Energy-efficient spectrum management for cognitive radio sensor networks 978-3-319-60318-6, 3319603183, 978-3-319-60317-9

Table of contents :
Front Matter ....Pages i-xv
Introduction (Ju Ren, Ning Zhang, Xuemin (Sherman) Shen)....Pages 1-13
Spectrum Resource Management for CRSNs (Ju Ren, Ning Zhang, Xuemin (Sherman) Shen)....Pages 15-22
Dynamic and Energy-Efficient Channel Access in Clustered CRSNs (Ju Ren, Ning Zhang, Xuemin (Sherman) Shen)....Pages 23-50
Secure and Energy-Efficient Collaborative Spectrum Sensing (Ju Ren, Ning Zhang, Xuemin (Sherman) Shen)....Pages 51-79
Joint Channel Allocation and Sampling Rate Control in EH-CRSNs (Ju Ren, Ning Zhang, Xuemin (Sherman) Shen)....Pages 81-106
Concluding Remarks and Future Directions (Ju Ren, Ning Zhang, Xuemin (Sherman) Shen)....Pages 107-110

Citation preview

Ju Ren · Ning Zhang Xuemin (Sherman) Shen

Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks

Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks

Ju Ren • Ning Zhang • Xuemin (Sherman) Shen

Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks

123

Ju Ren School of Information Science and Engineering Central South University Changsha, Hunan, China

Ning Zhang Department of Electrical and Computer Engineering University of Waterloo Waterloo, ON, Canada

Xuemin (Sherman) Shen Department of Electrical and Computer Engineering University of Waterloo Waterloo, ON, Canada

ISBN 978-3-319-60317-9 ISBN 978-3-319-60318-6 (eBook) DOI 10.1007/978-3-319-60318-6 Library of Congress Control Number: 2017943815 © Springer International Publishing AG 2018 This work is subject to copyright. All rights are reserved 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, express 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. Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Dedicated to my wife and parents. —Ju Ren

Preface

As a promising and fundamental networking solution for data collection in the Internet-of-Things (IoT) era, Wireless Sensor Network (WSN) has been widely applied in various fields and scenarios. However, due to the explosion of mobile devices and wireless services during recent years, the unlicensed spectrum, especially the Industrial, Scientific, and Medical (ISM) bands, is becoming increasingly crowded. Consequently, WSNs, which are generally operating over the ISM spectrum bands, have to face significant and uncontrollable interference from the overlapping wireless systems and hence to experience great network performance degradation. By introducing Cognitive Radio (CR) technology into WSN, it brings a new and advanced networking solution, named Cognitive Radio Sensor Network (CRSN), to address the spectrum scarcity problem in the IoT era. In CRSNs, sensor nodes can sense the availability of the licensed spectrum and opportunistically access the vacant spectrum bands for data transmission to avoid the inter-network interference. However, CRSN brings not only new opportunities but also new challenges of spectrum resource management under the energy-constrained and self-organized multi-hop sensor networks. The “double-edged sword” motivates us to design energy-efficient spectrum resource management schemes for CRSNs with full consideration of the inherent network characteristics of WSNs. In this monograph, we first introduce the architecture, benefits, and some potential applications of CRSN, as well as the arisen challenges in the spectrum resource management for CRSN. Then, we provide a comprehensive review and indepth discussion of the state-of-the-art research literature and propose an integrated energy-efficient spectrum resource management solution for CRSNs, covering spectrum access decision, spectrum sensing, and spectrum resource allocation. Since the energy consumption of spectrum sensing and channel switching is critical for sensor nodes, we first present a dynamic channel sensing and accessing scheme for CRSNs to determine when sensor nodes should sense and access a licensed channel for data transmission to improve energy efficiency. Moreover, to improve the accuracy and security of spectrum sensing, a secure and energy-efficient collaborative spectrum scheme is proposed in detail to resist Spectrum Sensing Data Falsification (SSDF) attacks and enhance the energy efficiency of spectrum sensing vii

viii

Preface

in CRSNs. In addition, we investigate the spectrum resource allocation problem in energy-harvesting CRSNs and present a joint channel access and sampling rate control scheme to maximize the network utility characterized by the sensed data amount of sensor nodes. Finally, some potential future research directions are envisioned to attract continuous research efforts in this emerging and evolving field of study. We hope that this monograph can provide some valuable insights for practical CRSN design and motivate new ideas for future IoT networking. Changsha, China Waterloo, ON, Canada Waterloo, ON, Canada March 2017

Ju Ren Ning Zhang Xuemin (Sherman) Shen

Acknowledgements

We would like to express our sincere thanks to the friends and colleagues at the BroadBand Communication Research (BBCR) group, University of Waterloo, Canada, for their invaluable comments and suggestions. Special thanks are also due to the staff at Springer Science + Business Media: Susan Lagerstrom-Fife and Patrick Carr, for their help throughout the publication preparation process. The author, Ju Ren, would also like to thank his Ph.D. supervisor, Prof. Yaoxue Zhang, for his guidance, and the group members at the Transparent Computing Laboratory, Central South University, China, for their help and contributions to this monograph.

ix

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.1 Wireless Sensor Networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.2 Cognitive Radio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.3 Cognitive Radio Sensor Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Applications and Challenges of CRSN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 CRSN Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Arisen Challenges for CRSNs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Aim and Outline of This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 1 1 3 4 7 7 8 10 12

2

Spectrum Resource Management for CRSNs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Dynamic Spectrum Access Decision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Secure Collaborative Spectrum Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Joint Energy Management and Spectrum Resource Allocation. . . . . . . 2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15 15 16 18 19 20

3

Dynamic and Energy-Efficient Channel Access in Clustered CRSNs . . 3.1 System Model and Objective Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1 Network Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2 Cognitive Radio Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3 Energy Consumption Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4 Objective Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Dynamic Channel Access for Intra-Cluster Data Transmission . . . . . . 3.2.1 Energy Consumption Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Optimal Transmission Time Allocation . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Analysis of Channel Sensing and Switching Decision. . . . . . . . 3.2.4 Dynamic Channel Sensing and Accessing Scheme . . . . . . . . . . . 3.3 Joint Power Allocation and Channel Access for Inter-Cluster Data Transmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Analysis of Channel Sensing and Switching Decision. . . . . . . .

23 24 24 25 27 28 30 30 31 33 35 35 36 xi

xii

4

5

Contents

3.3.2 Joint Transmission Power and Time Allocation . . . . . . . . . . . . . . . 3.3.3 Joint Power Allocation and Channel Accessing Scheme . . . . . 3.4 Performance Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1 Simulation Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2 Performance of Intra-Cluster Data Transmission . . . . . . . . . . . . . 3.4.3 Performance of Inter-Cluster Data Transmission . . . . . . . . . . . . . 3.4.4 Impacts of System Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39 42 43 43 44 45 47 48 49

Secure and Energy-Efficient Collaborative Spectrum Sensing . . . . . . . . . 4.1 System Model and Design Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1 Network Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2 Channel Sensing Model and Decision Rule . . . . . . . . . . . . . . . . . . . 4.1.3 Attack Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.4 Energy Consumption Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.5 Design Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Analysis on Trade-Off Between Security and Energy-Efficiency . . . . 4.2.1 Attack Impact Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 Security and Energy-Efficiency Trade-Off . . . . . . . . . . . . . . . . . . . . 4.3 FastDtec: Trust Evaluation for Fast Compromised Node Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1 Trust Evaluation Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.2 Analysis on Optimal Detection Threshold. . . . . . . . . . . . . . . . . . . . . 4.3.3 The Proposed FastDtec Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Secure and Energy-Efficient Collaborative Spectrum Sensing . . . . . . . 4.4.1 The Proposed Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2 Further Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Performance Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.1 Simulation Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.2 Attack Impacts Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.3 Performance of the Proposed Scheme . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51 53 53 54 55 57 58 58 59 61

Joint Channel Allocation and Sampling Rate Control in EH-CRSNs . 5.1 System Model and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.1 Network Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.2 Communication Link Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.3 Energy Harvesting and Consumption Model . . . . . . . . . . . . . . . . . . 5.1.4 Channel Interference Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.5 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Problem Decomposition and Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1 Problem Decomposition and Dual Problem . . . . . . . . . . . . . . . . . . . 5.2.2 Subgradient Method for Solving Dual Problem . . . . . . . . . . . . . . .

81 82 82 85 85 86 87 88 88 91

63 63 65 68 69 69 69 71 71 73 77 77 78

Contents

6

xiii

5.3 Subproblem Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1 Optimal Solution for Sampling Rate Control . . . . . . . . . . . . . . . . . 5.3.2 Computation Complexity Analysis on Channel Allocation . . 5.3.3 Suboptimal Solution for Channel Allocation . . . . . . . . . . . . . . . . . . 5.4 JASC: Joint Channel Allocation and Sampling Rate Control Scheme 5.4.1 Algorithm for Solving Primal Problem . . . . . . . . . . . . . . . . . . . . . . . . 5.4.2 The Proposed JASC Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 Performance Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.1 Utility Comparison and Efficiency Evaluation . . . . . . . . . . . . . . . . 5.5.2 Channel Allocation and Interference Evaluation . . . . . . . . . . . . . . 5.5.3 Impacts of System Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

91 91 92 95 97 97 98 99 100 102 103 104 105

Concluding Remarks and Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Future Research and Development Directions. . . . . . . . . . . . . . . . . . . . . . . . . 6.2.1 Distributed Spectrum Resource Management for Large-Scale CRSNs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.2 Cross-Layer Design for Opportunistic Spectrum Access . . . . . 6.2.3 RF Energy Harvesting and Transfer in CRSNs . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

107 107 108 108 109 109 110

Acronyms

CAD CEHA CH CM CR CRN CRSN C-SSDF DP-NUMP EH EH-CRSN FAP IoT ISM I-SSDF JASC MAC MDP MINLP NUMP PTAP PU QoS RF RSSI SNR SSDF SU TAP TDMA WSN

Channel Available Duration Cross-Entropy based Heuristic Algorithm Cluster Head Cluster Member Cognitive Radio Cognitive Radio Network Cognitive Radio Sensor Network Collaborative Spectrum Sensing Data Falsification Dual Problem of Network Utility Maximization Problem Energy Harvesting Energy Harvesting Cognitive Radio Sensor Network False Alarm Probability Internet-of-Things Industrial, Scientific, and Medical Independent Spectrum Sensing Data Falsification Joint Channel Access and Sampling Rate Control Media Access Control Missed Detection Probability Mixed-Integer Nonlinear Optimization Problem Network Utility Maximization Problem Joint Transmission Power and Time Allocation Problem Primary User Quality of Service Radio Frequency Received Signal Strength Indicator Signal-to-Noise Ratio Spectrum Sensing Data Falsification Secondary User Transmission Time Allocation Problem Time Division Multiple Access Wireless Sensor Network xv

Chapter 1

Introduction

1.1 Background 1.1.1 Wireless Sensor Networks With the fast development of sensor, wireless communication and micro-electronics technologies, WSN that integrates sensing, computing and communicating technologies has emerged as a promising networking solution to revolutionize the field of information sensing and collection [1, 2]. Benefited by the features of selforganization, distributed operation, low cost, and low-power, etc., WSN also attracts significant research attention from the academic during the past decade [3, 4]. A typical WSN consists of multiple sensor nodes deployed in an interested area for information sensing, converting, processing and transmission. These distributed sensor nodes can sense and collect the environmental data within certain range (such as pictures, temperature, humidity, etc.) before sending it to the sink via multihop data transmission. After simple data aggregation and processing, the sink will subsequently transmit the aggregated sensed information to the remote data center via an access point for specific applications. With the low-cost and self-organized WSNs, humans are enabled to transfer various information data from the physical world to the cyber world in an efficient manner and thereby promote the great fusion of cyber physical systems. For example, animalists have no need to personally track animals but know their living habits and health status through implanting sensor nodes into animals or deploying monitoring node in the grassland [5]. Another instance is that researchers can acquire real-time monitoring data without visiting mines and volcanoes by deploying sensor networks in these dangerous areas [6]. The promising advantages of WSNs have greatly driven its applications to penetrate into today’s informatized and technological society. The earliest research on WSN can be dated back to the tactics of US army during the Vietnam War. The US army arranged sensor nodes on “tropical trees” around Ho Chi Minh trail, to monitor whether there were Vietnamese troops passing and © Springer International Publishing AG 2018 J. Ren et al., Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks, DOI 10.1007/978-3-319-60318-6_1

1

2

1 Introduction

send the monitored data to the command center, which would decide whether to wage attack against the passing troops. This successful application considerably promoted the development of early-phase WSN in the military area, and brought birth to the navy cooperative engagement system, remote battlefield sensor system and so on [7]. It also played a significant role in enhancing battlefield monitoring and the combating reaction capability. However, as the study on WSNs develops, researchers discover the huge commercial values of WSNs and gradually employ them into civil domains. Business Week listed WSN among the 21 most influential technologies in the twenty-first century in 1999, and the US journal Technology Review also rated WSN as the top 1 of ten emerging technologies to impose profound influences on future human life in 2003. Since then, countries across the world have increased their inputs in the research and industrial application of WSN, making it the hottest research area for academic and industry. In view of the characteristics of sensor nodes and user demands for WSN applications, the majority of commercial sensor nodes remain to develop for smaller size, lower cost and power consumption and so on. Short-range and low-power communication protocols like ZigBee operating in the unlicensed spectrum turn to be suitable communication technologies for sensor nodes. These protocols can effectively reduce the cost of network deployment and increase the energy efficiency of sensor nodes, thus they are wildly adopted by most WSN applications. However, due to the explosion of wireless services and applications, the spectrum for wireless communication becomes a type of scarce resources. To avoid huge expense of using licensed spectrum, most newly-arisen wireless services, including WiFi, Ad Hoc networks, Bluetooth, etc., all operate on the unlicensed spectrum, especially the Industrial, Scientific, and Medical, (ISM) spectrum bands. But, the problem is that the rapid growth of wireless services makes the unlicensed spectrum increasingly crowded, resulting in unavoidable interferences to WSN applications working on the same spectrum bands. Such inter-network interferences caused by spectrum scarcity are significantly uncontrollable to WSNs, greatly degrading the network performances of WSNs. Table 1.1 shows some working frequencies of typical sensor nodes as well as somen overlapping wireless services.

Table 1.1 Work frequencies of typical sensor nodes Sensor nodes Mica [8] Mica 2 [9]

Working frequencies 916.3–916.7 MHz 315, 433, 868, 916 MHz

Mica Z [10], IMote [11], Iris [12] EyesIFX [13]

2.4 GHz 868–870 MHz

Overlapping wireless services Telemetry, ZigBee, Cellular Cellular, Satelite, Telemetry, ZigBee Cellular, IEEE 802.11b/g/n, Bluetooth, ZigBee Telemetry, ZigBee, Cellular

1.1 Background

3

1.1.2 Cognitive Radio Cognitive Radio (CR) is a form of wireless communication in which a radio can intelligently detect the radio environment and be programmed and configured dynamically to use the best vacant wireless channels in its vicinity [14]. It has been expected as a promising solution to address the spectrum scarcity problem. The concept of CR was first proposed by Joseph Mitola III in a seminar at KTH (the Royal Institute of Technology in Stockholm) in 1998 and published in an article by Mitola and Gerald Q. Maguire, Jr. in 1999 [15]. In Mitola’s work, CR was defined as an intelligent wireless communication system, which can sense the ambient radio environment and use artificial intelligence to analyze the sensing results. Based on the analysis results, CR can dynamically configure the radio-system parameters, including “waveform, protocol, operating frequency and networking”, to make devices access different spectrum bands for communication at different times. CR enables communicating devices to timely detect and opportunistically access the “spectrum holes” for data transmission, significantly enhancing the spatiotemporal utilization of spectrum resources. During the past a few years, CR has been evolving in terms of both concept and implementation technologies. In December 2003, a formal definition of CR is announced by the Federal Communications Commission (FCC). “A radio or system that senses its operational electromagnetic environment and can dynamically and autonomously adjust its radio operating parameters to modify system operation, such as maximize throughput, mitigate interference, facilitate interoperability, access secondary markets.” Based on this definition, FCC also aims to use the CR technology to create an open spectrum market, where Primary Users (PUs) have the priority to access the licensed spectrum, but Secondary Users (SUs) empowered by CR can opportunistically access the licensed spectrum for data transmission under the guarantee of no interference to PUs. Although there have been a number of different CR definitions [16], some common characteristics and basic functionalities of CR can be summarized as follows. 1. Spectrum Sensing. CR-enabled devices can sense and analyze the signal strength over a specific spectrum band, and determine whether “spectrum holes” exist. Moreover, the spectrum sensing capability also requires CR-enabled devices to timely sense if PU has returned back the licensed channel, after they access the vacant licensed channel. 2. Spectrum Decision. According to the spectrum sensing results, CR-enabled devices can further determine which vacant channel should be accessed based on specific application requirements or optimization targets. 3. Spectrum Sharing. Spectrum sharing allows SUs to fairly share the licensed channels of PUs. However, SUs have to restrict their transmit power so that the interference caused to PUs can be kept below a certain threshold. 4. Spectrum Mobility. Spectrum mobility should be supported by CR-enabled devices by dynamically changing their frequency of operation. Such that, cognitive-radio networks (CRNs) can use the spectrum in a dynamic manner

4

1 Introduction

by allowing radio terminals to operate in the best available frequency band, and maintaining seamless communication requirements during transitions to better spectrum. With the four basic functionalities, CR can significantly improve the efficiency of spectrum utilization. Promoted by the advocacy of FCC and continuous research efforts from the academic, CR has experienced a big step forward from theory to practice. Several standards have also been proposed for guiding the design of CR devices and applications. In October 2004, the Institute of Electrical and Electronic Engineers (IEEE) initializes a special group, named IEEE 802.22 work group, to develop the international standards for CR. IEEE 802.22, is a standard for wireless regional area network (WRAN) using white spaces in the television (TV) frequency spectrum. The development of the IEEE 802.22 WRAN standard aims at using cognitive radio (CR) techniques to allow sharing of geographically unused spectrum originally assigned to TV broadcast services, on a non-interfering basis, to bring broadband access to hard-to-reach, low population density areas, typical of rural environments. The standards of IEEE 802.22 are developed according to the general requirements and rules of IEEE 802 groups, and is compatible with the IEEE 802 serial protocols. The standards were published in July 2011 and attracted a large number of CR test-beds and products in the past several years.

1.1.3 Cognitive Radio Sensor Networks By applying CR capabilities into WSN, Cognitive Radio Sensor Network (CRSN) emerges to be a promising solution for addressing the spectrum-scarcity problem in traditional WSNs. In general, a CRSN can be defined as a distributed network of wireless cognitive radio sensor nodes, which can sense event signals and leverage CR technology to communicate their readings dynamically over available spectrum bands in a multihop manner to ultimately satisfy the application-specific requirements. In CRSN, sensor nodes act as SUs to sense the availability of licensed spectrum and opportunistically access the vacant ones for data transmission. In such a way, CRSN can significantly improve the network performance and the spectrum utilization efficiency. In this section, we briefly introduce a typical architecture of CRSN as well as its potential benefits.

1.1.3.1

A Typical Architecture of CRSN

In 2009, the concept of CRSN was first proposed by Prof. Ozgur B. Akan and published on IEEE Network [17], with a detailed introduction to its architecture and associated challenges. Figure 1.1 shows the proposed architecture of CRSN, where a large number of distributed sensor nodes equipped with cognitive radio modules can opportunistically access vacant licensed channels to transmit their sensed data

1.1 Background

5

CR Sensor

CR Cluster Head

Spectrum Management Center

PU Communication Link

PU

Base Station

Opportunistic Communication Link

Mobile CR Sensor

Sink

Fig. 1.1 The architecture of CRSN

to the sink node via a multihop manner. The sink may also be equipped with CR capabilities to receive data via different channels. Besides, to efficiently manage the dynamic channel access in CRSN, there should be a default control channel for communicating the control information and coordinating the whole process. Figure 1.2 shows the hardware architecture of a cognitive radio sensor node. As shown in the figure, the main difference between the hardware structure of a classical sensor node and a CR sensor node is the cognitive radio transceiver. The cognitive radio unit enables the sensor nodes to dynamically adapt their communication parameters such as carrier frequency, transmission power, and modulation. Cognitive radio sensor nodes also inherit the limitations of conventional sensor nodes in terms of power, communication, processing, and memory resources, which consequently restricts the features of cognitive radio.

1.1.3.2

Benefits of CRSN

Compared to traditional WSNs with fixed spectrum allocation, CRSN can benefit from the following potential advantages. • Dynamic Spectrum Access. The existing WSN deployments assume fixed spectrum allocation over very crowded unlicensed bands that are also used by other devices. Nevertheless, a spectrum lease for a licensed band amplifies the overall deployment cost. Hence, to be able to cooperate efficiently with other types of users, opportunistic spectrum access may be utilized in WSNs.

6

1 Introduction

Processing Unit

CR Transceiver Spectrum sensing results, spectrum ploicy

RF front-end

Processor

PU statistics, spectrum management rules

Operating frequency, modulation, transmit power, coding scheme

Demodulator (QAM,QPSK, etc.)

Decoder (FEC,block codes,LDPC ,etc)

Memory

Battery Power Unit

Recharging unit

Fig. 1.2 The hardware architecture of a cognitive radio sensor node

• Opportunistic Channel Usage for Bursty Traffic. A large number of sensor nodes detecting an event generate bursty traffic and try to acquire the channel to send their readings. This increases the probability of collisions and packet losses, which decreases the overall communication reliability with excessive power consumption. Opportunistic access to multiple alternative channels may alleviate these potential challenges. • Adaptability for Reducing Power Consumption. The dynamic nature of the wireless channel causes energy consumption due to packet losses and retransmissions. Cognitive radio capable sensor nodes may be able to adapt to varying channel conditions, which would increase transmission efficiency, and hence help reduce power used for transmission and reception. • Overlaid Deployment of Multiple Concurrent WSNs. Dynamic spectrum management may significantly contribute to the efficient coexistence of spatially overlapping sensor networks in terms of communication performance and resource utilization. • Communication Under Different Spectrum Regulations. A certain band available in one specific region or country may not be available in another due to varying spectrum regulations. Sensor nodes equipped with cognitive radio capability may overcome this potential problem.

1.2 Applications and Challenges of CRSN

7

1.2 Applications and Challenges of CRSN 1.2.1 CRSN Applications Enabled by the CR capability, CRSNs can dynamically access available licensed spectrum to perform monitoring and surveillance tasks when facing interferences from the overlapping wireless applications, such as Wi-Fi and Bluetooth applications. Therefore, CRSNs can significantly improve the applicability of WSNs in a variety of real-life scenarios.

1.2.1.1

E-Healthcare Applications

Electronic health-care (E-healthcare) system, also called wireless body area network (WBAN), consists of several biomedical sensors implanted in/on the patient’s body to connect the human body and information world [18]. These sensors can sense diverse health information from human, such as pressure and blood oxygen, motions and location. The sensed information is compressed by the embedded low-power computation modules and sent to the central controller for prevention or early detection of diseases. However, due to the inter-network interference and congestion caused by the spectrum-overlapped wireless applications in the context of e-healthcare system, the ISM spectrum becomes no longer sufficient for life-critical health monitoring. Moreover, the operation of body sensors is also facing the electromagnetic interference (EMI) to the biomedical devices, such as electrocardiograph monitor and electromyography, making the stability and dependability of data transmission be a critical challenge in e-healthcare system. By enabling body sensors to opportunistically access a larger range of spectrum, CRSN is a promising alternative of WBAN to mitigate the inter-network interferences and avoid the EMI to ambient biomedical devices.

1.2.1.2

Smart City Applications

The emerging smart city requires advanced sensors that are integrated in the real-time monitoring systems, including transportation system and power grid, to intelligently manage city’s assets and tackle inefficiency [19]. However, the wireless channel condition on the ISM spectrum bands is quite harsh and perverse due to the complex propagation environment brought by densely situated buildings. On the other hand, large portions of the licensed spectrum are underutilized even in the populated urban area, e.g., the TV white space. CRSN can enable sensors to access the temporally available spectrum, and hence decrease data transmission delay in real-time monitoring.

8

1.2.1.3

1 Introduction

Indoor Surveillance and Localization Applications

In order to provide intelligent and personalized services, numerous indoor scenarios, such as houses, shopping malls and factories, deploy a large number of sensors for surveillance and localization. Meanwhile, a significant portion of indoor scenarios have been covered by Wi-Fi hotspots to satisfy the rapid increasing demand for convenient Internet access. Since the transmission power of Wi-Fi hotspots is much higher than the power of sensor networks, the coexistence of WiFi and indoor WSNs leads to severe interferences for the latter, significantly degrading the network performance of indoor WSNs. By applying CRSNs into indoor surveillance and localization, it can mitigate or even avoid the interference by dynamically exploiting the under-utilized licensed spectrum, and thus become more flexible to complex spectrum environment in indoor scenarios.

1.2.1.4

Multimedia Applications

Reliable and timely delivery of event features in the form of multimedia by resource-constrained sensor nodes under varying channel conditions is an extremely challenging objective due to the inherent high bandwidth demands of multimedia. CRSNs may provide sensor nodes the freedom to dynamically change communication channels according to the environmental conditions and application-specific QoS requirements in terms of bandwidth, bit error rates, and access delay. For example, as the packet travels through multiple hops, each relaying node may use higher frequencies and the highest possible data rate to provide required bandwidth.

1.2.2 Arisen Challenges for CRSNs Despite the aforementioned benefits and strong potentials, CRSN is facing significant challenges in energy-efficient spectrum resource management by the integration of CR and WSN technologies, in terms of dynamic spectrum access decision and control, spectrum resources allocation and security. 1. Energy-Efficient Channel Access Control. Due the limitations of size and cost, sensor nodes are generally powered by batteries, making energy efficiency be an important focus of CRSNs. Meanwhile, different from the sensor nodes in traditional WSNs, the battery-powered sensor nodes in CRSNs have to consume considerable energy for supporting the CR functionalities, e.g., spectrum sensing and channel switching. It consequently brings additional energy burden for CRSNs. According to the experimental results from [20–22], the power of channel switching is close to 1000 mW, and the delay of switching a channel over 1 MHz is nearly 0.1 ms. Thus, the energy consumption of channel switching can be approximately calculated as 104 J. Additionally, the average energy

1.2 Applications and Challenges of CRSN

9

Table 1.2 Energy consumption rates of typical sensor nodes

Sensor nodes uAMP [23] Mica 2 [9] Mica 2C [24]

Idling power (mW) 400 960 960

Energy consumption rate for data transmission (J/bit) 0.4 3.12 4.62

Energy consumption rate for data reception (J/bit) 0.27 2.34 2.34

consumption of spectrum sensing is non-neglectable (calculated as 1:31  104 J) and highly dependent on the radio environment. When a CRSN application is deployed in a poor radio environment, the energy consumption of spectrum sensing may grow rapidly. Since sensor nodes have a much lower energy capacity and working power than the cognitive devices in CRNs (e.g., smartphone), such part of energy consumption should be carefully considered in CRSNs. Table 1.2 shows the energy consumption rates of three typical sensor nodes. It can be seen from the table that the energy consumption of spectrum sensing and channel switching is comparable to the energy consumption of data transmission and reception, and is an important part in the whole energy consumption of CRSN. Therefore, how to efficiently control the dynamic channel sensing and switching becomes a critical and challenging research topic in CRSNs. 2. Secure and Energy-Efficient Collaborative Spectrum Sensing. Ensuring the accuracy of spectrum sensing is the foundation to guarantee the network performance of CRSN. However, due to channel fading and shadowing, spectrum sensing by individual sensor node has inevitable sensing errors, which adversely impact the performance of both the CRSN and PUs of the licensed channels. To overcome the limitation of individual spectrum sensing, collaborative spectrum sensing is generally employed in CRSNs to improve the spectrum sensing accuracy. On the other hand, due to the limited hardware capabilities, powerful security mechanisms cannot be deployed in lightweight sensor nodes, making them easily compromised by adversaries to launch various attacks [4, 25]. Spectrum sensing data falsification (SSDF) attack is one type of spectrum sensing attacks, where compromised sensor nodes may independently or collaboratively report false sensing results to mislead the channel availability decision, which can significantly reduce spectrum utilization and degrade overall network performance. To resist SSDF attacks, a number of countermeasures have been proposed in CRNs, which could also be applied into CRSNs. But, few related works consider energy efficiency in SSDF attack countermeasure design. Roughly speaking, when a CRSN has only a small portion of attackers if we recruit more sensor nodes in spectrum sensing, we can make a more accurate decision for channel availability. But, as we have mentioned above, the energy consumption of spectrum sensing is considerable for energy-sensitive sensor nodes, especially when collaborative spectrum sensing is adopted. When more sensor nodes are chosen for spectrum sensing, the energy-efficiency will be

10

1 Introduction

simultaneously degraded. Therefore, how to design a secure and energy-efficient collaborative spectrum sensing scheme to find the trade-off between security and energy efficiency becomes a challenging research focus in CRSNs. 3. Joint Energy Management and Spectrum Resource Allocation. As a type of self-organized multihop data collecting networks, WSN has some unique characteristics in terms of energy consumption and resource allocation. One of the hottest research topics in WSNs is to maximize the network utility by controlling the data sampling rate and managing the residual energy of sensor nodes. However, when sensor nodes are empowered the CR capability, the accessible spectrum resources of CRSN will be spatiotemporally-changed, leading to a dynamic network capacity. It consequently makes the network utility maximization in CRSN face the dual challenges of energy management and dynamic spectrum resource allocation. Moreover, different from the spectrum resource allocation in traditional CRNs where CR devices can directly communicate with each other or the base station via a one-hop wireless network, spectrum resource allocation in CRSNs becomes more challenging due to the many-toone multihop network topology. We have to avoid the co-channel interference of neighboring sensor nodes and consider the changing network topology caused by dynamic routing. These challenges require researchers to design dedicated energy management and spectrum resource allocation schemes for CRSNs.

1.3 Aim and Outline of This Book As a promising networking solution to address the spectrum scarcity problem in the IoT era, CRSN can provide numerous opportunities and benefits, and has the potential be used in a wide variety of IoT applications. However, challenges are always followed by opportunities. The integration of CR and WSN also poses great challenges for CRSNs in energy efficiency and spectrum resource management. Without some sophisticated solutions to address these challenging problems, the development and flourish of CRSN applications will be significantly impeded. In this monograph, we aim to provide a comprehensive and in-depth study on the energy-efficient spectrum management in CRSNs. We introduce the overall architecture of CRSN and some promising CRSN applications, and then clearly identify the arisen challenges of CRSNs with full consideration of the characteristics of WSNs. To address these challenges, an integrated energy-efficient spectrum management solution, consisting of energy-efficient spectrum sensing and accessing control, secure and energy-efficient collaborative spectrum sensing, and available spectrum resource allocation, is presented to show some design insights and guidelines for CRSN applications. The reminder of this monograph is organized as follows. Chapter 2 provides a comprehensive review on the state-of-the-art research literature, which covers most of existing works in energy-efficient spectrum resource management in CRSNs, including dynamic spectrum access decision, secure cooperative spectrum sensing and spectrum resource allocation.

1.3 Aim and Outline of This Book

11

Chapter 3 studies the dynamic spectrum access decision problem and presents two dynamic channel sensing and accessing schemes to improve energy efficiency in clustered CRSNs [26]. Since the energy consumption of spectrum sensing and channel switching is considerable for sensor nodes, we carefully consider it to make channel sensing and accessing decisions for energy efficiency improvement. We determine the conditions when sensor nodes should sense and switch to a licensed channel for improving the energy efficiency for both intra-cluster and inter-cluster data transmission, according to the packet loss rate of the license-free channel. Finally, two dynamic channel accessing schemes are developed to identify the channel sensing and switching sequences for intra-cluster and inter-cluster data transmission, respectively. Chapter 4 investigates the joint channel access and sampling rate control problem to maximize the network utility in EH-CRSNs [27]. Different from traditional sensor networks, sensor nodes in CRSNs are equipped with cognitive radio modules, enabling them to dynamically access the licensed channels. Since the dynamic channel access is critical to guarantee the network capacity for CRSNs, existing solutions without considering the dynamic channel access cannot be directly applied into CRSNs. To this end, we aim at maximizing the network utility by jointly controlling the sampling rates and channel access of sensor nodes, under the energy consumption, channel capacity and interference constraints. With the consideration of fluctuated energy harvesting rates and channel switching costs, we formulate the network utility maximization as a mix-integer non-linear programming problem and solve it in an efficient and decoupled way by means of dual decomposition. A joint channel access and sampling rate control scheme, named JASC, is then designed to maximize the network utility by utilizing the real-time channel sensing results and energy harvesting rates. Chapter 5 focuses on how to design a secure and energy-efficient collaborative spectrum scheme to resist Spectrum Sensing Data Falsification (SSDF) attacks and enhance the energy efficiency in CRSNs [28]. Specifically, we theoretically analyze the impacts of two types of attacks, i.e., independent and collaborative SSDF attacks, on the accuracy of collaborative spectrum sensing in a probabilistic way. To maximize the energy efficiency of spectrum sensing, we calculate the minimum number of sensor nodes needed for spectrum sensing to guarantee the desired accuracy of sensing results. Moreover, a trust evaluation scheme, named FastDtec, is developed to evaluate the spectrum sensing behaviors and fast identify compromised nodes. Based on that, a secure and energy-efficient collaborative spectrum sensing scheme is then presented to further improve the energy efficiency of collaborative spectrum sensing, by adaptively isolating the identified compromised nodes from spectrum sensing. Finally, Chap. 6 points out some open problems and future research directions to foster continuous studies in CRSNs.

12

1 Introduction

References 1. I.F. Akyildiz, W. Su, Y. Sankarasubramaniam, E. Cayirci, Wireless sensor networks: a survey. Comput. Netw. 38(4), 393–422 (2002) 2. Y. Zhang, J. Ren, J. Liu, C. Xu, H. Guo, Y. Liu, A survey on emerging computing paradigms for big data. Chin. J. Electron. 26(1), 1–13 (2017) 3. J. Ren, Y. Zhang, K. Zhang, A. Liu, J. Chen, X. Shen, Lifetime and energy hole evolution analysis in data-gathering wireless sensor networks. IEEE Trans. Ind. Inf. 12(2), 788–800 (2016) 4. J. Ren, Y. Zhang, K. Zhang, X. Shen, Adaptive and channel-aware detection of selective forwarding attacks in wireless sensor networks. IEEE Trans. Wirel. Commun. 15(5), 3718–3731 (2016) 5. A.-J. Garcia-Sanchez, F. Garcia-Sanchez, F. Losilla, P. Kulakowski, J. Garcia-Haro, A. Rodríguez, J.-V. López-Bao, F. Palomares, Wireless sensor network deployment for monitoring wildlife passages. Sensors 10(8), 7236–7262 (2010) 6. M. Li, Y. Liu, Underground coal mine monitoring with wireless sensor networks. ACM Trans. Sensor Netw. 5(2), 10 (2009) 7. A. Liu, J. Ren, X. Li, Z. Chen, X.S. Shen, Design principles and improvement of cost function based energy aware routing algorithms for wireless sensor networks. Comput. Netw. 56(7), 1951–1967 (2012) 8. J.L. Hill, D.E. Culler, Mica: a wireless platform for deeply embedded networks. IEEE Micro 22(6), 12–24 (2002) 9. J. Polastre, J. Hill, D. Culler, Versatile low power media access for wireless sensor networks, in Proceedings of the 2nd International Conference on Embedded Networked Sensor Systems (ACM, New York, 2004), pp. 95–107 10. MEMSIC Inc., MICAz 2.4 GHz sensor node, http://www.memsic.com/userfiles/files/ DataSheets/WSN/micazdatasheet-t.pdf 11. CROSSBOW Inc., IMote sensor node, http://wsn.cse.wustl.edu/images/e/e3/Imote2Datasheet. pdf 12. MEMSIC Inc., Iris sensor node, http://www.memsic.com/userfiles/files/Datasheets/WSN/ IRISDatasheet.pdf 13. Eyes project, www.tinyos.net/ttx-02-2005/platforms/ttx2005-eyesIFX.ppt 14. E.Z. Tragos, S. Zeadally, A.G. Fragkiadakis, V.A. Siris, Spectrum assignment in cognitive radio networks: a comprehensive survey. IEEE Commun. Surv. Tutorials 15(3), 1108–1135 (2013) 15. J. Mitola, G.Q. Maguire, Cognitive radio: making software radios more personal. IEEE Pers. Commun. 6(4), 13–18 (1999) 16. M. Naeem, A. Anpalagan, M. Jaseemuddin, D.C. Lee, Resource allocation techniques in cooperative cognitive radio networks. IEEE Commun. Surv. Tutorials 16(2), 729–744 (2014) 17. O.B. Akan, O. Karli, O. Ergul, Cognitive radio sensor networks. IEEE Netw. 23(4), 34–40 (2009) 18. D. Zhang, Z. Chen, H. Zhou, X.S. Shen, Resource Management for Energy and Spectrum Harvesting Sensor Networks (Springer, Berlin, 2017) 19. K. Zhang, J. Ni, K. Yang, X. Liang, J. Ren, X.S. Shen, Security and privacy in smart city applications: challenges and solutions. IEEE Commun. Mag. 55(1), 122–129 (2017) 20. S. Bayhan, F. Alagoz, Scheduling in centralized cognitive radio networks for energy efficiency. IEEE Trans. Veh. Technol. 62(2), 582–595 (2013) 21. S. Maleki, A. Pandharipande, G. Leus, Energy-efficient distributed spectrum sensing for cognitive sensor networks. IEEE Sensors J. 11(3), 565–573 (2011) 22. E. Chatziantoniou, Spectrum sensing and occupancy prediction for cognitive machine-tomachine wireless networks (2014) 23. J. Hill, R. Szewczyk, A. Woo, S. Hollar, D. Culler, K. Pister, Architecture directions for networked sensors, in Nine International Symposium on Architectural Support for Programming Languages and Operating Systems (ASPLOS IX) (2000)

References

13

24. V. Shnayder, M. Hempstead, B.-R. Chen, G.W. Allen, M. Welsh, Simulating the power consumption of large-scale sensor network applications, in Proceedings of the 2nd International Conference on Embedded Networked Sensor Systems (ACM, New York, 2004), pp. 188–200 25. J. Ren, Y. Zhang, K. Liu, Multiple k-hop clusters based routing scheme to preserve sourcelocation privacy in WSNs. J. Cent. South Univ. 21, 3155–3168 (2014) 26. J. Ren, Y. Zhang, N. Zhang, D. Zhang, X. Shen, Dynamic channel access to improve energy efficiency in cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 15(5), 3143–3156 (2016) 27. J. Ren, Y. Zhang, R. Deng, N. Zhang, D. Zhang, X. Shen, Joint channel access and sampling rate control in energy harvesting cognitive radio sensor networks, in IEEE Transactions on Emerging Topics in Computing (2016). doi:10.1109/TETC.2016.2555806 28. J. Ren, Y. Zhang, Q. Ye, K. Yang, K. Zhang, X.S. Shen, Exploiting secure and energyefficient collaborative spectrum sensing for cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 15(10), 6813–6827 (2016)

Chapter 2

Spectrum Resource Management for CRSNs

The coexistence of various overlapping wireless applications causes severe inter-network interference to each other and leads to increasingly poor network performance. It motivates researchers to find new solutions for wireless sensor networking. The emergence of CRSN is a consequent result, which attracts growing attentions from both the academia and industry and experiences rapid development during the recent years. But different from traditional cognitive radio networks (CRNs), the integration of CR and WSN poses some new challenges in CRSNs. One of the most critical challenges is to design energy-efficient spectrum resource management solutions for CRSNs by taking the energy consumption and network characteristics of WSNs into consideration. To accelerate the flourish of CRSN applications, researchers have devoted great efforts on tackling this critical and urgent challenge. In this chapter, we provide a comprehensive review on the state-ofthe-art research literature, which covers most of existing works in energy-efficient spectrum resource management in CRSNs, including dynamic spectrum access decision, secure cooperative spectrum sensing and spectrum resource allocation.

2.1 Dynamic Spectrum Access Decision The most significant benefit of CRSN is enabling sensor nodes to opportunistically access the under-utilized licensed spectrum to avoid inter-network interference and improve the network performance. Therefore, after the emergence of CRSN, a large number of existing works focus on improving the network performance, such as end-to-end delay and network throughput. Liang et al. [1] analyze the delay performance to support real-time traffic in CRSNs. They derive the average packet transmission delay for two types of channel switching mechanisms, namely periodic switching and triggered switching, under two kinds of real-time traffic, including periodic data traffic and Poisson

© Springer International Publishing AG 2018 J. Ren et al., Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks, DOI 10.1007/978-3-319-60318-6_2

15

16

2 Spectrum Resource Management for CRSNs

traffic, respectively. Bicen et al. [2] provide several principles for delay-sensitive multimedia communication in CRSNs through extensive simulations. A greedy networking algorithm is proposed in [3] to enhance the end-to-end delay and network throughput for CRSNs, by leveraging distributed source coding and broadcasting. Since the QoS performances of sensor networks can be significantly impacted by routing schemes, research efforts are also devoted in developing dynamic routing for CRSNs [4, 5]. Quang and Kim [4] propose a throughput-aware routing algorithm to improve network throughput and decrease end-to-end delay for a large-scale clustered CRSN based on ISA100.11a. In addition, opportunistic medium access (MAC) protocol design and performance analysis of existing MAC protocols for CRSNs are studied in [6, 7]. Most of the aforementioned works can effectively improve the network performances for various WSNs applications, and also provide a foundation for spectrum management and resource allocation in CRSNs. However, as a senor network composed of resource-limited and energy-constrained sensor nodes, CRSN is still facing an inherent challenge on energy efficiency, which attracts increasing attention to study the energy efficiency enhancement. Han et al. [8] develop a channel management scheme for CRSNs, which can adaptively select the operation mode of the network in terms of channel sensing, channel switching, and data transmission/reception, for energy efficiency improvement according to the outcome of channel sensing. The optimal packet size is studied in [9] to maximize energy efficiency while maintaining acceptable interference level for PUs and achieving reliable event detection in CRSNs. The transmission power of sensor nodes can also be adjusted for improving the energy efficiency of data transmission. In [10], Chai et al. propose a power allocation algorithm for sensor nodes to achieve satisfactory performance in terms of energy efficiency, convergence speed and fairness in CRSNs. Meanwhile, since spectrum sensing accounts for a certain portion of energy consumption for CRSNs, energyefficient spectrum sensing schemes are also studied in CRSNs to improve the spectrum detection performance [11–13]. Furthermore, motivated by the superior energy efficiency of clustered WSNs, spectrum-aware clustering strategies are investigated in [14, 15] to enhance energy efficiency and spectrum utilization for CRSNs. However, a comprehensive study on energy-efficient data gathering is particularly important for CRSNs and is still missing in the literature. It requires CRSN to jointly consider the energy consumption in channel sensing and switching, channel detection probability and PU protection to make dynamic channel sensing and switching decision.

2.2 Secure Collaborative Spectrum Sensing Spectrum sensing is one of the basic functionalities of CRSNs to accurately identify available spectrum bands. However, due to channel fading and shadowing, spectrum sensing by individual sensor node has inevitable sensing errors, which adversely

2.2 Secure Collaborative Spectrum Sensing

17

impact the performance of both the CRSN and the primary users (PUs) of the licensed channels. To overcome the limitation of individual spectrum sensing, collaborative spectrum sensing is generally employed in CRSNs to improve the spectrum sensing accuracy [16]. But since sensor nodes may be compromised by adversaries, these nodes can send false sensing results, making CRSNs vulnerable to spectrum sensing data falsification (SSDF) attack [17, 18]. Recently, increasing attention has been paid to securing spectrum sensing against SSDF attacks [16]. As most of the SSDF attack countermeasures in CRNs can be effectively applied to CRSNs, we here briefly review some existing countermeasures in CRNs by dividing them into three main categories: trust/reputation based approaches, abnormal statistical-behavior detection based approaches, and clustering based approaches. Trust and reputation based approaches are the most widely studied techniques in the literature [19–22]. The main idea of these approaches is to update the trust values of spectrum sensing nodes according to their historical sensing behaviors, and design weighted decision making strategies to resist SSDF attacks based on the evaluated trust values. Qin et al. [19] propose a trust-based model and design a weighted sensing result aggregation scheme to remove attackers from the decision making process. In [21], Rawat et al. analyze the performance limits of collaborative spectrum sensing under independent and collaborative SSDF attacks, respectively, and then propose a simple reputation-based scheme to resist SSDF attacks. They prove that, by employing the theory of Kullback-Leibler divergence metric, a certain fraction of attackers can make collaborative spectrum sensing no better than random guess. Another category of promising countermeasures is to identify SSDF attackers by detecting their abnormal statistical spectrum sensing behaviors [23–27]. From this perspective, two hidden Markov models (HMMs), with respect to honest and malicious users, are adopted in [23] to characterize their different sensing behaviors. Attackers are identified by the difference in the corresponding HMM parameters. He et al. [25] use two conditional frequency check statistics to identify SSDF attackers based on the Markovian model of spectrum state. In addition, several recent research works focus on securing the collaborative spectrum sensing based on clustering the cooperating sensors [28–30]. Hyder et al. [29] develop a reputation-based clustering algorithm to divide nodes into a number of virtual clusters based on their evaluated reputation values. A bi-level voting strategy, consisting of intra-cluster and intercluster decision voting, is then proposed to make the final decision. There are also some works resisting SSDF attacks by using consensus-based approaches [31– 33], radio propagation characteristics [34], data cleansing approaches [35], and incentive-based mechanisms [36]. Despite of their effectiveness in resisting SSDF attacks, most of existing works do not consider the energy efficiency of SSDF attack countermeasures [37, 38]. In a CRSN, sensor nodes have to consume considerable energy for collaborative spectrum sensing, which may degrade the energy efficiency of the network. However, if the unlicensed channel is facing significant interference when a CRSN starts to collect data, the performance of data collection would be hardly guaranteed by keeping working on the channel. It makes sensing and accessing a licensed channel as a necessary way to guarantee the network performance. Meanwhile, it

18

2 Spectrum Resource Management for CRSNs

also motivates us to carefully consider the energy efficiency in spectrum sensing and SSDF countermeasure design. Several recent works pay attention to the significance of energy efficiency in resisting SSDF attacks for CRNs. In [39], a low-overhead scheme is proposed for CRNs to address the always-1 SSDF attack under a trade-off between security and the energy efficiency of sensing report transmission. Recently, Mousavifar and Leung [38] develop a collaborative spectrum sensing scheme based on trust management to resist independent SSDF attacks in a CRN with a fixed number of compromised nodes and honest nodes. It is verified as effective to reduce the required sensing reports to achieve a targeted accuracy requirement. However, as more powerful SSDF attacks than independent SSDF attacks, collaborative SSDF attacks, where compromised nodes can collaboratively launch SSDF attacks, are not studied in both of the aforementioned works.

2.3 Joint Energy Management and Spectrum Resource Allocation For most of data gathering CRSN applications, network utility is an important metric to evaluate data collection efficiency. According to the definition, network utility greatly depends on the sampling rates of sensor nodes and network capacity. Meanwhile, in CRSNs, sampling rate control is constrained by the limited energy of sensor nodes and network capacity is determined by the dynamically-changing spectrum holes. Thus, how to jointly manage the energy of sensor nodes and allocate the available spectrum resource for network utility maximization becomes a key research focus in CRSNs. In the past decade, network utility maximization by energy management attracted considerable attention in traditional energy harvesting WSNs. Liu et al. [40] propose a QuickFix algorithm to maximize the network utility by determining the sampling rates and routes for sensor nodes. To adjust the sampling rates adaptive to battery levels, a local algorithm, called SanpIt, is developed to sustain the network operation. Deng et al. [41] investigate the network utility maximization problem with spatiotemporally-coupled constraints in rechargeable sensor networks. To address the problem, they propose a distributed algorithm to jointly optimize the sampling rates and battery levels to achieve the globally optimal solution. Different from the aforementioned works maximizing the network utility based on predictable energy harvesting rates, Chen et al. [42] propose an online solution to address the energy allocation and routing problem for maximizing the network utility without prior knowledge of the replenishment profile. Huang and Neely [43] investigate the general network utility maximization problem in energy harvesting networks. They propose an online algorithm to jointly manage the harvested energy and adjust transmit power to optimize the time-average expected network utility and guarantee the network stability and energy supply. Besides the discussed works, Zhao et al. [42] develop a distributed algorithm to adjust the sampling rates of

2.4 Summary

19

sensor nodes for maximizing network utility, by leveraging a mobile entity as mobile data collector and energy transporter. In [44] and [45], limited battery capacity is considered and studied for the optimization of sampling rates in energy harvesting WSNs. Most of the aforementioned solutions can effectively improve the network utility and manage the harvested energy in traditional energy harvesting WSNs [46]. However, since the available spectrum resources that are dynamically changing in both space and time becomes an important impact factor of maximizing the network utility in CRSNs, existing solutions cannot be directly applied into CRSNs. Although there is little attention paid to studying the joint energy management and spectrum resource allocation for network utility maximization in CRSNs, some research works focusing on dynamic spectrum resource allocation and channel access can provide valuable insights for data gathering CRSN applications. Liang et al. [1] analyze the delay performance of dynamic channel access to support real-time traffic in CRSNs. They derive the average packet transmission delay for two types of channel switching mechanisms, called periodic switching and triggered switching, under two kinds of real-time traffic, including periodic data traffic and Poisson traffic, respectively. Quang and Kim [4] develop a throughputaware routing scheme to improve network throughput and decrease end-to-end delay for a large-scale clustered CRSN based on ISA100.11a. Han et al. [8] propose a channel management scheme for CRSNs, which can adaptively choose the operation mode of the network in terms of channel sensing, channel switching, and data transmission/reception, for energy efficiency improvement according to the outcome of channel sensing. In addition, opportunistic medium access (MAC) protocol design and performance analysis are studied to support the dynamic channel access for CRSNs in [6, 7, 47]. Motivated by the superior energy efficiency of clustering, dynamic spectrum-aware clustering strategies are also investigated in [14, 15] to improve energy efficiency and spectrum utilization for CRSNs.

2.4 Summary This chapters presents a comprehensive review on energy-efficient spectrum resource management in CRSNs. As energy efficiency becomes a more and more important concern in the IoT era, growing research efforts will be continuously devoted in this field of study. Through the overview of the state-of-the-art literature, we can find that few of existing works has fully considered the energy consumption and network characteristics of WSN in managing the spectrum resources to improve the energy efficiency and network performance of CRSNs. For example, sensor nodes have to consume considerable energy in spectrum sensing and channel switching, which is neglected by most of existing works. Moreover, as a manyto-one (or many-to-several) multihop wireless network, CRSN has some unique network characteristics in terms of energy consumption and dynamic routing, etc. Since available spectrum resource of CRSN is spatiotemporally stochastic, it

20

2 Spectrum Resource Management for CRSNs

consequently makes the spectrum management in CRSN more challenging. In the following chapters of this monograph, we will present several promising solutions to address these challenges and provide some insights and guidances for designing energy-efficient CRSN applications.

References 1. Z. Liang, S. Feng, D. Zhao, X. Shen, Delay performance analysis for supporting real-time traffic in a cognitive radio sensor network. IEEE Trans. Wirel. Commun. 10(1), 325–335 (2011) 2. A.O. Bicen, V.C. Gungor, O.B. Akan, Delay-sensitive and multimedia communication in cognitive radio sensor networks. Ad Hoc Netw. 10(5), 816–830 (2012) 3. S.-C. Lin K.-C. Chen, Improving spectrum efficiency via in-network computations in cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 13(3), 1222–1234 (2014) 4. P.T.A. Quang, D.-S. Kim, Throughput-aware routing for industrial sensor networks: application to ISA100. 11a. IEEE Trans. Ind. Inf. 10(1), 351–363 (2014) 5. P. Spachos D. Hantzinakos, Scalable dynamic routing protocol for cognitive radio sensor networks. IEEE Sensors J. 14(7), 2257–2266 (2014) 6. G.A. Shah, O.B. Akan, Performance analysis of CSMA-based opportunistic medium access protocol in cognitive radio sensor networks. Ad Hoc Netw. 15, 4–13 (2014) 7. G. Shah, O. Akan, Cognitive adaptive medium access control in cognitive radio sensor networks. IEEE Trans. Veh. Technol. 64(2), 757–767 (2015) 8. J.A. Han, W.S. Jeon, D.G. Jeong, Energy-efficient channel management scheme for cognitive radio sensor networks. IEEE Trans. Veh. Technol. 60(4), 1905–1910 (2011) 9. M.C. Oto, O.B. Akan, Energy-efficient packet size optimization for cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 11(4), 1544–1553 (2012) 10. B. Chai, R. Deng, P. Cheng, J. Chen, Energy-efficient power allocation in cognitive sensor networks: a game theoretic approach, in IEEE Proceedings of GLOBECOM (2012), pp. 416–421 11. G. Ding, J. Wang, Q. Wu, F. Song, Y. Chen, Spectrum sensing in opportunity-heterogeneous cognitive sensor networks: how to cooperate? IEEE Sensors J. 13(11), 4247–4255 (2013) 12. S. Maleki, A. Pandharipande, G. Leus, Energy-efficient distributed spectrum sensing for cognitive sensor networks. IEEE Sensors J. 11(3), 565–573 (2011) 13. J. Ren, Y. Zhang, Q. Ye, K. Yang, K. Zhang, X.S. Shen, Exploiting secure and energyefficient collaborative spectrum sensing for cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 15(10), 6813–6827 (2016) 14. G.A. Shah, F. Alagoz, E.A. Fadel, O.B. Akan, A spectrum-aware clustering for efficient multimedia routing in cognitive radio sensor networks. IEEE Trans. Veh. Technol. 63(7), 3369–3380 (2014) 15. M. Ozger O. Akan, Event-driven spectrum-aware clustering in cognitive radio sensor networks, in IEEE Proceedings of INFOCOM (2013), pp. 1483–1491 16. R. Sharma D. Rawat, Advances on security threats and countermeasures for cognitive radio networks: a survey. IEEE Commun. Surv. Tutorials 17(2), 1023–1043 (2015) 17. Y. Cai, Y. Mo, K. Ota, C. Luo, M. Dong, L. Yang, Optimal data fusion of collaborative spectrum sensing under attack in cognitive radio networks. IEEE Netw. 28(1), 17–23 (2014) 18. Z. Gao, H. Zhu, S. Li, S. Du, X. Li, Security and privacy of collaborative spectrum sensing in cognitive radio networks. IEEE Wirel. Commun. 19(6), 106–112 (2012) 19. T. Qin, H. Yu, C. Leung, Z. Shen, C. Miao, Towards a trust aware cognitive radio architecture. ACM SIGMOBILE Mob. Comput. Commun. Rev. 13(2), 86–95 (2009) ˇ 20. K. Zeng, P. Pawełczak, D. Cabri´ c, Reputation-based cooperative spectrum sensing with trusted nodes assistance. IEEE Commun. Lett. 14(3), 226–228 (2010)

References

21

21. A.S. Rawat, P. Anand, H. Chen, P.K. Varshney, Collaborative spectrum sensing in the presence of Byzantine attacks in cognitive radio networks, IEEE Trans. Signal Process. 59(2), 774–786 (2011) 22. R. Chen, J.-M. Park, K. Bian, Robust distributed spectrum sensing in cognitive radio networks, in Proceedings of IEEE INFOCOM (2008) 23. X. He, H. Dai, P. Ning, HMM-based malicious user detection for robust collaborative spectrum sensing. IEEE J. Sel. Areas Commun. 31(11), 2196–2208 (2013) 24. J. Wang, J. Yao, Q. Wu, Stealthy-attacker detection with a multidimensional feature vector for collaborative spectrum sensing. IEEE Trans. Veh. Technol. 62(8), 3996–4009 (2013) 25. X. He, H. Dai, P. Ning, A Byzantine attack defender in cognitive radio networks: the conditional frequency check. IEEE Trans. Wirel. Commun. 12(5), 2512–2523 (2013) 26. F. Penna, Y. Sun, L. Dolecek, D. Cabric, Detecting and counteracting statistical attacks in cooperative spectrum sensing. IEEE Trans. Signal Process. 60(4), 1806–1822 (2012) 27. Z. Qin, Q. Li, G.-C. Hsieh, Defending against cooperative attacks in cooperative spectrum sensing. IEEE Trans. Wirel. Commun. 12(6), 2680–2687 (2013) 28. M. Ghaznavi, A. Jamshidi, A reliable spectrum sensing method in the presence of malicious sensors in distributed cognitive radio network. IEEE Sensors J. 15(3), 1810–1816 (2015) 29. C. Hyder, B. Grebur, L. Xiao, M. Ellison, ARC: adaptive reputation based clustering against spectrum sensing data falsification attacks. IEEE Trans. Mob. Comput. 13(8), 1707–1719 (2014) 30. G. Ding, Q. Wu, Y.-D. Yao, J. Wang, Y. Chen, Kernel-based learning for statistical signal processing in cognitive radio networks: theoretical foundations, example applications, and future directions. IEEE Signal Process. Mag. 30(4), 126–136 (2013) 31. S. Liu, H. Zhu, S. Li, X. Li, C. Chen, X. Guan, An adaptive deviation-tolerant secure scheme for distributed cooperative spectrum sensing, in Proceedings of IEEE GLOBECOM (2012), pp. 603–608 32. Q. Yan, M. Li, T. Jiang, W. Lou, Y.T. Hou, Vulnerability and protection for distributed consensus-based spectrum sensing in cognitive radio networks, in Proceedings of IEEE INFOCOM (2012), pp. 900–908 33. H. Tang, F. R. Yu, M. Huang, Z. Li, Distributed consensus-based security mechanisms in cognitive radio mobile ad hoc networks. IET Commun. 6(8), 974–983 (2012) 34. S. Liu, Y. Chen, W. Trappe, L.J. Greenstein, ALDO: an anomaly detection framework for dynamic spectrum access networks, in Proceedings of IEEE INFOCOM (2009), pp. 675–683 35. G. Ding, J. Wang, Q. Wu, L. Zhang, Y. Zou, Y.-D. Yao, Y. Chen, Robust spectrum sensing with crowd sensors. IEEE Trans. Commun. 62(9), 3129–3143 (2014) 36. S. Sodagari, A. Attar, V. Leung, S.G. Bilén, Denial of service attacks in cognitive radio networks through channel eviction triggering, in Proceedings of IEEE GLOBECOM (2010), pp. 1–5 37. S. Althunibat, V. Sucasas, H. Marques, J. Rodriguez, R. Tafazolli, F. Granelli, On the trade-off between security and energy efficiency in cooperative spectrum sensing for cognitive radio. IEEE Commun. Lett. 17(8), 1564–1567 (2013) 38. S.A. Mousavifar, C. Leung, Energy efficient collaborative spectrum sensing based on trust management in cognitive radio networks. IEEE Trans. Wirel. Commun. 14(4), 1927–1939 (2015) 39. S. Althunibat, R. Palacios, F. Granelli, Energy-efficient spectrum sensing in cognitive radio networks by coordinated reduction of the sensing users, in Proceedings of IEEE ICC (2012), pp. 1399–1404 40. R.-S. Liu, P. Sinha, C.E. Koksal, Joint energy management and resource allocation in rechargeable sensor networks, in IEEE Proceedings of INFOCOM (2010), pp. 1–9 41. R. Deng, Y. Zhang, S. He, J. Chen, X. Shen, Maximizing network utility of rechargeable sensor networks with spatiotemporally-coupled constraints. IEEE J. Sel. Areas Commun. (2016, to appear). doi:10.1109/JSAC.2016.2520181 42. S. Chen, P. Sinha, N.B. Shroff, C. Joo, A simple asymptotically optimal joint energy allocation and routing scheme in rechargeable sensor networks. IEEE/ACM Trans. Netw. 22(4), 1325–1336 (2014)

22

2 Spectrum Resource Management for CRSNs

43. L. Huang, M.J. Neely, Utility optimal scheduling in energy-harvesting networks. IEEE/ACM Trans. Netw. 21(4), 1117–1130 (2013) 44. Z. Mao, C.E. Koksal, N.B. Shroff, Near optimal power and rate control of multi-hop sensor networks with energy replenishment: basic limitations with finite energy and data storage. IEEE Trans. Autom. Control 57(4), 815–829 (2012) 45. Y. Zhang, S. He, J. Chen, Y. Sun, X. Shen, Distributed sampling rate control for rechargeable sensor nodes with limited battery capacity. IEEE Trans. Wirel. Commun. 12(6), 3096–3106 (2013) 46. C. Huang, R. Zhang, S. Cui, Optimal power allocation for outage probability minimization in fading channels with energy harvesting constraints. IEEE Trans. Wirel. Commun. 13(2), 1074–1087 (2014) 47. J. Ren, Y. Zhang, N. Zhang, D. Zhang, X. Shen, Dynamic channel access to improve energy efficiency in cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 15(5), 3143–3156 (2016)

Chapter 3

Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

Cognitive Radio (CR) has emerged as a promising technology to improve the spectrum utilization by enabling opportunistic access to the licensed spectrum bands [1]. This technology can also be applied to WSNs, which leads to Cognitive Radio Sensor Networks (CRSNs) [2]. Sensor nodes in CRSNs can sense the availability of licensed channels and adjust the operation parameters to access the idle ones, when the condition of the licensed-free channel degrades. However, since the energy consumption for supporting the CR functionalities, e.g., channel sensing and switching, is critical for battery-powered sensor nodes [3, 4], the opportunistic channel access should be carefully studied to improve the energy efficiency in CRSNs. As we mentioned in Chap. 2, existing works provide a comprehensive and in-depth investigation on optimizing the quality-of-service (QoS) performances for CRSNs, such as reducing the transmission delay [5–7] or increasing the network capacity [8, 9]. However, few of them have paid attention to improving the energy efficiency for CRSNs, with a delicate consideration of the energy consumption in channel sensing and switching. In order to enhance energy efficiency, the key issue is to determine when the energy consumption of transmitting a fixed amount of data can be reduced by sensing and accessing a licensed channel, compared with the energy consumption when only using the default license-free channel. It is very challenging since the decision depends on different factors, including the packet loss rate of the license-free channel, the probabilities for accessing licensed channels, as well as the protection for primary users (PUs). Moreover, due to the dynamic availability of licensed channels, when sensor nodes decide to sense and access a licensed channel, another challenge lies in identifying the best licensed channel to sense and access to optimize the energy efficiency for data transmission. In this chapter, we investigate the opportunistic channel accessing problem to improve energy efficiency in clustered CRSNs [10]. Sensor nodes form a number of clusters and periodically transmit their sensed data to the sink via hierarchical routing. They work on a license-free channel but are also able to access idle licensed © Springer International Publishing AG 2018 J. Ren et al., Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks, DOI 10.1007/978-3-319-60318-6_3

23

24

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

channels when the packet loss rate over the license-free channel increases. To protect the PUs sufficiently, the channel available duration (CAD) is limited for each licensed channel when it is detected as idle. Then, we analyze the expected energy consumption to determine if sensor nodes can reduce their energy consumption by accessing a licensed channel, considering the energy consumption in channel sensing and switching. Furthermore, to tackle the opportunistic availability of licensed channels, two sequential channel sensing and accessing schemes with the resource allocation of an accessed channel are exploited for minimizing the energy consumption in both intra- and inter-cluster data transmission. Specifically, the main content and contributions of this chapter are threefold. • For both intra-cluster and inter-cluster data transmission, we determine the condition when sensor nodes should sense and switch to a licensed channel for potential energy consumption reduction. • We propose a dynamic channel accessing scheme to reduce the energy consumption for intra-cluster data transmission, which identifies the sensing and accessing sequence of the licensed channels within each cluster. • Based on the analysis of intra-cluster data transmission, a joint power allocation and channel accessing scheme is developed for inter-cluster data transmission, which can dynamically adjust the transmission power of cluster heads and determine the channel sensing and accessing sequence to reduce energy consumption. The remainder of this chapter is organized as follows. Section 3.1 introduces the system model and problem statement. In Sect. 3.2, we provide a detailed analysis of energy consumption for channel sensing decision and propose a dynamic channel sensing and accessing scheme for intra-cluster data transmission. Section 3.3 presents a joint power allocation and channel accessing scheme for inter-cluster data transmission. Simulation results are provided in Sect. 3.4 to evaluate the performance of the proposed schemes. Finally, Sect. 3.5 concludes the chapter.

3.1 System Model and Objective Statement 3.1.1 Network Model Consider a cognitive radio sensor network, where a set of cognitive sensor nodes N D fs1 ; : : : ; sn g are distributed to monitor the area of interest, as shown in Fig. 3.1. According to the application requirements, sensor nodes periodically sense the environment with different sampling rates and then report their sensed data to the sink node [11]. We divide the operation process of the network into a large number of data periods. A data period is composed of data sensing, data transmission, and sleeping durations, where sensor nodes sense the monitored area, transmit the sensed data to the sink node, and then sleep, respectively. Motivated by the benefits of hierarchical data gathering, sensor nodes form a number of clusters, denoted by L D fL1 ; : : : ; Lm g, to transmit the sensed data to the sink [12]. Denote the cluster head (CH) of Li as Hi , and the set of cluster members (CMs) in Li as Ni .

3.1 System Model and Objective Statement

25

PU

CM Sink

Base Station

CM CH Cluster CH Cluster CH PU

PU

Cluster

Primary Network

Fig. 3.1 The architecture of CRSN

The data transmission is further divided into two phases: intra-cluster data transmission and inter-cluster data transmission. In the intra-cluster data transmission, CMs directly transmit their sensed data to the cluster heads in a Time Division Multiple Access (TDMA) manner. During the inter-cluster data transmission, CHs aggregate the sensed data and directly send the aggregated intra-cluster data to the sink. The inter-cluster data transmission is also based on a TDMA manner, coordinated by the sink. The sensor network operates on a license-free channel C0 for data transmission, which may occasionally suffer from uncontrolled interference causing a significant packet loss rate. Enabled by the cognitive radio technique, sensor nodes can sense the licensed channels and access the vacant ones, when the packet loss rate of C0 is fairly high. There is only one radio within each sensor node for data communication, which means sensor nodes can only access one channel at a time. Moreover, similar to most existing works [13, 14], we assume that sensor nodes use a network-wide common control channel for control signaling and channel access coordination.

3.1.2 Cognitive Radio Model Suppose that there are k different licensed data channels C D fC1 ; : : : ; Ck g with different bandwidths fB1 ; : : : ; Bk g in the primary network. The PU’s behavior is assumed to be stationary and ergodic over the k channels. The cognitive sensor nodes in the primary network are secondary users (SUs) that can opportunistically access the idle channels. A fixed common control channel is considered to be available to exchange the control information among the sensor nodes and the sink. We model the PU traffic as a stationary exponential ON/OFF random process [1, 15].

26

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

The ON state indicates that channel is occupied by PUs and the OFF state implies that the channel is idle. Let Vx and Lx be the exponential random variables, describing the idle and occupancy durations of Cx with means vx and lx , respectively. Thus, for each channel Cx , the probability of channel being idle pxoff and the probability of channel occupancy pxon are (

pxoff D vx =.vx C lx /;

H0;x

pxon D lx =.vx C lx /;

H1;x

;

(3.1)

where H0;x and H1;x represent the hypothesis that Cx is idle and occupied, respectively. Sensor nodes are assumed to sense channel by the energy detectionbased spectrum sensing approach [14]. When sj adopts energy detector to sense Cx , the detection probability pd;x;j (i.e., the probability of an occupied channel being determined to be occupied correctly) and the false alarm probability pf ;x;j (i.e., the probability of an idle channel being determined as occupied) are defined as pd;x;j D Pr.Dx  ıx jH1;x / and pf ;x;j D Pr.Dx  ıx jH0;x /, where ıx is the detection threshold and Dx is the test statistic for Cx . And the misdetection probability can be calculated as pm;x;j D Pr.Dx < ıx jH1;x / D 1  pd;x;j . According to the analysis of Liang the false alarm probability of sj  et al. [16], p  ıx for Cx can be given by pf ;x;j D Q  2  1 'fs , where x2 is the variance of x the Gaussian noise; ' is the sensing duration; fs is the sampling frequency and Q./ is the complementary distribution function ofthe standard Gaussian. The  detection q 'fs ıx , where  x;j probability of sj for Cx is given by pd;x;j D Q  2   x;j  1 2 x;j C1 x is the average received signal-to-noise ratio (SNR) over channel Cx at sj . To enhance the accuracy of sensing results, sensor nodes collaboratively perform channel sensing. Specifically, sensor nodes in the same cluster send the individual sensing results to the cluster head to make a combined decision. The decision rules at the cluster head can include AND rule, OR rule, etc. When OR rule is adopted, PUs are considered to be present if at least one sensor claims the presence of PUs. Then, if we use a number of sensor nodes, e.g., a set of sensor nodes y, to cooperatively sense a channel, the cooperative detection probability Fdx and the cooperative false alarm probability Ffx for channel Cx are Fdx D 1 

Y

.1  pd;x;j /;

Nj 2y

Ffx D 1 

Y

.1  pf ;x;j /

(3.2)

Nj 2y

The cooperative misdetection probability Fmx is defined as the probability that the presence of the PU is not detected, i.e., Fmx D 1  Fdx . In order to guarantee the accuracy of spectrum sensing, channel sensing should satisfy a requirement that the probability of interfering with PUs should be below a predefined threshold FI . In other words, there is a constraint on y such that Y pxon  Fmx D pxon  .1  pd;x;j /  FI : (3.3) Nj 2y

3.1 System Model and Objective Statement

27

Given the signal transmission power Pj of sj , the noise power x2 over Cx , and the average channel gain h2j;i;x of the link between j and its destination node i over Cx , the transmission rate Rj;i;x from j to i can be given as [17]:   2 Pj Rj;i;x D Bx log 1 C hj;i;x 2 : (3.4) x We consider that data transmission over each licensed channel Cx is error-free with the available channel capacity in Eq. (3.4). During the intra-cluster data transmission, the transmission power of each sensor node is fixed to avoid co-channel interference among neighboring clusters [5]. The inter-cluster data transmission is also performed in TDMA, but CHs can adjust their transmission power for inter-cluster transmission when accessing a licensed channel. However, we assume that CHs do not adjust their power when they transmit data over C0 , to avoid potential interference to other applications operating on this license-free channel [18]. The determination of the transmission power over the default license-free channel can be referred to existing solutions [19, 20], which is out of the scope of this chapter.

3.1.3 Energy Consumption Model The energy consumption of sensor nodes mainly includes four parts: the energy consumption for spectrum sensing, spectrum switching, data transmission and reception. For each sensor node, we use es to denote the energy consumption for sensing a licensed channel, which is fixed and the same for different channels. Meanwhile, sensor nodes need to consume energy to configure the radio and switch to a new channel. Therefore, we use ew to denote the energy consumption that a sensor node consumes for channel switching. For sj , the data transmission energy consumption Ej;t is based on the classic energy model [21], i.e., Ej;t D .Pj CPj;c /tj;x , where tj;x is the data transmission time, Pj is the transmission power and Pj;c is the circuit power at sj . Following a similar model in [22], Pj;c can be calculated as 1 Pj;c D ˛j C .  1/  Pj , where ˛j is a transmission-power-independent component  that accounts for the power consumed by the circuit, and  is the power amplifier efficiency. Physically,  is determined by the drain efficiency of the RF power amplifier and the modulation scheme [21, 22]. Therefore, we have the energy consumption of data transmission at sj is Ej;t D

1 1  Pj  tj;x C ˛j  tj;x D .Pj C ˛c;j /  tj;x ;  

(3.5)

where ˛c;j D   ˛j is defined as the equivalent circuit power consumption for data transmission. The energy consumption for data receiving is related to the data that a sensor node receives [12]. If sj receives l bits data, the energy consumption is Ej;r D ec  l, where ec is the circuit power for data receiving.

28

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

A data period

...

...

Data transmission Data sensing

Intra-cluster Inter-cluster data transmission data transmission

Time

Sleep

Time flow in Intra- or Intercluster data transmission

Transmitting data over C0 Transmitting data over C0 Channel sensing Decision making

Channel accessing Decision making

Channel sensing

Channel accessing

...

Fig. 3.2 The time flow of CRSN

3.1.4 Objective Statement Figure 3.2 shows the time flow of the CRSN to illustrate the temporal relationship of different actions. As shown in the figure, a data period consists of three phases, i.e., data sensing, data transmission and sleeping. At the beginning of each data period, sj senses the monitored area and generates Aj sensed data to report to the sink. Once the sensed data is successfully transmitted to the next hop, it will turn into sleep mode for energy saving and wait for the next data period. Since data transmission is independent among different data periods, our objective is to efficiently transmit P A D sj 2N Aj data to the sink within a data transmission period, by determining the channel sensing and accessing decision according to the channel condition of C0 . As an indicator of the time-varying channel condition, the packet loss rate of C0 is measured/estimated at the beginning of each transmission period, by the RSSI (Received Signal Strength Indicator) and SNR (Signal-to-Noise Ratio) during the communications of each pair CM-CH and CH-Sink [23, 24], and assumed to be stable in a data transmission period but may vary over different periods [23]. According to the network model, the data transmission consists of two phases: intra-cluster data transmission and inter-cluster data transmission [25]. Therefore, we focus on reducing the energy consumption during the two phases, respectively. Figure 3.2 also shows the time flow of the two phases, which also describes the objectives of this work. Specifically, we aim to address the following two issues. 1. During the intra-cluster data transmission, each cluster Li should determine whether to sense and access a licensed channel according to the packet loss rate of C0 . When Li decides to sense and access a license channel, the channel sensing and accessing sequence should be determined for Li to minimize the energy consumption of intra-cluster data transmission in a probabilistic way.

3.1 System Model and Objective Statement

29

2. During the inter-cluster data transmission, the channel sensing and accessing decision should also be carefully determined for potential energy consumption reduction. Since CHs can adjust their transmission power when accessing a licensed channel, the transmission power control and dynamic channel accessing should be jointly considered to minimize the energy consumption of inter-cluster data transmission. To ease the presentation, the key notations are listed in Table 3.1.

Table 3.1 The key notations Notation N, L Hi , Ni C

C0 Bx es , ew Rj;i;x h2j;i;x , x2 j;i;0 y Pj Aj E1;0 .i/ pr Tx tj;x E1;x .i/, E1;x .i/ Ai E2;x , E2;x 0 E2;x Pi;x , ti;x

Definition Set of sensor nodes fs1 ; s2 ; : : : ; sn g, set of clusters fL1 ; L2 ; : : : ; Lm g Cluster head of Li , set of sensor nodes in Li Set of licensed channels fC1 ; C2 ; : : : ; Ck g The default license-free channel Bandwidth of channel Cx Energy consumption for channel sensing and switching Data transmission rate from sj to Hi over Cx Average channel gain between sj and Hi over Cx , average noise power over Cx Packet loss rate between sj and Hi over C0 A fixed number of sensor nodes chosen for cooperative channel sensing Transmission power of sj over C0 during intra-cluster data transmission sj ’s sensed data needed to be transmitted during the intra-cluster data transmission Energy consumption of Li by performing intra-cluster data transmission over C0 PU protection requirement The determined maximum channel available time of Cx under the required pr Allocated transmission time of sj over an accessed licensed channel Cx Energy consumption and expected energy consumption of Li by performing intra-cluster data transmission over Cx Aggregated data of Hi needed to be transmitted during inter-cluster data transmission Energy consumption and expected energy consumption of inter-cluster data transmission over Cx Equivalent energy consumption for optimizing E2;x Allocated transmission power and time of Hi over Cx during inter-cluster data transmission

30

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

3.2 Dynamic Channel Access for Intra-Cluster Data Transmission In this section, we propose a dynamic channel access solution for intra-cluster data transmission to improve the energy efficiency, according to the temporally fluctuated packet loss rate over C0 . Specifically, we adopt a four-step analysis to introduce the main ideas of the proposed solution: (1) We analyze the energy consumption E1;0 .i/ for intra-cluster data transmission over C0 in a cluster Li ; (2) We calculate the optimal energy consumption E1;x .i/ in a cluster Li , if Li accesses a licensed channel Cx for intra-cluster data transmission; (3) Since there are different idle probabilities for the licensed channels, we further calculate the expected energy consumption E1;x .i/ for the intra-cluster data transmission in Li by accessing Cx , taking the energy consumption of channel sensing and switching into consideration. Only if the packet loss rate over C0 increases to a value making E1;0 .i/ > E1;x .i/, Cx has the potential to improve the energy efficiency of intra-cluster data transmission; (4) When there are multiple licensed channels in C can potentially improve the energy efficiency, we propose a sequential channel sensing and   accessing strategy, where the licensed channel Cx with a larger E1;0 .i/  E1;x .i/ has a higher priority to be sensed and accessed by Li , to achieve the highest energy efficiency improvement. In the following, we will detail the main ideas and mathematical analysis of each step, respectively.

3.2.1 Energy Consumption Analysis Since each cluster aims to opportunistically access a licensed channel for intra-cluster data transmission to reduce the energy consumption of transmitting intra-cluster data over C0 , the original energy consumption should be calculated first if a cluster Li (Li 2 L) gathers the intra-cluster data over C0 . According to the system model, the packet loss rate over C0 can be measured for each communication link at the beginning of each data period. Given the measured packet loss rate of C0 , Proposition 1 analyzes the energy consumption of the clusters. Proposition 1 For any cluster Li (Li 2 L), if the data amount of a cluster member sj .sj 2 Ni / is Aj , and the packet loss rate between sj and the cluster head Hi over C0 is j;i;0 , the energy consumption for intra-cluster data transmission is E1;0 .i/ D

X Aj  ER1;j ; .1  j;i;0 / s 2N j

i

(3.6)

3.2 Dynamic Channel Access for Intra-Cluster Data Transmission

31

  Rj;i;0  ec C Pj C ˛c;j means the energy consumption rate of sj   Rj;i;0   for transmitting intra-cluster data, Rj;i;0 D B0 log 1 C h2j;i;0 Pj =02 and Pj is sj ’s transmission power. where ER1;j D

Proof For each sj 2 Ni , it generates Ai data to transmit during a data transmission period. Since the packet loss rate of C0 is j;i;0 , the expected number of transmission attempts for each packet is 1=.1  j;i;0 /. Therefore, the expected transmitted data is Aj =.1  j;i;0 /. If the transmission power of sj is Pj , the data transmission time is Aj . Therefore, for all the sensor nodes in Li , the energy consumption .1  j;i;0 /Rj;i;0 for data transmission is X  Aj  .Pj C ˛c;j / e1;t .i/ D   .1  j;i;0 /  Rj;i;0 s 2N j

D

i

X

2 4

sj 2Ni

3

Aj  .Pj C ˛c;j / 5 :  .1  j;i;0 /B0 log 1 C h2j;i;0 Pj =02

(3.7)

Additionally, the energy consumption for receiving the sensed data is e1;r .i/ D

X Aj  sj 2Ni

 1  ec : 1  j;i;0

(3.8)

Therefore, the total energy consumption of intra-cluster data transmission over C0 is E1;0 .i/ D e1;t .i/ C e1;r .i/, which can be transformed to Eq. (3.6). It completes the proof.

3.2.2 Optimal Transmission Time Allocation According to Eq. (3.6), the energy consumption for intra-cluster data transmission in Li grows sharply with the increasing packet loss rate of C0 . If we aim to access licensed channel Cx to reduce the intra-cluster energy consumption in Li , we should first address the problem: how to allocate the transmission time of CMs to minimize the energy consumption with the consideration of PU protection. In this section, we focus on determining the optimized energy consumption if Li accesses Cx for data transmission. When Li accesses to Cx , the channel available duration (CAD) of Cx , denoted by Tx , is limited to control the interference probability to PUs, due to the fact that PUs may return at any time point and cause an interference with a certain probability. We define pr as the PU protection requirement, which means the interference probability

32

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

to PU during Tx should be no larger than pr . According to the cognitive radio model, the PU traffic is an independent and identically distributed ON/OFF process, with vx as the mean idle time. Thus, if Cx is accessed for Tx , the interference probability of Cx is 1  evx Tx . Meanwhile, the probability that Cx is idle and detected as idle is pxoff  .1  Ffx /. Therefore, the interference probability during Tx is pxoff  .1  Ffx /  .1  evx Tx /, and the PU protection requirement is pxoff  .1  Ffx /  .1  evx Tx /  pr . Based on that, the maximum CAD of Cx is ! 1 pr Tx D  ln 1  x : vx poff  .1  Ffx /

(3.9)

If Tx is large enough to guarantee the complement of the intra-cluster data transmission in Li , all the data of CMs in Li can be transmitted over Cx . Otherwise, Tx should be carefully allocated to the CMs of Li to minimize the energy consumption, since CMs have different amounts of sensed data and different transmission rates, both of which can directly impact the energy consumption of intra-cluster data transmission. In the following, we mathematically formulate the transmission time allocation problem as an optimization problem, which will be solved to minimize the energy consumption of intra-cluster data transmission. For channel Cx and cluster Li , let tj;x be the allocated transmission time of sj (sj 2 Ni ) over Cx . Then, the energy consumption of sj for data transmission over Cx is 1 ej;x D .Pj C˛c;j /tj;x . The residual data of sj , if any, will be transmitted over C0 , with  the amount of Aj  Rj;i;x  tj;x . The associated energy consumption for transmitting the  Aj  Rj;i;x tj;x  ER1;j residual data over C0 is ej;0 D . Let E1;x .i/ be the total energy 1  j;i;0 consumption   data transmission in Li by accessing Cx . Then, we have P for intra-cluster E1;x .i/ D sj 2Ni ej;x C ej;0 . There are also some constraints for the transmission time allocation of Tx . For each CM sj 2 Ni , the successfully transmitted data of sj during the allocated time tj;x should be no larger than the generated data, which means Rj;i;x  tj;x  Ai ; 8sj 2 Ni :

(3.10)

Meanwhile, the allocated transmission time tj;x of sj should be no less than 0 and the total allocated transmission time of Li should be no larger than Tx . Thus, we have (P

sj 2Ni tj;x

 Tx ;

tj;x  0; 8sj 2 Ni :

(3.11)

We aim to determine the time allocation vector tx D ft1 ; : : : ; tjNi j g to minimize the energy consumption of intra-cluster data transmission, which can be formulated as the following optimization problem:

3.2 Dynamic Channel Access for Intra-Cluster Data Transmission

.TAP/ minimize E1;x .i/ D tx

33

X  ej;x C ej;0 sj 2Ni

s:t: (3.10) and (3.11): It can be seen that (TAP) is a classic linear programming problem. The wellknown Simplex method can be directly applied to solve this problem [26]. In   the following, we use tx D ft1 ; : : : ; tjN g and E1;x .i/ to denote the optimal time ij allocation and energy consumption for intra-cluster data transmission by accessing Cx , respectively.

3.2.3 Analysis of Channel Sensing and Switching Decision In this section, we focus on determining the condition when sensor nodes should sense and switch to a licensed channel for intra-cluster data transmission. By solving (TAP), we can obtain the optimal energy consumption for transmitting intra-cluster data over Cx . However, due to the uncertain availability of Cx and the energy consumption for channel sensing and switching, we can only obtain the expected energy consumption of intra-cluster data transmission by accessing Cx , if considering these two factors. According to the cognitive radio model, once Li decides to sense a licensed channel, a number of CMs y should be chosen to perform cooperative sensing to achieve better sensing performance. Here, jyj is a system parameter to meet the constraint of Eq. (3.3), and we assume jyj  minCi 2C jNi j. Recall that, reducing the energy consumption of intra-cluster data transmission is the primary objective for channel sensing and switching. To determine if the energy consumption can be improved by sensing and switching to a licensed channel, we first define the expected accessible channel that is expectedly profitable for a cluster to sense and access. Definition 1 For cluster Li , an expected accessible channel is a channel, by accessing which the expected energy consumption for intra-cluster data transmission can be reduced, taking account of the energy consumption for channel sensing and switching, as well as the idle detection probability of this channel by cooperative sensing. According to the definition, the following proposition determines the expected accessible channels for a specific cluster. Proposition 2 For channel Cx , given detection probability Pxd and false alarm probability Pxf , the expected energy consumption for intra-cluster data transmission in Li by accessing Cx is  E1;x .i/ D E1;0 .i/ C Yj;i;x Fsx tj;x C 2jNi jew Fsx C jyjes ;

(3.12)

34

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

and Cx is an expected accessible channel for Li , if we have  Yj;i;x  Fsx  tj;x C 2jNi j  ew  Fsx C jyj  es < 0;

where Fsx D pxoff .1Ffx / and Yj;i;x D

P sj 2Ni

(3.13)

 .P C ˛ /.1   /  ER R  j c;j j;i;0 1;j j;i;x . .1  j;i;0 /  

Proof For channel Cx with available time Tx , if it is used for intra-cluster data transmission in Li , the optimal time allocation solution can be determined as  tx D ft1 ; : : : ; tjN g by solving (TAP). Then, the optimal energy consumption for ij intra-cluster data transmission is     Aj  Rj;i;x tj;x ER1;j i X h .Pj C ˛c;j /tj;x  (3.14) C : .i/ D E1;x  1  j;i;0 s 2N j

i

If we consider the energy consumption for channel sensing and switching, the  total energy consumption for using Cx in intra-cluster data transmission is E1;x .i/ C jyj  es C 2jNi j  ew . Meanwhile, if Li decides to sense Cx , the probability that Cx is detected as available is Fsx D pxoff  .1  Ffx /, according to the cognitive radio model.1 It means that we have a probability Fsx to use Cx and a probability 1  Fsx to stay in channel C0 . Therefore, the expected energy consumption for sensing and switching to Cx for intra-cluster data transmission is    E1;x .i/ D Fsx  E1;x .i/ C jyj  es C 2jNi j  ew C .1  Fsx /  .E1;0 .i/ C jyj  es / : (3.15)  .i/ according to Eqs. (3.6) and (3.14), respectively, Substituting E1;0 .i/ and E1;x then Eq. (3.12) can be proved. If Cx is an expected accessible channel for Li , the expected energy consumption should be less than E1;0 .i/, i.e., E1;x .i/ < E1;0 .i/. Substituting E1;0 .i/ and E1;x .i/ with Eqs. (3.6) and (3.15), we can obtain Eq. (3.13). Based on Proposition 2, we have the following corollary to determine the condition in which the cluster Li should sense licensed channels for intra-cluster data transmission.

Corollary 1 If there exists such channel Cx 2 C that is an expected accessible channel of Li , Li should sense new channels for intra-cluster data transmission. Proof According to Definition 1 and Proposition 2, the expected energy consumption for intra-cluster data transmission can be reduced in Li by sensing and switching to the channel Cx , if Cx is an expected accessible channel of Li . Therefore, if there exists such channel Cx 2 C that can meet the constraint of Eq. (3.13), Li should sense this licensed channel for the potential energy efficiency improvement.

1

When Cx is detected as idle by cooperative sensing, there is also a probability that Cx is not available at this time, which is pxon  Fmx . However, this probability is limited below FI by Eq. (3.3), thus, we ignore it in the analysis of this work.

3.3 Joint Power Allocation and Channel Access for Inter-Cluster Data. . .

35

3.2.4 Dynamic Channel Sensing and Accessing Scheme In this section, we propose a sequential channel sensing and accessing scheme for the intra-cluster data transmission of each cluster. With Corollary 1, each cluster Li can decide whether it should sense a licensed channel for intra-cluster data transmission according to the packet loss rate of the default channel C0 . However, if there exist a set of expected accessible channels C 0 (C 0 2 C) for Li , the problem is which one is the most profitable to sense and access for intra-cluster data transmission. Proposition 2 indicates that the channel with the lowest expected energy consumption E1;x .i/ should be sensed first. However, E1;x .i/ is only an expected value and the availabilities of licensed channels are totally opportunistic, which means the expected accessible channels may be detected as unavailable through spectrum sensing. Therefore, we arrange the expected accessible channel set Cx 2 C 0 according to the increasing order E1;x .i/, and Li senses the channels of C 0 one by one according to the order until detecting a channel as idle. Then, Li switches to this channel for intra-cluster data transmission. Specifically, we discuss the dynamic channel sensing and accessing for intra-cluster data transmission in Li in the following situations. 1. If C 0 D ;, it means that there is no expected accessible channel for Li . The cluster does not sense any licensed channel and uses C0 for intra-cluster data transmission. 2. If C 0 ¤ ; and all the channels of C 0 are sensed as unavailable, Li transmits the intra-cluster data over C0 . 3. If C 0 ¤ ; and Cx (Cx 2 C 0 ) is sensed as idle by Li , Li switches to Cx and transmits the intra-cluster data over Cx . If the intra-cluster data is not completed after Tx , the channel sensing and accessing decision should be performed again. For each CM sj 2 Ni , we denote the residual data of sj as A0j . Then, we use A0j in Propositions 1 and 2 to determine the set of expected accessible channels C 0 , and repeat the channel sensing and accessing according the three situations until the intra-cluster data transmission is finished in Li . Based on the discussion above, Fig. 3.3 shows a flow chart to illustrate the procedures. Algorithm 1 presents the main idea of the dynamic channel sensing and accessing scheme for intra-cluster data transmission.

3.3 Joint Power Allocation and Channel Access for Inter-Cluster Data Transmission After intra-cluster data transmission, CHs aggregate the received data, and then send the aggregated data to the sink. Based on the analysis of intra-cluster data transmission, in this section, we focus on the channel accessing problem to improve the energy efficiency of inter-cluster data transmission. Similar to the analytical

36

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

Start

Calculate E1,0(i) and E1,x(i)

Determine expected accessible channel set C

Calculate

yes Sense C one by one until find an idle channel

no |C |=0 ?

Ck is idle

yes

Access Ck and transmit data according to TAP solution

All channels in C’ is busy

Transmit all residual data over C0

Update residual data Still have residual data?

no

End

Fig. 3.3 Procedures of the dynamic channel sensing and accessing scheme

way of intra-cluster data transmission, we perform a four-step analysis to introduce the dynamic channel access solution for inter-cluster data transmission to improve the energy efficiency. If we consider all the CHs and the sink as a cluster where CHs are CMs and the sink is the CH, the inter-cluster data transmission is similar to the intra-cluster data transmission. However, since there is no interference for TDMA-based inter-cluster transmission over licensed channels, CHs can adjust their transmission power to transmit their data to the sink when accessing to a licensed channel.

3.3.1 Analysis of Channel Sensing and Switching Decision Following the analytical path of intra-cluster data transmission, we first obtain the energy consumption of inter-cluster data transmission over C0 in the following proposition. According to our model, CHs do not adjust their power when they transmit over C0 , to avoid potential interference to other applications transmitting over this license-free channel. Therefore, we have the following proposition. Proposition 3 Given the data aggregation rate of Hi .Li 2 L/ as i , the packet loss rate i;s;0 between a cluster head Hi and the sink over C0 , the energy con-

3.3 Joint Power Allocation and Channel Access for Inter-Cluster Data. . .

37

Algorithm 1 Dynamic channel sensing and accessing for intra-cluster data transmission Input: For each sj and Cx , the sampling rate srj , the expected transmission rate Rj;i;x , packet loss rate j;i;x , available transmission duration Tx , and other parameters in cognitive model and energy consumption model. Output: Channel sensing and accessing sequence for intra-cluster data transmission. 1: for all Li 2 L do 2: Calculate the energy consumption of intra-cluster data transmission E1;0 .i/ over C0 ; 3: for all Cx 2 C do 4: Determine Ei;x .i/ and Ei;x .i/ by solving (TAP) and according to Proposition 2, respectively; 5: end for 6: Determine the expected accessible channel set C 0 according to Proposition 2, and reorder 0 C 0 as C according to increasing order of Ei;x .i/; 7: k D 1; 0 8: while k  jC j do 0 9: Sense the k-th channel Ck of C ; 10: if Ck is idle then 11: Go to step 18; 12: end if 13: k D k C 1; 14: end while 0 0 15: if jC j DD 0 or k > jC j then 16: Transmit the residual intra-cluster data over the default channel C0 ; 17: else 18: Transmit the intra-cluster data over the channel Ck , and allocate the transmission time tj;k to each sensor node Nj 2 Li ; 19: if The CAD of Ck is expired and the intra-cluster data transmission of Li is not completed then 20: Go to step 2; 21: end if 22: end if 23: end for

sumption for inter-cluster data transmission over C0 is E2;0 D

P Hi 2L

Ai  ER2;i , .1  i;s;0 /

  Ri;s;0  ec C Pi;0 C ˛c;j means the energy consumption rate of   Ri;s;0 P Hi for transmitting inter-cluster data over C0 , Ai D i , Ri;s;0 D sj 2Ni Aj    2 2 B0 log 1 C hi;s;0 Pi;0 =0 and Pi;0 is the transmission power of Hi . where ER2;i D

Proof Similar to the proof of Proposition 1. We then determine the minimized energy consumption of inter-cluster data transmission by accessing licensed channel Cx . Based on Eq. (3.9), we can calculate the CAD of Cx as Tx . Note that, besides Tx , the transmission power of CHs can also be adjusted for the inter-cluster data transmission. For each Hi , let Pi;x and ti;x denote the allocated transmission power and transmission time over Cx , respectively. 1 The energy consumption of data transmission over Cx is ei;x D .Pi;x C ˛c;i /  ti;x , 

38

3 Dynamic and Energy-Efficient Channel Access in Clustered CRSNs

and the energy consumption of transmitting the residual data over C0 , if any, ! i h h2i;s;x Pi;x 1 is e0;x D Ai  Bx log 1 C . To minimize the ti;x  ER2;i  2 x 1  i;s;0 energy consumption, we can jointly determine the transmission power vector Px D fP1;x ; : : : ; Pm;x g and transmission time vector tx D ft1;x ; : : : ; tm;x g of the CHs, which can be formulated as the following optimization problem: .PTAP/ minimize E2;x D Px ;tx

X

.ei;x C ei;0 /

Hi 2L

! 8 h2i;s;x Pi;s ˆ ˆ ˆ Bx  log 1 C  ti;x  Ai ; 8Hi 2 L; ˆ ˆ x2 ˆ

Tsw , Proposition 7 provides an approximation solution. Proposition 7 To guarantee that the probability of a honest node being identified as compromised by FastDtec is no larger than ', the approximate optimal detection threshold  .T/, when Tsw < T  Tmax , is p 8 H z1' TPH ˆ e .1  Pe /  ˆ ; < .T/  V0  PH e 

2 H .V0  Vmin /  Pe ˆ ˆ :Tmax D : 1  PH e

(4.22)

4.3 FastDtec: Trust Evaluation for Fast Compromised Node Identification

67

where z1' is 1  ' quantile of standard normal distribution, Vmin is the minimum trust value. Proof Let X be the number of wrong sensing reports with X  B.T; PH e /. When T  20, the binomial distribution B.T; PH / can be approximated as a normal e H H ; TP .1  P //. On the basis of the normal distribution [31, 32], i.e., X  N.TPH e e e X  TPH e , the 1  ' quantile of B.T; PH approximation to the quantity p e /, H/ TPH .1  P e e denoted by L.T; PH e ; 1  '/, is given as H L.T; PH e ; 1  '/  TPe C z1'

q H TPH e .1  Pe /;

(4.23)

where z1' is the 1  ' quantile of the standard normal distribution. To guarantee  P.X  f . ; T//  ', we have f . ; T/  L.T; PH e ; 1  '/. Thus, f . .T/; T/ D H  L.T; Pe ; 1  '/. By substituting f . .T/; T/ according to Eq. (4.20), we have the optimal  .T/ for given T is 

.T/  V0 

z1'

p

H TPH e .1  Pe / : PH e

Noting that, since ' is usually a small value, such as 5%, z1' is a positive value. Therefore,  .T/ is a decreasing function with respect to T, which indicates that the optimal threshold  .T/ decrease with the increasing T. Since Vmin is the minimum trust value, we can calculate the maximum T, denoted can makes

by Tmax , which 2 H  V /  P .V 0 min e

 .T/ below Vmin . Let  .T/  Vmin , we have Tmax D . 1  PH e According to Proposition 7, we can adaptively determine the optimal detection threshold  .T/ for FastDtec when Tsw < T  Tmax . We define Tmax as a trust evaluation cycle of FastDtec. When node i is selected for spectrum sensing Tmax C 1 times, we consider it is the first evaluation period of a new evaluation cycle, and reset its trust value to V0 . It indicates that a compromised node would be missed by FastDtec if it can keep its trust value large than Vmin after Tmax times of spectrum sensing. Lemma 5 analyzes the probability that a compromised node is not identified by FastDtec within a trust evaluation cycle. Lemma 5 Given the required false identification probability ', the missed identification probability of FastDtec is MD.'/ D minfMD.'; Tmin /; : : : ; MD.'; Tmax /g, T P where MD.'; T/ is given as MD.'; T/ D Q.T; k; Pe /; Tmin and Tmax are kDbf .  .T/;T/c

in Eqs. (4.21) and (4.22), and Pe D PIe or PCe according to Eqs. (4.13) and (4.14) under I-SSDF or C-SSDF attacks, respectively. Proof We first focus on calculating the missed identification probability of FastDtec for given T. Let X be the number of wrong sensing reports with X  B.T; Pe /. If a

68

4 Secure and Energy-Efficient Collaborative Spectrum Sensing

1  PH e / PH e  

.T/. Thus, we have that X  f . .T/; T/ and the corresponding probability is T P MD.'; T/ D P.X  f .  .T/; T// D Q.T; k; Pe /.

compromised node is missed by FastDtec, it means V0 C.T X/X.1

kDbf .  .T/;T/c

If a compromised node is missed by FastDtec, it means that the node is not identified by FastDtec at any T 2 fTmin ; : : : ; Tmax g. Thus, we have that the missed identification probability of FastDtec is MD.'/ D minfMD.Tmin /; : : : ; MD.T/; : : : ; MD.Tmax /g.

4.3.3 The Proposed FastDtec Scheme Based on the preceding analysis, FastDtec can utilize asymmetric evaluation to stabilize the trust values of honest nodes and linearly reduce the trust values of compromised nodes. Moreover, an adaptive detection threshold is dynamically derived for FastDtec to accelerate the identification of compromised nodes while keeping the false identification probability below a specific requirement. We summarize the main ideas of FastDtec in Algorithm 4. Algorithm 4 The procedures of FastDtec Input: The parameters related to FastDtec, such as V0 , Vmin , Vmax and '. 1: Ti 0 for each node i 2 S , CM ;; 2: for each spectrum sensing node i 2 N do 3: Ti Ti C 1; 4: Evaluate a trust score and update its trust value table according to Eqs. (4.15) and (4.16), respectively; 5: if Tmin  Ti  Tsw then 6: Determine the optimal detection threshold  .T/ according to the numerical procedure; 7: else if Tsw < Ti  Tmax then 8: Determine the optimal detection threshold  .T/ according to Proposition 7; 9: end if 10: if Vi   .T/ then 11: CM CM C fig; 12: end if 13: if Ti  Tmax && i … CM then 14: Ti D 0, Vi D 0; 15: end if 16: end for

4.4 Secure and Energy-Efficient Collaborative Spectrum Sensing

69

4.4 Secure and Energy-Efficient Collaborative Spectrum Sensing In this section, we propose a secure and energy-efficient collaborative spectrum sensing scheme for CRSNs, which can utilize FastDtec to accurately identify the compromised nodes and adaptively reduce the number of spectrum sensing nodes for further enhancing energy efficiency.

4.4.1 The Proposed Scheme Based on our analysis, FastDtec can fast and accurately identify the compromised nodes launching SSDF attacks in CRSNs. If we isolate the identified compromised nodes from collaborative spectrum sensing, we can adaptively reduce the number of spectrum sensing nodes, i.e., N, to further improve the energy efficiency of collaborative spectrum sensing in CRSNs. For example, after M sensor nodes are identified as compromised nodes, the probability of selecting a compromised node into spectrum sensing changes to ˛S  M ˛0 D . Obviously, we have ˛ 0 < ˛. According to Proposition 6, the reduced SM ˛ 0 means the decreased false alarm probabilities PIF , PCF and missed detection I probabilities PIM , PCM under fixed L and N, which also indicates that reduced Nmin C and Nmin could be derived to guarantee the required ıF and ıM . Therefore, once there is a sensor node identified as compromised by FastDtec, we can redetermine the optimal number of spectrum sensing nodes. The main idea of the proposed secure and energy-efficient collaborative spectrum sensing scheme is illustrated in Fig. 4.3, and the detailed procedures are summarized in Algorithm 5.

4.4.2 Further Discussion In the proposed scheme, all the spectrum sensing nodes are randomly selected from the sensor nodes without being blacklisted. Meanwhile, the reports from the spectrum sensing nodes have the same weights for the global decision making. A number of existing works have done in-depth investigation to leverage the evaluated trust values in spectrum sensing node selection and global decision making. It can also be applied into the proposed scheme to improve the accuracy of global decision making. For example, we can allocate higher priority to the sensor nodes with higher trust values when selecting the spectrum sensing nodes. However, the energy consumption balancing issue should be further studied to keep the sensor nodes with high trust values from exhausting their energy quickly. Moreover, when the sink fuses local decisions to make a global decision, the reports from

70

4 Secure and Energy-Efficient Collaborative Spectrum Sensing

Isolate

CM

S

α’=(α|S|-|CM|)/(|S|-|CM|) Determine N with α’ based on Alg. 1

Randomly select N nodes

Spectrum sensing node

Identified compromised nodes for updating α

Local sensing reports and α’ Trust evaluation Determine ξ*(t)

Update

Trust table

Identify compromised nodes

FastDtec

Fig. 4.3 Illustration of the proposed secure and energy-efficient collaborative spectrum sensing scheme

Algorithm 5 The proposed secure and energy-efficient collaborative spectrum sensing scheme Input: The percentage of compromised nodes ˛, the required security levels ıF and ıM , the set of sensor nodes S , V0 , Vmin , Vmax and other parameters. 1: SN S , ˛0 ˛, CM ;; 2: Vi V0 for each i 2 S ; 3: for each time period t from 1 to C1 do 4: At the beginning of time period t : C I 5: Determine Nmin , LI , Nmin and LC with ˛ 0 , ıF and ıM by addressing Eqs. (4.11) and (4.12); C I 6: Under I-SSDF and C-SSDF attack scenarios, randomly select Nmin and Nmin sensor nodes I C from SN with L and L for spectrum sensing, respectively; 7: At the end of time period t : 8: Update the trust values of spectrum sensing nodes according to Algorithm 4; 9: if There exists node i 2 SN with Vi   then 10: SN SN  fig, CM CM C fig; ˛S  jCMj 0 11: ˛ ; jSN j 12: end if 13: end for

the sensor nodes with higher trust values can be allocated with higher weights for decision making. The related techniques and analysis can be referred to [8, 17, 19]. In [19], it is proved that the weighted decision making based on evaluated trust values has positive impacts on reducing the number of sensing reports. However,

4.5 Performance Evaluation

71

when weighted decision making is adopted, the designed SSDF countermeasure should be capable of resisting promotion attacks, where compromised nodes can collaboratively promote the trust values of several compromised nodes to mislead the decision making. In this work, we provide a probabilistic analysis for securing collaborative spectrum sensing, from the energy efficiency perspective. The proposed secure and energy-efficient collaborative spectrum sensing scheme can guarantee a desired security requirement and maximize the energy efficiency simultaneously. To the best of our knowledge, this is the first work to investigate the energy efficiency optimization for collaborative spectrum sensing under a desired security requirement in a probabilistic way.

4.5 Performance Evaluation In this section, we validate our theoretical analysis and evaluate the performance of our proposed schemes by OMNET++ [33, 34]. We setup a CRSN with jSj D 200 sensor nodes and a sink node. The network process is divided into a sequence of time periods. At the beginning of each time period, the sink randomly chooses a number of sensor nodes to sense a licensed channel. The idle and busy probabilities of the licensed channel are P0 D 90% and P1 D 10%, respectively. The parameters for individual spectrum sensing are set as Pcf D 20% and Pcm D 20%. The percentage of compromised nodes in the CRSN is ˛ D 20%, if it is not specified in the simulation figures. The desired missed detection probability and false alarm probability of collaborative spectrum sensing are set as ıM D 1% and ıF D 5%, respectively. For the settings of FastDtec, we have V0 D 200, Vmin D 0 and Vmax D 255. The requirement of the false identification probability of FastDtec is ' D 2:5%, and Tsw D 20.

4.5.1 Simulation Settings In this subsection, we first evaluate the impacts of I-SSDF and C-SSDF attacks on the accuracy of collaborative spectrum sensing. Then, we show the tradeoff between security and energy efficiency determined by addressing Eqs. (4.11) and (4.12). Figure 4.4 shows the accuracy of global decision by collaborative spectrum sensing under different percentages of SSDF attackers (i.e., compromised nodes). The accuracy of global decision is defined as P.D D 1jH1 / C P.D D 0jH0 /. 20 sensor nodes are involved in the collaborative spectrum sensing and the decision rule is “majority” rule, which mean N D 20 and L D 11. The individual sensing refers to that only one sensor node is selected for sensing the licensed channel. It can be seen from the figure that the accuracy of global decision is significantly reduced with the increasing percentage of compromised nodes. The collaborative spectrum

72

4 Secure and Energy-Efficient Collaborative Spectrum Sensing

Accuracy of Global Decision

100% Pa,0=Pa,1=Pa,c=100%

80% 60% 40%

50% - Random Guess I-SSDF Attack C-SSDF Attack Individual Sensing

20% 0% 0%

10% 20% 30% 40% 50% 60% 70% Percentage of Compromised Nodes

Number of Spectrum Sensing Nodes

Fig. 4.4 The impacts of I-SSDF and C-SSDF attacks on the accuracy of collaborative spectrum sensing

45 Optimal N under I-SSDF Attacks L under I-SSDF Attacks Optimal N under C-SSDF Attacks L under C-SSDF Attacks

40 35 30 25

Pa,0=Pa,1=Pa,c=80%

20 15 10 5 0 0%

5% 10% 15% 20% 25% Percentage of Compormised Nodes

Fig. 4.5 Required number of spectrum sensing nodes to guarantee a desired security level

sensing under C-SSDF attacks has a lower accuracy than that under I-SSDF attacks. Moreover, when the percentage of compromised nodes exceeds 39%, the global decision accuracy of collaborative spectrum sensing under C-SSDF attacks is below 50%. It means that collaborative spectrum sensing is no better than random guess for a network with more than 39% compromised nodes to launch C-SSDF attacks. A similar conclusion can be found in [9, 17]. In addition, under C-SSDF attacks, individual sensing is more accurate than collaborative spectrum sensing when the percentage of compromised nodes is larger than 37%. Figure 4.5 compares the required number of spectrum sensing nodes to guarantee desired security levels (i.e., ıF and ıM ) under different percentages of compromised nodes. It is shown that

4.5 Performance Evaluation

73

Trust Value

225

I-SSDF Attack with Pa,0=Pa,1=20%

210

C-SSDF Attack with Pa,c=20%

195

C-SSDF Attack with Pa,c=80%

180

I-SSDF Attack with Pa,0=Pa,1=80%

V0=200 identified at T=27 identified at T=43

165

identified at T=20

150

identified at T=19

135 0

5

10 15 20 25 30 35 40 Times of Participating Spectrum Sensing

45

Fig. 4.6 Expected trust values of compromised nodes under different attacking probabilities

the required number of spectrum sensing nodes increases sharply with increasing compromised nodes, especially for resisting C-SSDF attacks.

4.5.2 Attack Impacts Evaluation In this subsection, we evaluate the performance of the proposed FastDtec scheme, in terms of the speed and accuracy of compromised node identification. Figure 4.6 compares the expected trust values of compromised nodes with different attacking probabilities. It can be seen that the expected trust values of compromised nodes with a higher attacking probability drop more quickly as the times of participating spectrum sensing increases. Consequently, it leads to that the compromised nodes with higher attacking probability are identified by FastDtec faster than those with lower attacking probability. Moreover, with the same attacking probability, the compromised node launching C-SSDF attacks has a lower expected trust value and is identified by FastDtec faster than the one launching I-SSDF attacks. Figure 4.7 compares the determined threshold  .T/ and the optimal threshold under different T. It shows that the approximate optimal threshold determined by Proposition 7 is very close to the optimal threshold. Figure 4.8 shows the determined optimal detection thresholds of FastDtec under different attacking probabilities and T. We use X to denote the wrong sensing reports of a sensor node. These three figures show the probability mass functions (PMF, i.e., P.X D k/ where 0  k  T), of different sensor nodes, under different T and attacking probabilities. It can be seen from the figures that the overlap of the PMFs of a honest node and a compromised node becomes larger when the attacking probability is low. It indicates that the missed identification probability of FastDtec MD.'; T/ would be larger under a

74

4 Secure and Energy-Efficient Collaborative Spectrum Sensing

190 Approximate optimal threshold

Detection Threshold

185 180 175 170 165

Determined Threshold Optimal Threshold

160 155

10

15 20 25 30 35 40 45 50 Times of Participating Spectrum Sensing

Fig. 4.7 Comparison of determined detection threshold and optimal detection threshold 30%

30% PMF of Honest Node PMF of I-SSDF Attacker PMF of C-SSDF Attacker

Pa,0=Pa,1=Pa,c=80%

25%

Upper tolerance limit = 8 ξ*(20)=180

20% Probability

Probability

20% 15% 10% 5%

PMF of Honest Node PMF of I-SSDF Attacker PMF of C-SSDF Attacker

25%

Upper tolerance limit = 8 ξ*(20)=180

15% Pa,0=Pa,1=Pa,c=40%

10% MDP(20)

5%

MDP(20)

0%

0% 0

2 4 6 8 10 12 14 16 Number of Wrong Sensing Reports

18

20

0

5 10 15 Number of Wrong Sensing Reports

(a)

(b)

20%

PMF of Honest Node PMF of I-SSDF Attacker PMF of C-SSDF Attacker

15% Probability

20

Upper tolerance limit=13 ξ*(30)=175

10% Pa,0=Pa,1=Pa,c=80%

5% MDP(40)

0% 0

5

10 15 20 25 30 35 Number of Wrong Sensing Reports

40

(c) Fig. 4.8 Determined detection thresholds of FastDtec under different attacking probabilities and T. (a) T D 20, Pa;0 D Pa;1 D Pa;c D 80%. (b) T D 20, Pa;0 D Pa;1 D Pa;c D 40%. (c) T D 40, Pa;0 D Pa;1 D Pa;c D 40%

Percentage of Compromised Nodes

4.5 Performance Evaluation

75

20% 15% 10% Pa,0=Pa,1=Pa,c=80%

5%

I-SSDF Attack C-SSDF Attack

0% 0

20

40

60 80 100 Time Period (T)

120

140

Fig. 4.9 The speed of compromised node identification under I-SSDF and C-SSDF attacks

lower attacking probability when T is fixed. Meanwhile, as the number of times of participating spectrum sensing increases, the missed identification probability of FastDtec reduces significantly. Figure 4.9 shows the percentage of the compromised nodes that are not identified by FastDtec. It can be seen that the compromised nodes under C-SSDF attack scenarios are identified faster than those under I-SSDF attack scenarios. Furthermore, the percentage of the compromised nodes without being identified quickly plummets to 0 after a compromised node is identified by the proposed scheme. It also demonstrates that FastDtec can fast identify the compromised nodes in a small number of time periods. To show the superiority of FastDtec in compromised node identification, we compare FastDtec with an existing scheme [35], named CatchIt, in terms of the accuracy and speed of compromised node identification. Figure 4.10 shows the comparison of the identification accuracy of FastDtec and CatchIt under different attacking probabilities. The identification accuracy is evaluated by the sum of two probabilities. One is false identification probability (FIP), which means the probability that the trust system falsely identifies a honest node as a compromised node. And the other is missed identification probability (MIP), which means the probability that a compromised node is missed by the trust scheme within a certain number of periods. It can be seen from the figure that the identification accuracy increases with the attacking probability of compromised nodes and both of FastDtec and CatchIt can achieve a high identification accuracy under different attacking probabilities. However, FastDtec can still outperform CatchIt, i.e., having lower FIP+MIP, especially when the attacking probability is low. Figure 4.11 compares the identification speed of FastDtec and CatchIt under different attacking probabilities. We evaluate the identification speed by the number of evaluation periods that is required to identify a compromised node. It can be seen that, under both of FastDtec and CatchIt, C-SSDF attackers can be identified faster than I-SSDF attackers at the same attacking probability. Meanwhile, FastDtec has a faster identification

76

4 Secure and Energy-Efficient Collaborative Spectrum Sensing

12% C-SSDF Attacker under FastDtec C-SSDF Attacker under CatchIt I-SSDF Attacker under FastDtec I-SSDF Attacker under CatchIt

FIP+MIP

10% 8% 6% 4% 2%

10%

20% 30% Attacking Probability

40%

Identification Speed (Periods)

Fig. 4.10 Compromised node identification accuracy comparison under different attacking probabilities C-SSDF Attacker under FastDtec C-SSDF Attacker under CatchIt I-SSDF Attacker under FastDtec I-SSDF Attacker under CatchIt

60 50 40 30 20 10 0

20%

40% 60% Attacking Probability

80%

Fig. 4.11 Compromised node identification speed comparison under different attacking probabilities

speed than CatchIt. The comparison results of the two figures also demonstrate that by guaranteeing the identification accuracy and optimizing the identification speed, FastDtec can significantly improve the performance of compromised node identification for collaborative spectrum sensing.

4.6 Summary

16 Number of Spectrum Sensing Node

Fig. 4.12 Required number of spectrum sensing nodes under different time periods

77

C-SSDF Attack I-SSDF Attack

14 12

Pa,0=Pa,1=Pa,c=80%

10 8 6 4 2 0 40

50

60

70

80 90 100 110 120 130 Time Period

4.5.3 Performance of the Proposed Scheme In this subsection, we evaluate the performance of the proposed secure and energyefficient collaborative spectrum sensing scheme. Since the proposed collaborative spectrum sensing scheme can guarantee the desired accuracy requirements by Eqs. (4.11) and (4.12), we mainly focus on evaluating the energy efficiency enhancement. As the energy consumption of collaborative spectrum sensing is high and increases linearly with the number of spectrum sensing nodes, we use the required number of sensor nodes that can guarantee the accuracy requirements to evaluate the energy efficiency. Figure 4.12 shows the required numbers of spectrum sensing nodes during different time periods. It can be seen that the number of required spectrum sensing nodes is significantly reduced with the increasing time period, under both types of SSDF attacks. And the number of spectrum sensing nodes decreases earlier and faster under C-SSDF attacks, as the compromised nodes with C-SSDF attacks are identified faster by the proposed scheme.

4.6 Summary In this chapter, we have investigated the secure collaborative spectrum sensing for CRSNs, from the energy efficiency perspective. We theoretically analyze the impacts of independent and collaborative SSDF attacks on the accuracy of collaborative spectrum sensing. Our analysis and simulations show that the number of spectrum sensing nodes and associated global decision rule have significant impacts on the accuracy of collaborative sensing results. To achieve the tradeoff between security and energy efficiency, we determine the minimum number of spectrum sensing nodes to guarantee desired security requirements. Moreover, we have developed a trust evaluation scheme, named FastDtec, to evaluate the

78

4 Secure and Energy-Efficient Collaborative Spectrum Sensing

spectrum sensing behaviors and identify the compromised nodes. By determining an adaptive and optimal detection threshold, FastDtec can fast and accurately identify compromised nodes. In addition, taking advantage of FastDtec to isolate the identified compromised nodes from spectrum sensing, we have proposed a secure and energy-efficient collaborative spectrum sensing scheme to further enhance the energy efficiency of collaborative spectrum sensing. Extensive simulation results demonstrate that the proposed collaborative spectrum sensing scheme can effectively resist SSDF attacks, and fast and accurately identify compromised nodes, as well as improving energy efficiency.

References 1. R. Sharma, D. Rawat, Advances on security threats and countermeasures for cognitive radio networks: a survey. IEEE Commun. Surv. Tutorials 17(2), 1023–1043 (2015) 2. N. Zhang, H. Zhou, K. Zheng, N. Cheng, J.W. Mark, X. Shen, Cooperative heterogeneous framework for spectrum harvesting in cognitive cellular network. IEEE Commun. Mag. 53(5), 60–67 (2015) 3. Y. Cai, Y. Mo, K. Ota, C. Luo, M. Dong, L. Yang, Optimal data fusion of collaborative spectrum sensing under attack in cognitive radio networks. IEEE Netw. 28(1), 17–23 (2014) 4. Z. Gao, H. Zhu, S. Li, S. Du, X. Li, Security and privacy of collaborative spectrum sensing in cognitive radio networks. IEEE Wirel. Commun. 19(6), 106–112 (2012) 5. N. Zhang, N. Cheng, N. Lu, H. Zhou, J.W. Mark, X. Shen, Risk-aware cooperative spectrum access for multi-channel cognitive radio networks. IEEE J. Sel. Areas Commun. 32(3), 516–527 (2014) 6. G. Baldini, T. Sturman, A.R. Biswas, R. Leschhorn, G. Gódor, M. Street, Security aspects in software defined radio and cognitive radio networks: a survey and a way ahead. IEEE Commun. Surv. Tutorials 14(2), 355–379 (2012) 7. T. Qin, H. Yu, C. Leung, Z. Shen, C. Miao, Towards a trust aware cognitive radio architecture. ACM SIGMOBILE Mob. Comput. Commun. Rev. 13(2), 86–95 (2009) ˇ 8. K. Zeng, P. Pawełczak, D. Cabri´ c, Reputation-based cooperative spectrum sensing with trusted nodes assistance. IEEE Commun. Lett. 14(3), 226–228 (2010) 9. A. S. Rawat, P. Anand, H. Chen, P.K. Varshney, Collaborative spectrum sensing in the presence of Byzantine attacks in cognitive radio networks. IEEE Trans. Signal Process. 59(2), 774–786 (2011) 10. R. Chen, J.-M. Park, K. Bian, Robust distributed spectrum sensing in cognitive radio networks, in Proceedings of IEEE INFOCOM (2008) 11. X. He, H. Dai, P. Ning, HMM-based malicious user detection for robust collaborative spectrum sensing. IEEE J. Sel. Areas Commun. 31(11), 2196–2208 (2013) 12. J. Wang, J. Yao, Q. Wu, Stealthy-attacker detection with a multidimensional feature vector for collaborative spectrum sensing. IEEE Trans. Veh. Technol. 62(8), 3996–4009 (2013) 13. X. He, H. Dai, P. Ning, A Byzantine attack defender in cognitive radio networks: the conditional frequency check. IEEE Trans. Wirel. Commun. 12(5), 2512–2523 (2013) 14. F. Penna, Y. Sun, L. Dolecek, D. Cabric, Detecting and counteracting statistical attacks in cooperative spectrum sensing. IEEE Trans. Signal Process. 60(4), 1806–1822 (2012) 15. Z. Qin, Q. Li, G.-C. Hsieh, Defending against cooperative attacks in cooperative spectrum sensing. IEEE Trans. Wirel. Commun. 12(6), 2680–2687 (2013) 16. M. Ghaznavi, A. Jamshidi, A reliable spectrum sensing method in the presence of malicious sensors in distributed cognitive radio network. IEEE Sensors J. 15(3), 1810–1816 (2015)

References

79

17. C. Hyder, B. Grebur, L. Xiao, M. Ellison, ARC: adaptive reputation based clustering against spectrum sensing data falsification attacks. IEEE Trans. Mob. Comput. 13(8), 1707–1719 (2014) 18. S. Althunibat, V. Sucasas, H. Marques, J. Rodriguez, R. Tafazolli, F. Granelli, On the trade-off between security and energy efficiency in cooperative spectrum sensing for cognitive radio. IEEE Commun. Lett. 17(8), 1564–1567 (2013) 19. S.A. Mousavifar, C. Leung, Energy efficient collaborative spectrum sensing based on trust management in cognitive radio networks. IEEE Trans. Wirel. Commun. 14(4), 1927–1939 (2015) 20. J. Ren, Y. Zhang, N. Zhang, D. Zhang, X. Shen, Dynamic channel access to improve energy efficiency in cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 15(5), 3143–3156 (2016) 21. S. Althunibat, R. Palacios, F. Granelli, Energy-efficient spectrum sensing in cognitive radio networks by coordinated reduction of the sensing users, in Proceedings of IEEE ICC (2012), pp. 1399–1404 22. Y.-C. Liang, Y. Zeng, E. Peh, A.T. Hoang, Sensing-throughput tradeoff for cognitive radio networks. IEEE Trans. Wirel. Commun. 7(4), 1326–1337 (2008) 23. M. Timmers, S. Pollin, A. Dejonghe, L. Van der Perre, F. Catthoor, A distributed multichannel MAC protocol for multihop cognitive radio networks. IEEE Trans. Veh. Technol. 59(1), 446–459 (2010) 24. J. Ren, Y. Zhang, K. Zhang, X. Shen, Adaptive and channel-aware detection of selective forwarding attacks in wireless sensor networks. IEEE Trans. Wirel. Commun. 15(5), 3718–3731 (2016) 25. J. Ren, Y. Zhang, K. Zhang, X. Shen, Exploiting mobile crowdsourcing for pervasive cloud services: challenges and solutions. IEEE Commun. Mag. 53(3), 98–105 (2015) 26. J. Ren, Y. Zhang, Q. Ye, K. Yang, K. Zhang, X.S. Shen, Exploiting secure and energyefficient collaborative spectrum sensing for cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 15(10), 6813–6827 (2016) 27. D. Zhang, Z. Chen, J. Ren, N. Zhang, M.K. Awad, H. Zhou, X.S. Shen, Energy-harvestingaided spectrum sensing and data transmission in heterogeneous cognitive radio sensor network. IEEE Trans. Veh. Technol. 66(1), 831–843 (2017) 28. W. Zhang, R.K. Mallik, K. Letaief, Optimization of cooperative spectrum sensing with energy detection in cognitive radio networks. IEEE Trans. Wirel. Commun. 8(12), 5761–5766 (2009) 29. S. Althunibat, R. Palacios, F. Granelli, Energy-efficient spectrum sensing in cognitive radio networks by coordinated reduction of the sensing users, in Proceedings of IEEE ICC (2012), pp. 1399–1404 30. S. Althunibat, Towards energy efficient cooperative spectrum sensing in cognitive radio networks. Ph.D. dissertation, University of Trento (2014) 31. T.T. Cai, H. Wang, Tolerance intervals for discrete distributions in exponential families. Stat. Sinica 19(3), 905 (2009) 32. K. Krishnamoorthy, Y. Xia, F. Xie, A simple approximate procedure for constructing binomial and Poisson tolerance intervals. Commun. Stat. Theory Methods 40(12), 2243–2258 (2011) 33. J. Ren, Y. Zhang, K. Zhang, A. Liu, J. Chen, X. Shen, Lifetime and energy hole evolution analysis in data-gathering wireless sensor networks. IEEE Trans. Ind. Inf. 12(2), 788–800 (2016) 34. Q. Ye, W. Zhuang, L. Li, P. Vigneron, Traffic load adaptive medium access control for fully-connected mobile ad hoc networks. IEEE Trans. Veh. Technol. (2016, to appear). doi: 10.1109/TVT.2016.2516910 35. W. Wang, H. Li, Y. Sun, Z. Han, CatchIt: detect malicious nodes in collaborative spectrum sensing, in Proceedings of IEEE GLOBECOM (2009), pp. 1–6

Chapter 5

Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

Sensor nodes in CRSNs can opportunistically access licensed channels for data transmission and reception by adjusting their radio configuration, when the licensed channels are sensed as available. As a result, the dynamic channel access becomes critical to guarantee the quality-of-service (QoS) for CRSNs [1]. In this regard, a number of existing works provide comprehensive investigation to reduce transmission delay [2–5] and improve network capacity [6, 7], laying a solid foundation for studying the dynamic channel access in CRSNs. Meanwhile, for most of datagathering CRSN applications, network utility is an important indicator to evaluate data collection efficiency. Since network utility depends on the sampling rates of sensor nodes and network capacity, it motivates us to jointly study the sampling rate control and dynamic channel access to optimize the network utility in CRSNs. To this end, we focus on the network utility maximization problem in EH-CRSNs. Different from the existing works that maximize the network utility for the WSNs with a static network capacity, the dynamic channel access leads to a dynamic network capacity in CRSNs. Therefore, network utility maximization in EH-CRSNs faces a great challenge of scheduling channel access for sensor nodes, in addition to the inherent challenges such as sampling rate controlling and stochastic renewable energy constraints. In this chapter, we investigate the network utility maximization problem in EH-CRSNs by jointly considering the sampling rate control and channel access [8]. The network operation is divided into different energy harvesting cycles, each of which is time slotted for dynamic channel access. Since the energy consumption for sensing the availability of licensed channels is considerable for sensor nodes, a spectrum market is introduced to take charge of channel sensing at each time slot and use the historical sensing results to predict the channel availability information for the CRSN. Sensor nodes are scheduled to access the available channels under the channel interference constraint at each time slot. The network utility is defined as increasing with the amount of sensory data collected by the CRSN and decreasing with the capacity of the accessed channels. Consequently, the network utility would © Springer International Publishing AG 2018 J. Ren et al., Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks, DOI 10.1007/978-3-319-60318-6_5

81

82

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

be greatly impacted by sampling rate control and channel access schedule, under the harvested energy, channel capacity and interference constraints. Therefore, we jointly optimize the channel access and sampling rates and propose an efficient algorithm to maximize the network utility. Specifically, the main content and contributions of this chapter can be summarized as follows. • With the consideration of fluctuated energy harvesting rates and energy consumption in channel switching, we formulate the network utility maximization problem as a mixed-integer nonlinear optimization problem (MINLP), to determine the optimal channel access and sampling rate for each sensor node at different time slots. • To address the MINLP, we decouple the primal problem into two independent subproblems by dual decomposition and present two efficient solutions for the subproblems. Based on the subproblem solutions, a subgradient method based algorithm is proposed to solve the network utility maximization problem. • To mitigate the impact of prediction error, we further propose a Joint channel Access and Sampling rate Control scheme, named JASC, to maximize the network utility, which can determine the sampling rates and accessed channels of sensor nodes at each time slot, adapting to the real-time channel sensing results and energy harvesting rates. The remainder of this chapter is organized as follows. The system model and problem formulation are introduced in Sect. 5.1. In Sect. 5.2, we decouple the network utility maximization problem into two subproblems by dual decomposition. Section 5.3 presents the solutions of the two subproblems. The joint channel access and sampling rate control scheme is proposed in Sect. 5.4. Simulation results are provided in Sect. 5.5 to evaluate the performance of the proposed scheme. Finally, Sect. 5.6 concludes the chapter.

5.1 System Model and Problem Formulation 5.1.1 Network Model Consider a cognitive radio sensor network with a set of cognitive sensor nodes N D f1; : : : ; Ng distributed to monitor an area of interest, as shown in Fig. 5.1. Each sensor node is equipped with energy harvesting devices to harvest energy from ambient environment and a rechargeable battery to store the harvested energy. The time cycle of the network is defined as an energy harvesting period T , and each time cycle is divided into a set of time slots T D f1; : : : ; Tg. During each time slot t, sensor node i senses the monitored area with a sampling rate si;t and transmits the sensory data to the sink via a static multi-hop routing path, which makes the network topology be a tree with the sink as the root.

5.1 System Model and Problem Formulation

Spectrum Market

83

PU

Sink Base Station

PU

Sink

PU Sensor node

Primary user

Primary Network

Licensed channel

Fig. 5.1 The architecture of CRSN

The sensor nodes originally work on a unlicensed channel (e.g., 2.4 GHz), but they are embedded with cognitive radio modules, which enable them to opportunistically access licensed channels by adjusting their radio parameters. Meanwhile, there are a number of overlapping wireless applications working on the same unlicensed channel, causing significant and uncontrollable interference to the CRSN. As a result, sensor nodes have to sense and access licensed channels for data transmission to guarantee the required network performances (e.g., delay and throughput). There are a set of licensed channels C D f1; : : : ; Kg with different channel capacities fC1 ; : : : ; CK g in the primary network coexisting with the CRSN .1 The primary user (PU) behavior is assumed to be stationary and ergodic over the K channels [9]. The cognitive sensor nodes of the CRSN are secondary users (SUs) in the primary network and can opportunistically access the idle licensed channels. There exists a spectrum market that is in charge of channel sensing and providing available licensed channel information to SUs (e.g., sensor nodes in the CRSN). At the beginning of each time slot, the spectrum market senses the licensed channels and provide the available channels that can be accessed during the time slot. Denote the available licensed channel set at time slot t as Ct , and we have Ct C. Due to the stochastic PU traffic over the licensed channels, the available licensed channel

1 TV white spaces are the considered licensed channels in our network scenario. A TV tower is the primary user in the primary network, which can make the availability of each licensed channel the same to all the sensor nodes.

84

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

Table 5.1 The key notations Notation N ,T ,C i or j t or h, k Ct A, R , B

si;t , S zi;t;k , Z Di;t Ck es , et , er , ew i;t w Ei;t u Ri;t t Ii , Si i;t

I.a/ ' ˛, ˇ , ,  yi;t;k , Y Qp , N p

Definition Set of sensor nodes, time slots and licensed channels Index of sensor node Index of time slot and licensed channel Available channel set at time slot t Set of ancestors, relays and neighbors i’s sampling rate at time slot t, sampling rate matrix 0-1 variable denoting whether i accesses channel k at time slot t, channel access decision matrix i’s data transmission rate at time slot t Capacity of channel k Energy consumption for data sensing, transmitting, receiving and channel switching Energy consumption rate of node i at time slot t i’s channel switching energy consumption at t Energy consumption upper bound of i at time slot t Duration of a time slot Interference set of node i, set of node i’s sons i’s energy harvesting rate at time slot t Indicator function, if a is true, I.a/ D 1; otherwise, I.a/ D 0 Channel cost for the channel with unit capacity Lagrangian multiplier matrices Intermediate variable matrices 0-1 variable denoting whether TRS p accesses channel k at time slot t, channel access decision matrix Set of transmitters in TRS p, Set of sensor nodes (including transmitters and the receiver) in TRS p

set Ct may vary at different time slots. Moreover, at each time slot t, the spectrum market can predict the availability of each licensed channel k 2 C for the next T time slots based on historical channel sensing results with high accuracy, by hidden Markov models [10] or neural networks [11]. In other words, the CRSN can obtain the predicted available channel sets fC1 ; : : : ; CT g over the whole period T . During each time slot t, the CRSN schedules each node i to access a licensed channel k 2 Ct to transmit its data. To coordinate the channel accessing, a common channel is assumed to be available to exchange control information among sensor nodes and the sink. To ease the presentation, the key notations are listed in Table 5.1.

5.1 System Model and Problem Formulation

85

5.1.2 Communication Link Model According to the network model, node i senses data at si;t and transmits its sensory data via a fixed routing path to the sink. Let Ri be the set of relays for node i, Ai , fj j i 2 Rj g be the set of ancestors that use node i as a relay along their routing paths, and Bi denote the set of brothers that have the same next hop as node i. There are only a transmitter and a receiver embedded in each sensor node as well as the sink. Both of transmitter and receiver can independently adjust their radio parameters to use a specific channel independently. It indicates that each sensor node should transmit or receive data over a single channel, but the channels used for data transmission and reception can be different. We use a binary variable zi;t;k to denote whether node i’s transmitter is scheduled to access channel k during time slot t. If i is scheduled to access channel k, zi;t;k D 1; otherwise, zi;t;k D 0. Since each transmitter can only access a specific channel, we have X zi;t;k D 1; 8i 2 N ; 8t 2 T : (5.1) k2Ct

Meanwhile, each data receiver, including each relay node and the sink, can only receive data over a specific channel, which indicates that node i and its brothers should access the same channel. Thus, we have, zi;t;k D zj;t;k ; 8j 2 Bi ; 8k 2 C; 8i 2 N ; 8t 2 T :

(5.2)

Furthermore, if node i and the nodes of Bi are scheduled to access channel k, the channel capacity should be no less than the sum of node i’s transmission rate and its brothers’ transmission rates to avoid link congestion P [12]. Let Di;t be the data transmission rate of node i, then we have Di;t D si;t C j2Ai;t sj;t , and the link capacity constraint X X Dj;t  zi;t;k Ck ; 8i 2 N ; 8t 2 T : (5.3) Di;t C j2Bi;t

k2Ct

Actually, the utilization of channel capacity Ck depends on the number of nodes in Bi;t and the adopted MAC protocol. To focus on the network utility optimization in high network layers, we consider that the MAC protocol can provide full utilization of the channel capacity.

5.1.3 Energy Harvesting and Consumption Model Let i;t denote the energy harvesting rate of node i at time slot t, which is assumed to be stable during the time slot and can be predicted with high accuracy for the next T time slots. Each sensor node consumes energy in data sensing, transmitting and

86

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

receiving. For each node i, let es , et and er be the energy consumption rates for data sensing, transmitting and receiving, respectively. Thus, during each time slot t, the energy consumption rate i;t of node i is defined as X i;t , .es C et /  si;t C .er C et /  sj;t : (5.4) j2Ai

In addition, there is energy consumption associated with node i for channel switching, if its accessed channel changes from time slot t  1 to t [9]. Let Si be the set of i’s sons who use i as the next hop, and s be one of the sensor nodes in Si . Therefore, for each i 2 N , i’s energy consumption for channel switching is P P w Ei;t , k2Ct I.zs;t1;k ¤ zs;t;k /ew zi;t;k C k2Ct I.zi;t1;k ¤ zi;t;k /ew zi;t;k , where ew is the energy consumption for channel switching, and I.a/ is an indicator function such that if a is true, I.a/ D 1; otherwise, I.a/ D 0. Note that, if node i is a leaf node, Si D ; and I.zs;t1;k ¤ zs;t;k / D 0. We assume that the battery capacity is large enough to store the harvested energy, such that we only focus on controlling the energy consumption rates to guarantee the sustainability of sensor nodes. Let t be the duration of each time slot and ri;t be i’s residual energy at the end of time slot t. Then, we have ri;t D ri;t1 C i;t t  w i;t t  Ei;t . Denote ri;0 as i’s initial energy, which is known in advance. ri;t can be recursively calculated as ri;t D ri;0 C

t X

i;h t

hD1



t X

i;h t 

hD1

t X

w Ei;h :

(5.5)

hD1

In order to guarantee that sensor nodes have enough energy to switch channel at the beginning of each time slot t, we have ri;t  ri;w , where ri;w is i’s reserved energy for channel switching. If i is a leaf node, ri;w D ew ; otherwise ri;w D 2ew . Therefore, for each node i at time slot t, we have the following energy consumption constraint: ri;0 C

t X

i;h t 

hD1

t X

i;h t 

hD1

t X

w Ei;h  ri;w :

(5.6)

hD1

P Since ri;0 , ri;w and thD1 i;h t are constant in the constraint, we define the P energy consumption upper bound as Rui;t , ri;0 C thD1 i;h t  ri;w and rewrite the energy consumption as t X hD1

i;h t C

t X

w Ei;h  Rui;t ; 8i 2 N ; 8t 2 T :

(5.7)

hD1

5.1.4 Channel Interference Model Due to the broadcast nature of wireless channels, sensor nodes may cause interference to the sensor nodes within their interference range, if they are scheduled to access the same channel [13]. Since all the sensor nodes have a fixed transmission

5.1 System Model and Problem Formulation

Routing path

87

Interference range

Sink a

c b

d

g

e

h f

Fig. 5.2 Interference set illustration

power, the interference range of each sensor node is also fixed. For two communication links a ! b and c ! d, if node d (or b) is in the interference range of node a (or c), a and c cannot operate on the same channel to avoid interference. Let Ii be the interference set of node i, which consists of the sensor nodes that cannot access the same channel as node i. Therefore, for 8i 2 N , we have X  zi;t;k  zj;t;k D 0; 8j 2 Ii ; 8t 2 T ; (5.8) k2Ct

where the interference set Ii has no intersection with Bi , i.e., Ii \ Bi D ;. Figure 5.2 illustrates a specific case with channel interference. According to the interference definition, we have Ia D ff g, Ib D fe; g; hg, Ic D ff g, Id D ff g, Ie D ff g, If D fa; c; d; e; g; hg, Ig D fb; f g and Ih D ff g.

5.1.5 Problem Formulation The objective is to maximize the network utility by controlling the sampling rate si;t and determining the channel access zi;t;k for each node i over a period T . The network utility is defined as the utility of sensory data subtracting the channel cost [14, 15], while the sensory data utility and channel cost are specifically defined as follows. Let V.i; t/ be the utility of i’s sensory data at time slot t. Then, the P P total utility of sensory data is V , t2T i2N V.i; t/. The utility function V./ is assumed to be increasing and strictly concave to guarantee the fairness for

88

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

sensor nodes [14]. For example, V.i; t/ D log.1 C si;t /. On the other hand, since the spectrum market periodically senses the licensed channels and provides the available channel information to the CRSN, the CRSN should pay for the channels accessed by sensor nodes, which is defined as channel cost. We consider that the channel cost depends on the total channel capacity of the accessed channels. Let ' be the then, the channel cost is W ,  capacity,P P P cost for Pthe channel with unit I. , where I. z ¤ 0/  'C k t2T k2Ct i2N i;t;k i2N zi;t;k ¤ 0/ is to indicate whether channel k is accessed by the CRSN. In summary, the network utility is defined as ! ! XX XX X U, log.1 C si;t /  zi;t;k ¤ 0 'Ck : I (5.9) t2T i2N

t2T k2Ct

i2N

Let a matrix S , fsi;t j 8i 2 N ; 8t 2 T g and a binary matrix Z D fzi;t;k j 8i 2 N ; 8t 2 T ; 8k 2 Ct g denote the allocated sampling rates and channels for sensor nodes over time period T , respectively, the network utility maximization problem can be formulated as .NUMP/ max U S;Z

8 ˆ ˆ < (5.1), (5.2), (5.3), (5.7) and (5.8) s:t:

si;t  0 ˆ ˆ :z D 0 or 1; 8k 2 C i;t;k t

; 8i 2 N ; 8t 2 T :

Since Eq. (5.8) is a non-linear constraint, (NUMP) is a mix-integer non-linear optimization problem, which is challenging to solve. In the following section, we will focus on solving this intractable optimization problem.

5.2 Problem Decomposition and Solution In this section, we aim to decouple (NUMP) into separable subproblems by dual decomposition, and then tackle the subproblems separately and efficiently.

5.2.1 Problem Decomposition and Dual Problem Since Eqs. (5.3) and (5.7) are the constraints coupling the two decision variable matrices S and Z, we can decouple them by relaxing these two constraints with Lagrangian relaxation. We introduce two Lagrangian multiplier matrices ˛ ,

5.2 Problem Decomposition and Solution

89

f˛i;t j 8i 2 N ; 8t 2 T g and ˇ , fˇi;t j 8i 2 N ; 8t 2 T g, and define the Lagrangian associated with the primal problem (NUMP) as L.˛; ˇ/ D

XX

log.1 C si;t / 

t2T i2N

C

XX

2 ˛i;t 4

t2T i2N

C

XX

XX

X

I

t2T k2Ct

X

X

3

Dj;t 5

j2Bi;t

t t h i X X w ; ˇi;t Rui;t  i;h t  Ei;h hD1

t2T i2N

!

zi;t;k ¤ 0 'Ck

i2N

zi;t;k Ck  Di;t 

k2Ct

!

(5.10)

hD1

where the Lagrangian multipliers ˛i;t and ˇi;t should be no less than 0, i.e., ˛i;t  0 and ˇi;t  0, 8i 2 N ; 8t 2 T . According to the communication link model, we have 8 ! ! ˆ P P P P ˆ ˆ ˛i;t Di;t C Dj;t D Di;t ˛i;t C ˛j;t ˆ < i2N j2Bi i2N j2Bi ! ! (5.11) ˆ P P P P ˆ ˆ ˛ C s s C ˛ s D ˛ ˆ i;t i;t j;t i;t i;t j;t : i2N

j2Ai

i2N

j2Ri

Thus, if we define the following intermediate variables 8 " # ˆ P P P ˆ ˆ i;t , ˛i;t C ˛j;t C ˛x;t ˛j;t C ˆ ˆ ˆ j2Bi j2Ri x2Bj ˆ ˆ # " < t P P i;t , ˇj;h .er C et /  t ˇi;h .es C et / C ˆ ˆ hD1 j2Ri ˆ ˆ ˆ t ˆ P ˆ ˆ ˇi;h :i;t ,

;

(5.12)

hD1

the Lagrangian can be rewritten as L.˛; ˇ; S; Z/ D

XXh

log.1 C si;t /  .i;t C i;t /si;t

t2T i2N



XX t2T k2Ct

C

XX t2T i2N

I 2

X

!

!

zi;t;k ¤ 0  'Ck

i2N

4˛i;t

i

X k2Ct

3 w zi;t;k Ck  i;t Ei;t C ˇi;t Rui;t 5 :

(5.13)

90

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

Based on the Lagrangian [i.e., Eq. (5.13)], the dual function (i.e., the objective function of the dual problem) is the maximum value of the Lagrangian over S and Z: D.˛; ˇ/ D sup L.˛; ˇ; S; Z/:

(5.14)

S;Z

Due to the separability of the two variable matrices, we define two subproblems to decouple S and Z: .SP1/ D1 .˛; ˇ/ , max S

XXh

log.1 C si;t /  .i;t C i;t /si;t

i

t2T i2N

s:t: si;t  0; 8i 2 N ; 8t 2 T : 2 3 XX X w5 4˛i;t .SP2/ D2 .˛; ˇ/ , max zi;t;k Ck  i;t Ei;t Z

 ( s:t:

XX t2T k2Ct

I

t2T i2N

X

k2Ct

!

!

zi;t;k ¤ 0  'Ck

i2N

(5.1), (5.2) and (5.8) zi;t;k D 0 or 1; 8k 2 Ct

; 8i 2 N ; 8t 2 T :

Based on the definition of (SP1) and (SP2), the dual function can be rewritten as D.˛; ˇ/ D D1 .˛; ˇ/ C D2 .˛; ˇ/ C

XX

ˇi;t Rui;t :

(5.15)

t2T i2N

The dual problem is to minimize the dual function over the Lagrangian multiplier matrices ˛ and ˇ: .DP  NUMP/ min D.˛; ˇ/ ˛;ˇ

s:t: ˛i;t ; ˇi;t  0; 8i 2 N ; 8t 2 T : According to the analysis and proof in [16, 17], only weak duality can be guaranteed by Lagrangian relaxation, which indicates that there exists a duality gap between the optimal solutions to the dual problem and the primal problem. Let Op and Od be the optimal results of (NUMP) and (DP-NUMP). We have, Op  Od holds for all feasible solutions and Od actually becomes the upper bound of Op [18].

5.3 Subproblem Solutions

91

5.2.2 Subgradient Method for Solving Dual Problem Given ˛ and ˇ, if we can address (SP1) and (SP2) separately, the dual problem (DPNUMP) can be iteratively solved using a subgradient method. Specifically, during each iteration, the Lagrangian multipliers are updated in an opposite direction to the partial gradient of the Lagrangian dual function [14, 19], i.e., (

˛i;t .m C 1/ D Œ˛i;t .m/  ˛  f˛;i;t .m/ C ˇi;t .m C 1/ D Œˇi;t .m/  ˇ  fˇ;i;t .m/ C

(5.16)

where m 2 NC is the iteration index; ˛ > 0 and ˇ > 0 are the step sizes adjusting the convergence rate; Œa C , maxfa; 0g and f˛;i .m/ and fˇ;i .m/ are subgradients of dual function with respect to ˛i;t and ˇi;t , respectively, 8 P @D.˛; ˇ/ ˆ ˆ D Ck  Di;t .m/  Dj;t .m/ 0. Since the feasible @si;t

set of f .si;t / is si;t  0, the optimal solution of f .si;t / is achieved at si;t j si;t D si;t j

@f @si;t

D0

D

1  1. ln 2  .i;t C i;t /

@f @si;t

D0

, i.e.,

5.3.2 Computation Complexity Analysis on Channel Allocation Given ˛ and ˇ, (SP2) is a channel access problem to maximize D2 .˛; ˇ/ by determining the optimal binary matrix Z . In the following, we prove that (SP2) is a NP-hard problem which cannot be optimally solved in polynomial time. We aim to reduce (SP2) to the vertex K-coloring problem (VKCP), which is a classic NP-hard problem. Before that, we first define VKCP as follows. Definition 2 The vertex K-coloring problem (VKCP): Given an undirected graph G D fV; Eg and K colors, where K is an integer with K  0, using at most K colors to find a coloring solution (i.e., assigning one of the K colors to each vertex) such that no two vertices sharing the same edge have the same color and the number of colors used to color the vertices is the smallest. Lemma 6 VKCP is a NP-hard problem when K  3.

5.3 Subproblem Solutions

93

c

Sink TRS 1

TRS 3

a

c

h

g

b a TRS 2

b

d

TRS 4

e

f

Fig. 5.3 Illustration of TRS definition

Proof When K  3, according to the analysis in [20] and [21], the VKCP can be reduced to the Boolean Satisfiability (SAT) problem, which is one of the first problems that is proven to be NP-complete. Thus, VKCP is NP-hard and its associated decision problem is NP-complete. The proof of Lemma 6 can be found in [20] and [21]. In order to reduce (SP2) into VKCP, we reconstruct the problem as follows. We divide the network routing tree into a number of transmitter-receiver sets (TRSs). Each TRS consists of a data receiver and its sons. For a specific network routing tree, data receivers include all non-leaf sensor nodes and the sink. To illustrate the definition of TRS, Fig. 5.3 shows an example of dividing the network case in Fig. 5.2 into four TRSs. The red dash link between two TRSs denotes that there exists interference between them, which means any pair of connected TRSs cannot access the same channel to avoid interference. According to the communication link model, the channel access in the CRSN is a receiver-based access problem, which means that the sensor nodes of each TRS should access the same channel for data transmission and reception. Let P denote the set of TRSs in the network. Then, the channel access problem changes to schedule each TRS p 2 P to access a specific channel k 2 Ct . We use a binary matrix Y D fyp;t;k j 8p 2 P; 8t 2 T ; 8k 2 Ct g to denote whether TRS p accesses channel k at time slot t. Meanwhile, if we let Qp be the set of transmitters in TRS p, respectively, we have zi;t;k D yp;t;k j 8p 2 P; 8t 2 T ; 8k 2 Ct for each i 2 Qp . Then, the constraints of (SP2) can be rewritten as follows: P (a) Eq. (5.1) changes to k2Ct yp;t;k D 1; 8 p 2 P; t 2 T ; (b) Eq. (5.2) is removed, since all the sensor nodes of a TRS can only access one channel; P (c) Eq. (5.8) changes to k2Ct .yp;t;k  yq;t;k / D 0; 8q 2 Ip ; 8 p 2 P; 8t 2 T , where Ip denotes the set of TRSs that may have interference with TRS p.

94

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

Meanwhile, the objective function D2 .˛; ˇ/ can be rewritten with respect to Y as TD2 .˛; ˇ/ D

XX

0 @

t2T p2P



X

˛i;t

i2Qp

XX

0 0 @I @

t2T k2Ct

X

yp;t;k Ck 

k2Ct

X

1

X i2Np

1 w A ˇi;t Ep;t

1

yp;t;k ¤ 0A  'Ck A ;

(5.20)

p2P

where Np denotes the set of sensor nodes in TRS p including the transmitters and P w receivers and Ep;t , k2Ct I.yp;t1;k ¤ yp;t;k /ew yp;t;k . Thus, (SP2) can be transformed to an equivalent problem (SP2-E), which is to determine Y to .SP2  E/ max TD2 .˛; ˇ/ Yt

8P ˆ ˆ ˆk2C yp;t;k D 1 ˆ < Pt s:t: .y  y / D 0; 8q 2 Ip ˆk2Ct p;t;k q;t;k ˆ ˆ ˆ : yp;t;k D 0 or 1; 8k 2 Ct

; 8p 2 P; 8t 2 T :

Theorem 1 In period T , if there exists a time slot t with jCt j  3, (SP2-E) is NP-hard. Proof Since ew , ' and Ck are independent and constant parameters in (SP2-E) with ew  0, '  0 and Ck  0, we can set ew D 0, ' D 1 and Ck D 1 for each k 2 Ct . Then, the problem becomes a specific case of (SP2-E) with the objective function

max Y

Since

P P P

XX X

˛i;t 

t2T p2P i2Qp

XX

0 I@

t2T k2Ct

˛i;t is constant and

t2T p2P i2Qp

P k2Ct

X

1 yi;t;k ; 0A :

p2P

P

I

! yi;t;k ; 0

is independent with

p2P

time slot t, the equivalent problem of the specific case of (SP2-E) is to determine Yt D fyp;t;k j 8p 2 P; 8k 2 Ct g for each time slot t 2 T to

min Yt

X k2Ct

0 I@

X

1 yi;t;k ; 0A

p2P

s:t: the constraints of (SP2-E):

5.3 Subproblem Solutions

95

Next, we prove that this problem is equivalent to VKCP. We create an undirected graph G D fV; Eg, where each vertex in V corresponds to a TRS in P. Meanwhile, for each vertex p 2 P and vertex q 2 Ip , there is an edge connecting p and q. Apparently, this creation can be achieved in polynomial time. Let Ct be the set of colors for coloring the vertices in G. The problem is equivalent to coloring the undirected graph G with jCt j colors, such that no two vertices sharing the same edge have the same color and the number of colors used to color the vertices is the smallest. Therefore, the problem is equivalent to VKCP and hence is NP-hard when jCt j  3. Since the specific case of (SP2-E) is to solve the problem at different time slots independently, it is NP-hard if there exists a time slot t with jCt j  3 over period T . Under this condition, as the specific case is NP-hard, it is sufficient to prove that (SP2-E) is NP-hard [22]. Corollary 3 In period T , if there exists a time slot t with jCt j  3, (SP2) is NP-hard. Corollary 3 always holds, since (SP2) is equivalent to (SP2-E). We consider that the number of available channels provided by the spectrum market is usually larger than 3 [2], which means that we cannot obtain the optimal solution of .SP2/ in polynomial time.

5.3.3 Suboptimal Solution for Channel Allocation In the following, we propose a cross-entropy based heuristic algorithm (CEHA) to determine a suboptimal solution for the problem in an efficient way. Due to the equivalence of (SP2) and (SP2-E), we focus on solving (SP2-E) and use the solution of (SP2-E) to derive the solution of (SP2). The main idea of cross-entropy methods is to transform a deterministic problem to an associated stochastic problem, which has been widely used in combinational optimization for suboptimal results [23]. CEHA defines a matrix H , fHp;t;k j 8p 2 P; 8t 2 T ; 8k 2 Ct g, where Hp;t;k denotes the probability of TRS p accessing channel k at time slot t. For each p 2 P at each t 2 T , Hp;t;k is initialized as a specific value for each k 2 Ct to make P k2Ct Hp;t;k D 1. As a result, Hp;t .k/ , Pr.X D k/ D Hp;t;k is actually a PMF function for k 2 Ct . Then, CEHA finds the solution by an iterative procedure, where each iteration consists of two phases. In the first phase, a set of random samples are generated according to the PMF for each p 2 P and t 2 T , which are used to calculate the corresponding solutions. In the second phase, the PMFs are updated based on some better solutions for the next iteration. Specifically, we describe the detailed procedures of CEHA as follows. 1. Constraint Relaxation—We introduce a constant to relax the second constraint of (SP2-E), which leads to the modified objective function 0 1 X XX X TD2 .˛; ˇ/ D TD2 .˛; ˇ/   I @ .yq;t;k  yq;t;k / ¤ 0A ; t2T p2P q2Ip

k2Ct

96

2.

3.

4.

5.

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

P P P P where , t2T p2P i2Qp k2Ct ˛i;t Ck is a penalty to objective function if the constraint of TD2 .˛; ˇ/ is violated. Notably, is an upper bound of TD2 .˛; ˇ/ to guarantee that every feasible solution without violating the constraint will achieve a lower objective value than the infeasible solutions. Initialization—Initialize the maximum iteration number  and a difference threshold . Set the iteration counter  D 1. For each time slot t, we index the available channel set Ct as f1; : : : ; jCt jg, and set TRS p accesses each channel k 2 Ct with equal probabilities, i.e., for each k 2 Ct , the PMF Hp;t is initialized  D jC1t j . as Hp;t .k/ D Hp;t;k Sample Generation—Use Hp;t to generate M values for each p and t. Each value denotes that p will access which channel at time slot t. For the m-th value Xm (1  m  M), we create a binary string xp;t;m containing jCt j digits, where the Xm -th digit is 1 and the other digits are 0. For example, if jCt j D 4 and X1 D 2, we have xp;t;1 D 0100 which means p will access the 2nd channel in Ct at time slot t. Such that, M samples can be generated by the strings, with the m-th sample as fxp;t;m j p 2 P; t 2 T g. Sample Evaluation—We use the M samples to calculate the corresponding values of TD2 .˛; ˇ/. Note that, for the m-th sample, if the k-th digit of xp;t;m is 1, it denotes yp;t;k D 1, and yp;t;l D 0 for each l 2 Ct  fkg. Let m be the value of TD2 .˛; ˇ/ using the m-th sample. Then, sort the M values of TD2 .˛; ˇ/ in a non-decreasing order and let  be the sorted set of the M values. We denote the corresponding samples of the first dMe values in  as a set M . /, which are called as “better” samples in this iteration. PMF Update—The PMF is updated based on M . / in the iteration. For each p 2 P and t 2 T , we update H C1 as P m2M . / G.xp;t;m ; k/  C1 ; k D 1; : : : ; jCt j; Hp;t;k D dMe

where G.a; b/ is a function to obtain the k-th digit of a, e.g., G.0100; 2/ D 1 and G.0010; 2/ D 0. 6. Stopping Rule—The iteration stops if either of the following two conditions is met. The first is the iteration count achieves the maximum iteration number  . The second is the difference of H and H C1 is lower than a required threshold, i.e., jjH C1  H jjF  , where jj  jjF is the Frobenius norm, and  is the required threshold. Otherwise, set  D  C 1 and go back to step (2).  7. Result Output—Generate a sample fxp;t j p 2 P; t 2 T g with the PMFs  C1 fHp;t j 8p 2 P; 8t 2 T g, and then use the sample to generate the solution Y  of (SP2-E). Since (SP2-E) is equivalent to (SP2), the solution of (SP2) is zi;t;k D yp;t;k ; 8p 2 P; 8t 2 T ; 8k 2 Ct ; 8i 2 Qp . Output Zt D fzi;t;k j 8i 2 N ; 8t 2 T ; 8k 2 Ct g. Note that, the stopping rules of ECHA can guarantee the computational complexity within a controllable range. Meanwhile, the update of PMF in each iteration can make the generated samples toward better solutions.

5.4 JASC: Joint Channel Allocation and Sampling Rate Control Scheme

97

5.4 JASC: Joint Channel Allocation and Sampling Rate Control Scheme From the discussion of the previous subsections, we can efficiently solve the subproblems of the dual problem (DP-NUMP) with given ˛ and ˇ. In this section, we summarize the steps of solving (NUMP) by the subgradient method and propose a joint channel access and sampling rate control scheme, named JASC, to maximize the network utility.

5.4.1 Algorithm for Solving Primal Problem We first focus on the primal problem (NUMP), which is to schedule the channel access and sampling rates of sensor nodes over a period. According to the system model, except the decision variables S and Z, all the parameters of (NUMP) are known in advance or can be predicted for the next T time slots, including the energy harvesting rate i;t , and the available channel set Ct . (NUMP) can be solved by addressing its dual problem (DP-NUMP) with the subgradient method discussed in Sect. 5.2.2. Since there is a duality gap between (NUMP) and (DP-NUMP), we introduce a duality gap threshold to terminate the iteration of solving (DP-NUMP). Moreover, a maximum iteration number ˘ is predefined to guarantee the efficiency of the algorithm when the convergence is slow. Specifically, the main idea of solving (NUMP) is described in Algorithm 6. Algorithm 6 Subgradient method for solving (NUMP) Input: The parameters of (NUMP), the maximum iteration number ˘ and duality gap threshold ı. Output: The optimal S and Y  . 1: Let m D 1; Initialize Lagrangian multipliers ˛i;t .m/ D 1 and ˇi;t .m/ D 1; 2: repeat 3: With given ˛.m/ and ˇ.m/, determine the optimal S .m/ and Y  .m/ by solving (SP1) and (SP2) according to Proposition 8 and the CEHA in Sect. 5.3.3, respectively; 4: With the derived S .m/ and Y  .m/, calculate the values of U and D.˛; ˇ/ as U  .m/ and D .˛.m/; ˇ.m//, according to Eqs. (5.9) and (5.15), respectively; 5: Generate ˛ .m C 1/ and ˇ  .m C 1/ for the next iteration according to Eq. (5.16); 6: m D m C 1; 7: until D .˛.m/; ˇ.m//  U  .m/  ! or m > ˘ ; 8: return S .m/ and Y  .m/;

98

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

5.4.2 The Proposed JASC Scheme In Algorithm 6, the channel access and sampling rate control solution is determined based on the predicted energy harvesting rates and available channel sets over the whole period. Although existing prediction algorithms can achieve an acceptable accuracy, the prediction error, especially in the availability of licensed channels, cannot be entirely avoided and will impact the solution of Algorithm 6. For example, if channel k is wrongly predicted as available at time slot t and is scheduled to be accessed by the CRSN, the CRSN will suffer from significant interference caused by PUs. Therefore, to avoid such situation, the CRSN should communicate with the spectrum market to require the real-time available channel set Ct0 at the beginning of each time slot t and use the real-time available set to adjust the channel access and sampling rate control solution. To this end, we propose a joint channel access and sampling rate control scheme, named JASC, which can provide a real-time solution based on the real-time Ct0 and 0 energy harvesting rate i;t to maximize the network utility. Figure 5.4 shows the main ideas of the JASC scheme and Algorithm 7 presents the detailed procedures.

Sensor nodes

Spectrum market Realtime sensing results

NUMP

α β

y

Realtime energy harvesting rates

NUMP-DP

α

+

β

+

+ +

SP1

SP2

Sampling Rate Control

Dynamic Channel Access

Fig. 5.4 Illustration of the JASC scheme

5.5 Performance Evaluation

99

Algorithm 7 Joint channel access and sampling rate control scheme (JASC) Input: The inputs of Algorithm 6. Output: The optimal sampling rates S0t and channel access schedule Y 0t at each time slot t. 1: t D 1; 2: repeat 3: At the beginning of time slot t, communicate with the spectrum market to require the realistic available channel set Ct0 and obtain the energy harvesting 0 0 rates of sensor nodes f 1;t ; : : : ; N;t g; 0 0 4: Use Ct and i;t , as well as the predicted available sets fCtC1 ; : : : ; CT g and energy harvesting rates f i;l j 1  i  N; t C 1  l  Tg, as input to run Algorithm 6 to obtain the real-time sampling rates S0t D fs01;t ; : : : ; s0N;t g and channel access schedule Y 0t D fyi;t;k j 8i 2 N ; 8k 2 Ct0 g for time slot t; 5: Adjust the sampling rates and schedule the channel access of sensor nodes according to S0t and Y 0t ; 6: t D t C 1; 7: until t  T; 8: return S0t and Y 0t for each t 2 T ; Table 5.2 Parameter settings

Parameters es et jN j ri;0 Cm

Settings 0:0013 J=Kb 0:0046 J=Kb 8 0:05 J 5  m Kb=s

Parameters er t jT j ew m

Settings 0:0024 J=Kb 300 s 288 104 J Œ1; : : : ; 10

5.5 Performance Evaluation In this section, we evaluate the performance of the proposed JASC scheme by extensive simulations on OMNET++ [12, 24]. We setup a network consisting of eight sensor nodes and a sink node. The network topology is the same as Fig. 5.2. Each sensor node has a 30  30 mm2 solar photovoltaic panel with an energy conversion efficiency 20%, and a rechargeable battery with enough capacity [14]. For example, if the solar radiation data during a time slot is 100 W=m2 , the corresponding energy harvesting rate is 0:018 J=s. The energy consumption rates for data sampling, receiving, and transmitting are 0:0013 J=Kb, 0:0024 J=Kb, and 0:0046 J=Kb, respectively, which are adopted from the measurements on the Mica 2 platform [25, 26]. The initial battery level is 0:05 J. The energy harvesting cycle is a day, and we divide each cycle into 24  6 time slots. Then, the duration of each time slot is 10 min. Meanwhile, there are ten licensed channels in the primary network. The capacity of the m-th (1  m  10) licensed channel is 5 m Kb=s, and the idle probability of each licensed channel is 65%, and the channel cost ' is 0.02. The main parameter settings are summarized in Table 5.2.

100

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs 0.21

Harvested Solar Energy (J)

0.18 0.15

Jan 7, 2014 Mar 30, 2014 June 25, 2014 Oct 15, 2014

0.12 0.09 0.06 0.03 0.00 4:00

6:00

8:00 10:00 12:00 14:00 16:00 18:00 20:00

Time of A Day

Fig. 5.5 Energy harvesting rates on different days

The energy harvesting rates of sensor nodes are set according to the real solar data collected by the NREL Solar Radiation Research Laboratory in Rancho Cordova, California [27]. We choose the solar radiation data on four different days, which can provide different amount of solar energy for sensor nodes, to generate the energy harvesting profiles. The energy harvesting rates of a sensor node in the 4 days are shown in Fig. 5.5. To evaluate the effectiveness of JASC, we compare it with an existing method, named DSCC [14]. Since DSCC is proposed for sampling rate control in traditional rechargeable sensor networks, we employ a dynamic channel allocation algorithm named GBCA [28], to collaborate with DSCC for sampling rate and channel access control in cognitive radio sensor networks. We first dynamically allocate the available licensed channels to sensor nodes by GBCA, and then use DSCC to schedule the sampling rates of sensor nodes for network utility maximization.

5.5.1 Utility Comparison and Efficiency Evaluation Figure 5.6 shows the network utility comparison on different days. It can be seen from the figure that JASC can achieve improved network utility than DSCC + GBCA on different days. The improved ratio is close to 10%. Especially, when the harvested solar energy is plentiful, e.g. on June 25, 2014, the improved ratio is approaching 15%. Figure 5.7 shows the network utility comparison under different network scales. It can be seen that JASC can outperform DSCC + GBCA under different network scales, in terms of network utility. Moreover, with the increasing number of sensor nodes, the network utility of both schemes gradually increases. However, it experiences very slight increment after the number of sensor nodes

5.5 Performance Evaluation

101

15

Network Utility (*100)

DSCC+GBCA JASC

12

9

6

3 Jan. 7

Mar. 30

June 25

Oct. 15

Different Days

Fig. 5.6 Network utility comparison on different days

Network Utility (*100)

35 DSCC+GBCA JASC

30

25

20

15

10 10

20

30

40

50

60

Total Number of Sensor Nodes

Fig. 5.7 Network utility comparison under different network scales

achieves 40. This is because the limited network capacity will significantly decrease the sampling rates of sensor nodes under a large network scale. Figure 5.8 shows the converge speed comparison of JASC under different network scales. As shown in the figure, the network utility of JASC can quickly converge to a certain value within 20 iterations, when there are eight sensor nodes in the network. But when the network scale increases to 20 sensor nodes, JASC has to experience nearly 50 iterations to achieve a converged network utility. Although the increasing network scale can degrade the converge speed of JASC, JASC still has a high efficiency to achieve a converged network utility.

102

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs 25

Network Utility (*100)

20 15 10 8 sensor nodes 20 sensor nodes

5 0 0

5

10

15

20 25 30 35 Iteration Number

40

45

50

Fig. 5.8 Converge speed comparison of JASC under different network scales

5.5.2 Channel Allocation and Interference Evaluation In this subsection, we aim to evaluate the dynamic channel access of sensor nodes and channel interference probability of JASC. Figure 5.9 shows the accessed channels of different sensor nodes over different time slots under the network topology (i.e., Fig. 5.2). It can be seen that the sensor nodes in TRS1 and TRS2 keep accessing the same channel due to no interference between them, while the sensor nodes in TRS3 and TRS4 access different channels with lower bandwidth. Since TRS1 is the bottleneck of the whole network capacity, it can greatly increase the network capacity by accessing a channel with large bandwidth. However, since the harvested energy is another key factor to limit the network throughput, TRS1 choose to access the channels with low bandwidth to save channel cost when the time is before 8:00 am and after 18:00. Because during those time slots, the harvested energy can only afford a low network throughput, which can be guaranteed by accessing the channels with low bandwidth. Furthermore, the stochastic channel availability makes the accessed channel of each TRS vary over different time slots. We also compare JASC with DSCC + GBCA under different network scenarios in Fig. 5.10, in terms of channel interference probability. It can be observed that the channel interference probability increases with the number of sensor nodes under a certain number of licensed channels. This is because a larger number of sensor nodes may produce a more complicated TRS partition, which requires more available channels to avoid interference. When the number of channels is fixed, interference probability may increase with the required number of available channels.

5.5 Performance Evaluation

103

TRS1=TRS2

TRS3

TRS4

Accessed Channel ID

12 10 8 6 4 2 0 4:00

6:00

8:00

10:00

12:00 14:00 Time of A Day

16:00

18:00

20:00

Fig. 5.9 Accessed channels over different time slots

Interference Probability

20%

16% DSCC+GBCA JASC

12%

8%

4%

0% 10

20 30 40 The Number of Sensor Nodes

50

Fig. 5.10 Channel interference comparison under different network scales

5.5.3 Impacts of System Parameters In this subsection, we evaluate the impacts of two system parameters, including energy harvesting capability and channel capacity, on the performance of JASC. Figure 5.11 shows the impacts of energy harvesting capability on network utility. The energy harvesting capability in this figure is defined as the energy harvesting rate of a sensor node when the solar radiation is 1 W=m2 . It can be observed that network utility increases with energy harvesting capability under both of JASC and DSCC + GBCA. However, it does not increase indefinitely but reaches the saturation points, which are caused by the limited channel capacity, after the energy harvesting capability becomes larger than 3:8  105 J=s. Figure 5.12 shows the impacts of channel capacity on network utility. The channel capacity here is defined as the channel capacity of channel 1. It means that, for each channel 1  m  10, its channel capacity is m d Kb/s, where d is the defined channel capacity in the figure.

104

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs 25

Network Utility (*100)

20 Limited by channel capacity

15

10 DSCC+GBCA JASC

5

0 0.6

1.0

1.4 1.8 2.2 2.6 3.0 3.4 3.8 4.2 Energy Harvesting Capability (*E-5 J/s)

4.6

Fig. 5.11 Impacts of energy harvesting capability on network utility 18

Network Utility (*100)

15 12

Limited by harvested energy

9 6

DSCC+GBCA JASC

3 1

2

3

4 5 6 7 8 Channel Capability (*m Kb/s)

9

10

Fig. 5.12 Impacts of channel capacity on network utility

Similar to the situation in Fig. 5.11, network utility keeps pace with the growing channel capacity and achieves its saturation point when the channel capacity is over 7 Kb=s, due to the limited harvested energy.

5.6 Summary In this chapter, we have investigated the network utility maximization problem in EH-CRSNs by jointly considering the sampling rate control and dynamic channel access. We have formulated the problem as a mix-integer non-linear programming problem. By employing dual decomposition, the joint optimization problem is

References

105

decoupled as two separable subproblems. We have also proposed a joint channel access and sampling rate control scheme, named JASC, for utilizing the realtime sensing results to adjust the dynamic channel access and sampling rates for sensor nods. Simulation results demonstrate that the proposed JASC can achieve an improved network utility.

References 1. N. Zhang, H. Liang, N. Cheng, Y. Tang, J.W. Mark, X. Shen, Dynamic spectrum access in multi-channel cognitive radio networks. IEEE J. Sel. Areas Commun. 32(11), 2053–2064 (2014) 2. Z. Liang, S. Feng, D. Zhao, X. Shen, Delay performance analysis for supporting real-time traffic in a cognitive radio sensor network. IEEE Trans. Wirel. Commun. 10(1), 325–335 (2011) 3. A.O. Bicen, V.C. Gungor, O.B. Akan, Delay-sensitive and multimedia communication in cognitive radio sensor networks. Ad Hoc Netw. 10(5), 816–830 (2012) 4. S.-C. Lin, K.-C. Chen, Improving spectrum efficiency via in-network computations in cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 13(3), 1222–1234 (2014) 5. J. Liu, J. Gao, X. Jiang, H. Nishiyama, N. Kato, Capacity and delay of probing-based two-hop relay in MANETs. IEEE Trans. Wirel. Commun. 11(11), 4172–4183 (2012) 6. P.T.A. Quang, D.-S. Kim, Throughput-aware routing for industrial sensor networks: application to ISA100.11a. IEEE Trans. Ind. Inf. 10(1), 351–363 (2014) 7. P. Spachos, D. Hantzinakos, Scalable dynamic routing protocol for cognitive radio sensor networks. IEEE Sensors J. 14(7), 2257–2266 (2014) 8. J. Ren, Y. Zhang, R. Deng, N. Zhang, D. Zhang, X. Shen, Joint channel access and sampling rate control in energy harvesting cognitive radio sensor networks. IEEE Trans. Emerg. Top. Comput. (2016). doi:10.1109/TETC.2016.2555806. 9. J. Ren, Y. Zhang, N. Zhang, D. Zhang, X. Shen, Dynamic channel access to improve energy efficiency in cognitive radio sensor networks. IEEE Trans. Wirel. Commun. 15(5), 3143–3156 (2016) 10. X. Xing, T. Jing, W. Cheng, Y. Huo, X. Cheng, Spectrum prediction in cognitive radio networks. IEEE Wireless Commun. 20(2), 90–96 (2013) 11. V.K. Tumuluru, P. Wang, D. Niyato, A neural network based spectrum prediction scheme for cognitive radioin, in IEEE Proceedings of ICC (2010), pp. 1–5 12. J. Ren, Y. Zhang, K. Zhang, X. Shen, Adaptive and channel-aware detection of selective forwarding attacks in wireless sensor networks. IEEE Trans. Wirel. Commun. 15(5), 3718–3731 (2016) 13. J. Liu, N. Kato, Device-to-device communication overlaying two-hop multi-channel uplink cellular networks, in Proceedings of ACM MobiHoc (2015), pp. 307–316 14. R. Deng, Y. Zhang, S. He, J. Chen, X. Shen, Maximizing network utility of rechargeable sensor networks with spatiotemporally-coupled constraints. IEEE J. Sel. Areas Commun. (2016, to appear) doi:10.1109/JSAC.2016.2520181 15. R.-S. Liu, P. Sinha, C.E. Koksal, Joint energy management and resource allocation in rechargeable sensor networks, in IEEE Proceedings of INFOCOM (2010), pp. 1–9 16. M. Guzelsoy, Dual methods in mixed integer linear programming. Lehigh University (2009) 17. I. Nowak, Relaxation and Decomposition Methods for Mixed Integer Nonlinear Programming, vol. 152 (Springer Science & Business Media, Berlin, 2006) 18. S. Boyd, L. Vandenberghe, Convex Optimization (Cambridge University Press, Cambridge, 2004) 19. R. Deng, G. Xiao, R. Lu, J. Chen, Fast distributed demand response with spatially and temporally coupled constraints in smart grid. IEEE Trans. Ind. Inf. 11(6), 1597–1606 (2015)

106

5 Joint Channel Allocation and Sampling Rate Control in EH-CRSNs

20. K.R. Duffy, N. O’Connell, A. Sapozhnikov, Complexity analysis of a decentralised graph colouring algorithm. Inf. Process. Lett. 107(2), 60–63 (2008) 21. T.R. Jensen, B. Toft, Graph Coloring Problems (Wiley, New York, 2011) 22. D.P. Bovet, P. Crescenzi, D. Bovet, Introduction to the Theory of Complexity (Prentice Hall, London, 1994) 23. P.-T. De Boer, D.P. Kroese, S. Mannor, R.Y. Rubinstein, A tutorial on the cross-entropy method. Ann. Oper. Res. 134(1), 19–67 (2005) 24. J. Ren, Y. Zhang, K. Zhang, A. Liu, J. Chen, X. Shen, Lifetime and energy hole evolution analysis in data-gathering wireless sensor networks. IEEE Trans. Ind. Inf. 12(2), 788–800 (2016) 25. M. Calle, Energy consumption in wireless sensor networks using GSP. Master’s thesis, University of Pittsburgh, Pittsburgh (2006) 26. J. Polastre, J. Hill, D. Culler, Versatile low power media access for wireless sensor networks, in Proceedings of ACM SenSys (2004), pp. 95–107 27. Nrel: Midc/srrl baseline measurement system (39.74 n, 105.18 w, 1829 m, gmt-7). [Online]. Available: http://www.nrel.gov/midc/srrl_bms/ 28. J. Chen, Q. Yu, P. Cheng, Y. Sun, Y. Fan, X. Shen, Game theoretical approach for channel allocation in wireless sensor and actuator networks. IEEE Trans. Autom. Control 56(10), 2332–2344 (2011)

Chapter 6

Concluding Remarks and Future Directions

6.1 Concluding Remarks In this monograph, we have provide a comprehensive investigation on energyefficient spectrum resource management for CRSNs. Based on the analysis and discussion given in this monograph, we present the following concluding remarks. • We have introduced the basic concepts and typical architectures of WSN, CR and CRSN. Moreover, the benefits of CRSNs and some typical CRSN applications, as well as the technical challenges especially in spectrum resource management, have also been clearly presented. In addition, we have provided a comprehensive literature survey of energy-efficient spectrum resource management in CRSNs. • For spectrum access decision, we have studied the dynamic channel access problem to improve the energy efficiency in clustered CRSNs, and theoretically derived the conditions of sensing and switching to a licensed channel for potential energy saving, with a dedicate consideration of the consumption in channel sensing and switching. The analysis results can provide some guidances for spectrum access decision in CRSNs, from the perspective of energy efficiency. Moreover, we have proposed two sequential channel sensing and accessing schemes for intra and inter-cluster data transmission, respectively, forming an integrated solution to control dynamic channel access in clustered CRSNs. Extensive simulation results demonstrate that the proposed solutions can significantly reduce the energy consumption of data transmission and outperform some existing works in terms of energy efficiency. • For spectrum sensing, we have investigated the secure and energy-efficient collaborative spectrum sensing for CRSNs. We have also provided a theoretical analysis on the impacts of independent and collaborative SSDF attacks on the accuracy of collaborative spectrum sensing. Our analysis and simulation results show that the number of spectrum sensing nodes and associated global decision rule can significantly impact the accuracy of collaborative spectrum sensing. © Springer International Publishing AG 2018 J. Ren et al., Energy-Efficient Spectrum Management for Cognitive Radio Sensor Networks, DOI 10.1007/978-3-319-60318-6_6

107

108

6 Concluding Remarks and Future Directions

Therefore, we theoretically derive the minimum number of spectrum sensing nodes to achieve the trade-off between security and energy efficiency. Moreover, we have developed a trust evaluation scheme, named FastDtec, to evaluate the spectrum sensing behaviors and identify the compromised nodes. In addition, by detecting and isolating the identified compromised nodes with FastDtec, a secure and energy-efficient collaborative spectrum sensing scheme is then proposed to further enhance the energy efficiency of collaborative spectrum sensing. Extensive simulation results demonstrate that the proposed scheme can effectively resist SSDF attacks, fast and accurately identify compromised nodes, as well as improving energy efficiency. • For spectrum resource allocation, we have investigated the NUM problem in EH-CRSNs by jointly controlling the sampling rate and dynamic channel access of sensor nodes. We have theoretically formulated the NUM problem as a mixinteger non-linear programming problem. By employing dual decomposition, the formulated joint optimization problem is decoupled as two independent subproblems that can be addressed separately and efficiently. Based on the solution of the NUM problem, we have further proposed a joint channel access and sampling rate control scheme, named JASC, to utilize the real-time channel sensing results to dynamically adjust the channel access and sampling rate control. The simulation results based on a realist energy harvesting dataset demonstrate that JASC can greatly improve the network utility while keep a low co-channel interference probability.

6.2 Future Research and Development Directions As an emerging solution to address the spectrum scarcity problem in the IoT area, CRSN is still in its fancy and has many obstacles hindering the flourish of its applications. The presented spectrum management solutions and some preliminary research results in this monograph can provide valuable insights and make a step forward in this emerging and evolving field of study. However, there are still enough research space for future studies. In the following, we outline some potential research topics to foster continuous research efforts.

6.2.1 Distributed Spectrum Resource Management for Large-Scale CRSNs In large-scale CRSNs, some sensor nodes have to forward a significant amount of sensed data to the sink node. In order to improve the network lifetime and performance, multiple sink nodes or mobile sink nodes are usually deployed, accompanying with dynamic routing schemes, for balanced energy consumption.

6.2 Future Research and Development Directions

109

It consequently makes the spectrum resource management increasingly complicated and challenging. For large-scale CRSNs, centralized mechanisms/algorithms can no longer meet the QoS requirements in terms of processing delay and scalability [1–3]. However, from the spectrum management perspective, centralized mechanisms/algorithms can provide more accurate control for spectrum sensing and accessing than distributed solutions [4, 5]. Therefore, how to design efficient distributed spectrum resource management for addressing this dilemma in largescale CRSNs becomes an challenging but meaningful research direction.

6.2.2 Cross-Layer Design for Opportunistic Spectrum Access Cross-layer design has been proved as an effective solution to improve the network performance during the past decade [6, 7]. The core idea is to maintain the functionalities associated to the original layers but to allow coordination, interaction and joint optimization of protocols crossing different layers. There have been a large number of existing works focusing on cross-layer design for both wired and wireless networks. However, since most of them are proposed for the networks with specific communication channels, they are not suitable for CRSNs with opportunistic spectrum access. It motivates researchers to devote more attention on the cross-layer design for CRSNs to improve the overall network performance.

6.2.3 RF Energy Harvesting and Transfer in CRSNs The flourish of WSN and CRSN applications is highly impeded by the limited energy supply of battery-powered sensor nodes. Thus, an emerging energy harvesting technology enables sensor nodes to absorb energy from ambient RF signals, which shows strong potentials to power future CRSNs [8]. By RF energy harvesting, sensor nodes have no need to add additional energy harvesting devices, such as solar photovoltaics and wind turbines, to get recharged. Meanwhile, since CRSN has the capability of cognitively deciding to access different spectrum bands, it can sense a channel with the strongest RF signals as the energy harvesting channel and let the charging sensor nodes access to improve the charging efficiency [9, 10]. Furthermore, with such a promising technology, sensor nodes can transfer energy among each other via RF signals to alleviate the unbalanced energy consumption in CRSNs [11, 12]. Therefore, CR and RF energy harvesting can be viewed as two complementary technologies and arise a number of interesting research topics in RF-powered CRSN applications.

110

6 Concluding Remarks and Future Directions

References 1. K.S. Challapali, J. Wang, Channel management method in a distributed spectrum cognitive radio network (2013), US Patent 8,451,751 2. J. Herzen, R. Merz, P. Thiran, Distributed spectrum assignment for home WLANs, in 2013 Proceedings IEEE INFOCOM (IEEE, New York, 2013), pp. 1573–1581 3. Q. Ye, M. Al-Shalash, C. Caramanis, J.G. Andrews, Distributed resource allocation in deviceto-device enhanced cellular networks. IEEE Trans. Commun. 63(2), 441–454 (2015) 4. N. Zhang, H. Liang, N. Cheng, Y. Tang, J.W. Mark, X.S. Shen, Dynamic spectrum access in multi-channel cognitive radio networks. IEEE J. Sel. Areas Commun. 32(11), 2053–2064 (2014) 5. N. Zhang, H. Zhou, K. Zheng, N. Cheng, J.W. Mark, X. Shen, Cooperative heterogeneous framework for spectrum harvesting in cognitive cellular network. IEEE Commun. Mag. 53(5), 60–67 (2015) 6. G.A. Shah, V.C. Gungor, O.B. Akan, A cross-layer QoS-aware communication framework in cognitive radio sensor networks for smart grid applications. IEEE Trans. Ind. Inf. 9(3), 1477–1485 (2013) 7. A. Cammarano, F. L. Presti, G. Maselli, L. Pescosolido, C. Petrioli, Throughput-optimal crosslayer design for cognitive radio ad hoc networks. IEEE Trans. Parallel Distrib. Syst. 26(9), 2599–2609 (2015) 8. X. Lu, P. Wang, D. Niyato, D.I. Kim, Z. Han, Wireless networks with RF energy harvesting: a contemporary survey. IEEE Commun. Surv. Tutorials 17(2), 757–789 (2015) 9. D.T. Hoang, D. Niyato, P. Wang, D.I. Kim, Opportunistic channel access and RF energy harvesting in cognitive radio networks. IEEE J. Sel. Areas Commun. 32(11), 2039–2052 (2014) 10. S. Lee, R. Zhang, K. Huang, Opportunistic wireless energy harvesting in cognitive radio networks. IEEE Trans. Wirel. Commun. 12(9), 4788–4799 (2013) 11. Q. Shi, L. Liu, W. Xu, R. Zhang, Joint transmit beamforming and receive power splitting for MISO SWIPT systems. IEEE Trans. Wirel. Commun. 13(6), 3269–3280 (2014) 12. H.H. Chen, Y. Li, Y. Jiang, Y. Ma, B. Vucetic, Distributed power splitting for SWIPT in relay interference channels using game theory. IEEE Trans. Wirel. Commun. 14(1), 410–420 (2015)