskits/ingenuity/landing/, Date last accessed on 18 Oct 2022 14. D.J. Bell, R. Cesarone, T. Ely, C. Edwards, S. Townes, Mars network: a Mars orbiting communications and navigation satellite constellation, in 2000 IEEE Aerospace Conference. Proceedings (Cat. No.00TH8484), vol. 7 (2000), pp. 75–88 15. M. Marcozzi, M. Ottavi, Evaluation of a multi-access communication architecture for future Mars exploration, in 38th International Communications Satellite Systems Conference (ICSSC 2021) (2021), pp. 27–30 16. K. Hill, K. Gagneja, Concept network design for a young Mars science station and Transplanetary communication, in 2018 Fourth International Conference on Mobile and Secure Services (MobiSecServ) (2018), pp. 1–8 17. C. Sacchi, S. Bonafini, From LTE-A to LTE-M: a futuristic convergence between terrestrial and martian mobile communications, in 2019 IEEE International Black Sea Conference on Communications and Networking (BlackSeaCom) (2019), pp. 1–5 18. V. Chukkala, P. De Leon, S. Horan, V. Velusamy, Modeling the radio frequency environment of Mars for future wireless, networked rovers and sensor Webs, in 2004 IEEE Aerospace Conference Proceedings, Big Sky, MT, vol. 2 (2004), pp. 1329–1336

10 Extraterrestrial Radio Access Network: The Road to Broadband. . .

239

19. V. Chukkala, P. De Leon, Simulation and analysis of the multipath environment of Mars, in 2005 IEEE Aerospace Conference, Big Sky, MT (2005), pp. 1678–1683 20. S. Bonafini, C. Sacchi, Building cellular connectivity on Mars: a feasibility study, in 2020 IEEE Aerospace Conference (2020), pp. 1–12 21. D. Wang, M. Giordani, M.-S. Alouini, M. Zorzi, The potential of multilayered hierarchical nonterrestrial networks for 6G: a comparative analysis among networking architectures. IEEE Veh. Technol. Mag. 16(3), 99–107 (2021) 22. R. Bassoli, F. Granelli, C. Sacchi, S. Bonafini, F.H. Fitzek, Cubesat-based 5g cloud radio access networks: a novel paradigm for on-demand anytime/anywhere connectivity. IEEE Veh. Technol. Mag. 15(2), 39–47 (2020) 23. S. Bonafini, C. Sacchi, R. Bassoli, F. Granelli, K. Kondepu, F.H.P. Fitzek, An analytical study on functional split in martian 3D networks. IEEE Trans. Aerosp. Electron. Syst. (2022). https:// doi.org/10.1109/TAES.2022.3187668 24. S. Bonafini, C. Sacchi, F. Granelli, R. Bassoli, F.H.P. Fitzek, K. Kondepu, 3D cloud-RAN functional split to provide 6G connectivity on Mars, in 2022 IEEE Aerospace Conference (AERO) (2022), pp. 1–13 25. S. Bonafini, C. Sacchi, R. Bassoli, K. Kondepu, F. Granelli, F.H.P. Fitzek, End-to-end performance assessment of a 3D network for 6G connectivity on Mars surface. Comput. Netw. 213, 109079 (2022). https://doi.org/10.1016/j.comnet.2022.109079

Part IV

New Space Applications

Chapter 11

Integration between Communication, Navigation and for Space Applications: Case Study on Lunar Satellite Navigation System with Focus on ODTS Techniques Cosimo Stallo, Henno Bookmap, Daniele Cretoni, Martina Cappa, Laura De Leo, Mattia Carosi, and Carmine Di Lauro

11.1 Introduction: From Earth to Space Applications Integration between communication, navigation and sensing technologies is the key to allow the provision of innovative and advanced services in different contexts of Earth-base applications (safety-critical (rail, road, maritime, aviation), massmarket (IoT (Internet of Things)), professional applications). In last years, in the context of land-based transportation the degree of automation of road vehicles has gradually increased [1]. Highly accurate positioning is the basis for intelligent vehicles to achieve path planning and motion trajectory tracking. GNSS (Global Navigation Satellite System) is widely adopted as primary method in vehicle localization. However, the positioning can be inaccurate or even denied in presence of tunnels and other occlusions or due to interferences. Hence, a GNSS-only positioning method cannot fully meet the needs of higher levels of SAE (Society of Automotive Engineers) not only in terms of accuracy, but also integrity [2]. Therefore, it is important to explore new vehicle localization methods to allow

C. Stallo () · D. Cretoni · M. Carosi · C. Di Lauro Thales Alenia Space, Roma, Italy e-mail: [email protected]; [email protected]; [email protected]; [email protected] H. Bookmap Telespazio, Darmstadt, Germany e-mail: [email protected] M. Cappa Ranstad Italia, Roma, Italy e-mail: [email protected] L. De Leo Ranstad, Roma, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 C. Sacchi et al. (eds.), A Roadmap to Future Space Connectivity, Signals and Communication Technology, https://doi.org/10.1007/978-3-031-30762-1_11

243

244

C. Stallo et al.

intelligent driving and fully autonomous transportation. Integration of different communication, navigation and sensing technologies on board of vehicle allows to autonomously sense the driving environment and cooperatively exchange data. Recently, the use of artificial intelligence on sensor data processing has emerged. Radar, LiDAR (Light Detection and Ranging), visual sensors, sonar systems, and INS (Inertial Navigation System) are mounted onboard smart and flexible platforms and on several types of unmanned vehicles in all types of environments. Particularly interesting is autonomous navigation for non-GNSS applications, such as underwater and indoor vehicle navigation [3, 4]. A last trend in the field of Position Navigation and Timing (PNT) is to use Signals Of OPportunity (SOOP) transmitted by existing satellite communication LEO (Low Earth Orbit) systems (like Iridium Next Communication System) [5] to provide PVT (Position, Velocity and Time) in areas where GNSS is severely degraded (like urban canyons) or denied due to intentional interferences (jamming, spoofing). Cellular signals are also attractive for PNT due to their inherently attractive characteristics (geometric diversity, larger bandwidth, signal availability, higher transmitted power, free to use, different frequency bands) [6]. However, these signals are not designed for PNT. Therefore, to use cellular signals for such purpose, several challenges must be addressed. This topic has been the subject of extensive research over the past few years. In [6] it is reported as cellular signals aided by INS are able to provide very promising performances. In order to achieve targeted accuracy levels required for specific applications (UAV (Unmanned Aerial Vehicle), automotive, maritime) the trend is to integrate SOOP with other sensors onboard of user terminal. However, the main disadvantages of SOOP stem from the fact that signals’ availability is not guaranteed everywhere and clock stability of LEO communication satellites is lower than that of GNSS satellites with higher error due to clock contributions in the PVT solution. In addition, if signals coming from different satellite communication systems are used, user navigation terminal needs multi-band antennas, multi-band RF front-end and a sufficient computing power. In order to overcome these issues, a dedicated LEO PNT constellation is needed to augment GNSS where their performance can be degraded in harsh environments (like urban canyons (for example to reduce TTFF (Time To First Fix) improving PPP (Precise Point Positioning)) [7]. Search and Rescue Service (SAR) offered by Galileo [8] is example of SoL (Safety of Life) service realized thanks to integration between communication and navigation that allows to save lives at sea. Galileo satellites are able to pick up emergency signals emitted from distress beacons at a frequency of 406 MHz and transmit a Return Link Message (RLM) signal back to the beacon through the Galileo Navigation Message (I/NAV E1). On January 21 2020, the SAR/Galileo Return Link Service (RLS) was declared operational. Galileo not only locates people in distress and makes their position known to the relevant authorities, but also SAR/Galileo RLS provides an automatic acknowledgement message back to the user informing them that their request for help has been received. This capability offered only by Galileo [9] with respect to other GNSS constellations to provide reassurance should deliver a valuable psychological lift to victims and further boost

11 Integration between Communication, Navigation and for Space. . .

245

survival rates. Recent studies [10] have been focused on the topic of integration of 5G (Fifth Generation) and HAPS (High Altitude Pseudo Satellites) based PNT in the new paradigm of Internet of Things (IoT), and then for a hybridisation with space-based PNT (being either from GNSS or LEO-PNT systems). The approach above described can be easily adapted to the space applications. In the context of interplanetary missions for lunar exploration, the integration of communication, navigation and sensing is the key to enable innovative and advanced services to allow return of humans to the Moon in the coming years and its colonization. The decade 2020–2030 will see the first phase of humanity’s return to the Moon, with different robotic missions under development around the world. While the first missions will focus on the South Pole, where most of Moon resources (ice water) are present, exploration ambitions will soon extend to the entire lunar surface. Many missions [11] currently show limitations due to by volume of spacecraft, their capacity to transfer data from the Moon, real time communication with the Earth or due to reduced PVT accuracy especially during some challenging operational phases as landing or surface navigation also impacting on astronauts safety and mission reliability and durability. In order to solve these issues, a dedicated constellation in lunar orbit [12] is therefore needed to permanently provide a range of communication and navigation services allowing to reduce cost and complexity of these missions.

11.2 Integrated Communication, Navigation and Sensing Systems for Space: Moon Case Study The last Global Exploration Roadmap has clearly identified the Moon as the short term priority in expanding the human presence in the Solar System [13, 14]. With more than a hundred lunar missions envisioned for this decade, a new Moon economy is emerging fast, posing technological challenges that are currently being tackled from both space agencies and an increasing number of privates. With Artemis, NASA will establish a long-term presence at the Moon, and it will require new, more robust communications, navigation, and networking capabilities. NASA’s Space Communications and Navigation (SCaN) program has developed the LunaNet architecture [15] to meet these needs and to establish a framework for systems interoperability and compatibility. LunaNet nodes will offer missions four distinct services: networking, navigation, detection and information, and radio/optical science services. LunaNet will start with a simple architecture of a few nodes to meet the needs of the early missions and evolve to meet the growing needs of a sustained lunar presence. All relay network services are not expected to be met by a single spacecraft, or node, but through a combination of interoperable systems supplied by commercial and government providers. The European Space Agency’s (ESA) Moonlight initiative [16, 17] aims to provide common Lunar Communication and Navigation Services (LCNS) to efficiently establish a reliable, sustainable and scalable network. A constellation of satellites in Elliptical Lunar Frozen Orbits

246

C. Stallo et al.

(ELFO) is being designed to satisfy those needs [18, 19]. The LCNS system high level architecture is decomposed in four main segments: • Earth Ground Segment (EGS): contains ground processing facilities dedicated to LCNS operations and User Requests management. • Lunar Space Segment (LSS): includes all the satellites orbiting the Moon for the provision of the communication, navigation and time services to the Lunar User Assets. • Moon Surface Segment (MSS): includes all LCNS elements that are located on the moon surface (i.e. beacon stations, wireless hotspots, etc.) to provide communication and navigation services to the surface LUS elements. • Lunar User Segment (LUS): includes the terminals that will be designed for the Lunar User Assets within the Orbit, Surface and Transportation scenario. Each terminal includes antennas, RF (Radio Frequency) equipment and electronics for establishing communication links with the LCNS Relay Satellites orbiting the Moon and for accessing the navigation and time services, if and when needed.

11.2.1 Lunar Communication System Architecture The Lunar Communications Architecture Working Group (LCAWG) has been tasked by the IOAG (Interagency Operations Advisory Group) to perform a study for defining a future Lunar Communications Architecture in order to support potential cross support to Lunar missions by communication assets owned and/or operated by the IOAG member agencies and their affiliated companies in the private sectors. Such an architecture aims at acting as a framework to ensure interoperability between different potential Lunar network(s) deployed by IOAG members. This architecture takes into account these elements: Lunar science orbiters, Lunar exploration orbiters, Lunar surface mobile and stationary vehicles, Lunar relay orbiters, Earth orbiting relays that provide service to lunar systems, Lunar Ascent and Descent modules, and associated Earth ground stations and mission operations centers. The communications links realized by the architecture includes EarthMoon link, Lunar proximity link, Lunar cross link, Lunar surface vicinity link, Earth orbiting relay link, and Earth space link extension. Scalability, expandability, interoperability, security, back-ward compatibility are main requirements to be considered in the design and deployment of this architecture. Figure 11.1 shows a concept overview of future communication network up to 2030 [20]. For what concerns the communication service, the lunar communication links will provide audio/voice and video streaming capabilities especially for the future Lunar manned missions. In the context of LCNS Programme, LCNS communications capabilities shall support at least user communications in S, X and Ka Bands and frequencies as defined and recommended in ITU (International Telecommunication Union) [21] and taking into account SFCG (Space Frequency Coordination Group) Recommendations [22] and ICSIS (International Communication System Interoperability

11 Integration between Communication, Navigation and for Space. . .

247

Fig. 11.1 Future Lunar communication network up to 2030 (courtesy by The Future Lunar Communications Architecture [20])

Standards) [23]. Different satellite communication architecture solutions can be identified based on the use of nodes for Communication from/to Earth and others for Communication from/to Moon. Figure 11.2 shows an example of a lunar satellite communication system architecture where different user needs are covered (LLO (Low Lunar Orbit), landing and surface user) and different frequency bands can be used (S, Ka bands). In this architecture, one of satellites can act as DTE (Direct To Earth) and as a potential hub (trunk link) for the others. Inter Satellite Link (ISL) between communication satellites can be foreseen.

11.2.2 Lunar Satellite Navigation System The requirements for a Lunar navigation system in terms of availability and accuracy are different from those for terrestrial GNSS, especially in the early stages of system deployment. Nonetheless, the baseline approach is to develop a positioning capability based on a constellation of lunar orbiting spacecrafts that transmit navigation messages, in combination with clock-based pseudorange observations collected by the user receiver. A baseline architecture consists of minimum 4 satellites deployed in ELFO orbits optimized to cover the South Pole, as shown in Fig. 11.3, that currently represents the most interesting Moon region

248

C. Stallo et al.

Fig. 11.2 Lunar satellite communication architecture high level concept Fig. 11.3 Example of Lunar navigation constellation satellite orbits

due to recent discoveries of ice water in that area [24]. Table 11.1 shows the main parameters of an example of potential candidate Elliptical Lunar Frozen Orbits (ELFO) for Lunar Satellite Navigation constellation. The number of satellites proposed is the minimum one for a PVT estimation in a GNSS-like localisation problem. A constellation with a low number of satellites is proposed to cover a

11 Integration between Communication, Navigation and for Space. . . Table 11.1 Example of ELFO orbital parameters of potential candidate orbits for lunar navigation

Parameters SMA(km) i(.◦ ) e Ar.Per.(.◦ ) RAAN(.◦ ) TA(.◦ )

1 9169.5 66.3 0.76 88.9 89 350

249 2 9172.4 57.9 0.76 97 335 350

3 9172.4 57.9 0.76 97 335 328

4 9149.6 56.1 0.76 84.1 208.37 352

specific area of the Lunar surface (i.e. South Pole). However, the availability of the service depends of the number of satellites in visibility and the orbits selected. A reduced ELFO constellation guarantees navigation performances suitable for an initial service. Preliminary studies and simulations about stable constellations of Lunar Frozen Orbits (LFO) have been conducted in the past years [25]. The need of a continuous coverage of the South Pole leads to a selection of orbits characterized by Ely and Lieb [26]: • large eccentricity (focused coverage near the apoapsis); • proper inclination and argument of periapsis set to 90.◦ (orientation of the apoapsis above the South Pole); • large semi-major axis (minimization of satellites on the same orbit). In this framework, LFOs are of special interest since the average value of eccentricity, argument of periapsis and semi-major axis of such orbits remain stationary for long periods (years) [27], this minimising fuel consumption required for orbit maintenance. However, LFOs are very sensitive to their initial conditions, where slightly different orbital elements from a restricted set of specific values do not allow maintaining frozen orbits conditions: conventionally, the dynamical model considered to study analytical relationships for initial orbital parameters retrieval includes a circular-orbit point-mass Earth and point-mass Moon, but results are shown to work quite well, even in presence of a more complex model, providing justification to use these orbits in real scenarios. In the baseline Lunar Satellite Navigation architecture here proposed, the satellites have one-way navigation payloads transmitting a ranging signal in the frequency bands [2483.5–2500] MHz [17]. Considering free space losses as the only source of propagation loss, the EIRP (Effective Isotropic Radiated Power) needed for an LFO satellite at boresight and at the maximum slant range from the user to close the link budget is slightly below 15 dBW (assuming at user receiver level a Noise Figure value of 1.5–2 dB, C/N0 (Carrier to Noise Ratio) Acquisition Threshold 30 dBHz and user antenna gain of .−2dB). The core of navigation system is the ODTS that is responsible for the generation of broadcast navigation messages to inform the user receiver about the predicted orbit and clock of the LNSS (Lunar Navigation Satellite System) satellites. Both during orbit propagation and for the generation of a lunar navigation message, Lunar

250

C. Stallo et al.

Navigation System [17] will use Moon Centered Reference Systems, both inertial and body-fixed. Trade-off among different techniques is reported in Sect. 11.3. A candidate Baseline ODTS concept is then presented, and it is a result of a dedicated trade-off in terms of solution complexity, cost, scalability, achievable performances and near-term implementation.

11.3 Orbit Determination Techniques for Lunar Satellite Navigation This section shows the main orbit determination methods and technologies applicable for a lunar satellite navigation system, and illustrates the main advantages and disadvantages for each of them. As far as Orbit Determination and Timing Synchronisation (ODTS) is concerned, there are just two kinds of tracking geometry: • Tracking from the Earth direction (S-band, terrestrial GNSS, SLR (Satellite Laser Ranging), VLBI (Very Long Baseline Interferometry)); • Tracking from within the LNSS constellation (ISL (Inter Satellite Link), LLO (Low Lunar Orbit) LNSS receivers, beacons on lunar surface) Only range and range-rate observations are available: • Range: GNSS code pseudorange, S-band range, ISL range, SLR. • Range-rate: GNSS Doppler, S-band Doppler, ISL Doppler. Location of the ODTS analysis process has impact on communication bandwidth, available update rates and SWaP (Size, Weight and Power Consumption): • Observation data must be transferred to the ODTS process (unless OD takes place at the tracking source); • Output products must be transferred to the LRNS payload (unless OD takes place on-board the LRNS satellites).

11.4 Tracking from Earth or from Earth Orbit The distance between the Moon and Earth is so large that any form of terrestrial tracking (regardless of the technique) deteriorates into range (or range-rate) observations along the line of sight, which is always approximately parallel to the Earth-Moon vector. Range observations are one-dimensional, but the orbit must be determined in three (inertial) dimensions. In order to achieve 3D observability of the orbit, the 1D observations must come from different directions in 3D space. It is important to understand what the main source of 3D information is in case of terrestrial range observations. The Moon, as seen from Earth, spans 0.5.◦ across the sky. A highly eccentric Lunar satellite orbit with semi-major axis of 6000 km (as

11 Integration between Communication, Navigation and for Space. . .

251

Fig. 11.4 Moon-Earth geometry in the case of tracking from Earth

example) would span around 1.8.◦ . The Earth as seen from the Moon spans only 1.9.◦ , and the ring of GNSS observations that can reach a lunar spacecraft spans just over 2.5.◦ . All these angles are so small that the observations from different terrestrial stations or GNSS satellites do not offer any significant 3D observability at a single observation epoch. However, the Moon itself moves around the Earth in about 28 days, which corresponds to an angular motion of 13.◦ per day. If we collect terrestrial range data over a 7 day arc, the data at the beginning of the arc makes an angle of 90.◦ with the data at the end of the arc, as shown in Fig. 11.4. It is this effect that is the main source of 3D observability of a lunar orbit. The position of a tracking station on Earth barely contributes to the 3D observability. Every Lunar satellite will need some housekeeping communication link to support its operations (for instance, to update the navigation message). At least in the early stages of the lunar satellite system, this communication could take place with terrestrial stations since there is no lunar relay station or similar available. All use of such communication links for range or range-rate tracking will here be referred to as “S-band” tracking, even if another frequency band may be used. The actual frequency of the carrier is almost irrelevant in comparison to the more operational properties of the data, such as temporal distribution or available contact windows between the spacecraft and a suitable ground station. S-band ranging was in fact a spin-off technology from the Apollo program, and the key properties of this tracking technique were defined in 1965 and have not changed by much since that time. It is clear that the technique itself can be used for lunar missions in general, however,

252

C. Stallo et al.

the accuracy requirements of the LNSS could be rather strict and not many lunar missions—if any—have used S-band as a precise tracking technique. Operational performance data for lunar S-band tracking are available from various missions, such as SMART-1 and LRO (Lunar Reconnaissance Orbiter). Typical noise levels of the S-band tracking residuals were reported for these missions as it follows: • SMART 140 m RMS (Root Mean Square) [28]; • LRO 55 m RMS (Root Mean Square) [29]. Reported issues with terrestrial S-band for lunar use are mainly: • seasonal blinding by the Sun; • occasional periods with very poor coverage due to unfortunate geometry. • unpredictable station availability windows. S-band is useful/easy/available, but could be unreliable since the distribution of the available data over time and space does not seem to be sufficiently dense to use S-band as the only tracking data type for the generation of the ODTS navigation message. Therefore, communication links that may be exploited in addition to housekeeping to terrestrial station network (typically S-band) could be: • LNSS payload communication links (including ISL communications); • Future communication with lunar surface stations.

11.5 Satellite Laser Ranging Satellite Laser Ranging (SLR) is one of the oldest satellite tracking techniques, and is also one of the most accurate techniques in terms of residual data noise. Lunar Laser Ranging (LLR) has been used since the days of the Moon landings, when reflector arrays were left behind by the Apollo astronauts and have been used ever since. The key significance of this is that especially thanks to this SLR data, we now have a highly accurate reference frame ephemerides and rotation model. Tracking a lunar orbiting spacecraft with SLR data is not straightforward. The first successful two-way range SLR passes to a lunar orbiter were performed in 2018, between the Grasse ground station in France and the LRO [30] (see Fig. 11.5). In [30] the authors provided the results of the first series of successful two-way laser ranging experiments from a ground station, the LLR station in Grasse, France, to a spacecraft at lunar distance, the LRO. Grasse station measured 67 returns in two 6-min sessions on September 4, 2018. SLR observation residuals were less than 3 cm RMS (Root Mean Square) range or 180 ps RMS over the 2-way flight time. In particular, SLR to lunar spacecraft faces two problems that terrestrial satellites do not have (or to less extent): • strongly varying geometry between satellite, Moon and rotating Earth requires the satellite to “assist” the SLR tracking by rotating its reflector array approximately towards the station;

11 Integration between Communication, Navigation and for Space. . .

253

Fig. 11.5 LRO from Lunar Laser Ranging Station in Grasse (France) (courtesy by Mazarico et al. [30])

• 2-way laser travel time to a lunar target is about 2.5 s, and during this time the line-of-sight should be free of atmospheric perturbations (troposphere jitter, turbulence etc.). Only a few telescopes have sufficiently stable weather conditions to have a predictable return on an SLR campaign to the Moon. Therefore, to consider SLR as a stand-alone baseline tracking technique for LNSS could be not sufficient. However, the great strength of SLR data is its high accuracy (cm level, even to lunar targets). This makes SLR the best candidate technique to perform independent validations of any ODTS system based on other tracking methods. It could be recommended to equip at least one of the LNSS spacecraft with an SLR retroreflector array for this purpose. In summary: • LLR can play similar role for the spacecraft as SLR for Galileo: validation, calibration of absolute antenna bias; • LLR can be very significant for lunar surface geodesy and reference frame. • LLR is too sparse to be the stand-alone baseline for ODTS message generation.

11.6 Very Long Baseline Interferometry Very long baseline interferometry (VLBI) is a observational technique based on combining geographically distributed radio telescopes to form an observation network, thereby enabling the highest possible angular resolution. The imaging resolution provided by VLBI is inversely proportional to the maximum baseline (the separation between two telescopes in the network) length and proportional to the observation wavelength. A Moon-based radio astronomical observatory is also under consideration [31]. Although there was subsequently intensive development of spacecraft navigation using VLBI, including tracking of both Cassini and Huygens at Saturn and Titan [32], VLBI was not applied again in lunar operations until 2007, when the Japanese Selene spacecraft carried two small subsatellites into lunar orbit. These small spacecraft, Rstar and Vstar, were used in the VRAD

254

C. Stallo et al.

(differential VLBI RADio sources) mission in order to improve the knowledge of the lunar gravity field, particularly on the far side. In the Chang’E-4 mission [33] (January 2019), the lander was equipped with a low-frequency radio telescope and it landed on the far side of the Moon. Thus, radio environment data were measured on the far side of the Moon for the first time. COMPASS (Combined Observational Methods for Positional Awareness in the Solar System) [34] will use beacons that emit coherent ultra-wideband signals designed to be interoperable with existing and future VLBI networks. Using differential VLBI, COMPASS will provide rapid determination of the interferometric phase delay with picosecond level accuracy during routine VLBI observing sessions.

11.7 High Sensitivity Spaceborn Receiver The use of terrestrial GNSS signals above the MEO (Medium Earth Orbit) orbits of the GNSS satellites themselves, and even up to lunar distance has attracted increased interest over recent years [35, 36]. Use of terrestrial GNSS weak signals for missions to the Moon has already been demonstrated as: • NASA Magnetospheric Multiscale (MMS) Mission (2019) uses GNSS signals half-way between Earth and Moon (highest ever use at 29 RE (Radius Earth) [35]; • PS-based onboard autonomous navigation sigma (3D)