Distributed Computing and Artificial Intelligence, Special Sessions II, 15th International Conference [1st ed.] 978-3-030-00523-8;978-3-030-00524-5

Table of contents :
Front Matter ....Pages i-viii
Image Analysis for Privacy Assessment in Social Networks (Joaquin Taverner, Ramon Ruiz, Elena del Val, Carlos Diez, Jose Alemany)....Pages 1-4
Rassel: Robot Assistant for the Elderly (Maite Giménez, Jaume Jordán, Javier Palanca, Jaime Rincon)....Pages 5-9
Domestic Violence Prevention System (Samuel Gallego Chimeno, Joaquín Delgado Fernández, Sergio Márquez Sánchez, Pablo Pueyo Ramón, Óscar Mauricio Salazar Ospina, Marcel Vicente Muñoz et al.)....Pages 10-14
LOWG – Intelligent Monitorization System with Custom Alerts to Avoid the Home Basics Services Related Risk (Carlos Peiró González, Jose Eduardo Reinoso Andrade, Alejandro Fuster Baggetro, Araceli Teruel Domenech)....Pages 15-19
Design Thinking for Social Challenges (Ana Gutiérrez Sanchis)....Pages 20-26
SiloMAS: A MAS for Smart Silos to Optimize Food and Water Consumption on Livestock Holdings (Sergio Marquez, Roberto Casado-Vara, Alfonso González-Briones, Javier Prieto, Juan M. Corchado)....Pages 27-37
Intelligent Livestock Feeding System by Means of Silos with IoT Technology (Alfonso González-Briones, Roberto Casado-Vara, Sergio Márquez, Javier Prieto, Juan M. Corchado)....Pages 38-48
Cooperative Algorithm to Improve Temperature Control in Recovery Unit of Healthcare Facilities (Roberto Casado-Vara, Fernando De la Prieta, Sara Rodriguez, Javier Prieto, Juan M. Corchado)....Pages 49-62
Back Matter ....Pages 63-63

Citation preview

Advances in Intelligent Systems and Computing 802

Sigeru Omatu Mohd Saberi Mohamad Paulo Novais Enrique Díaz-Plaza Sanz José Alberto García Coria Editors

Distributed Computing and Artificial Intelligence, Special Sessions II, 15th International Conference

Advances in Intelligent Systems and Computing Volume 802

Series Editor Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland Advisory Editors Nikhil R. Pal, Indian Statistical Institute, Kolkata, India Rafael Bello Perez, Faculty of Mathematics, Physics and Computing, Universidad Central de Las Villas, Santa Clara, Cuba Emilio S. Corchado, University of Salamanca, Salamanca, Spain Hani Hagras, School of Computer Science & Electronic Engineering, University of Essex, Colchester, UK László T. Kóczy, Department of Automation, Széchenyi István University, Gyor, Hungary Vladik Kreinovich, Department of Computer Science, University of Texas at El Paso, El Paso, TX, USA Chin-Teng Lin, Department of Electrical Engineering, National Chiao Tung University, Hsinchu, Taiwan Jie Lu, Faculty of Engineering and Information Technology, University of Technology Sydney, Sydney, NSW, Australia Patricia Melin, Graduate Program of Computer Science, Tijuana Institute of Technology, Tijuana, Mexico Nadia Nedjah, Department of Electronics Engineering, University of Rio de Janeiro, Rio de Janeiro, Brazil Ngoc Thanh Nguyen, Faculty of Computer Science and Management, Wrocław University of Technology, Wrocław, Poland Jun Wang, Department of Mechanical and Automation Engineering, The Chinese University of Hong Kong, Shatin, Hong Kong

The series “Advances in Intelligent Systems and Computing” contains publications on theory, applications, and design methods of Intelligent Systems and Intelligent Computing. Virtually all disciplines such as engineering, natural sciences, computer and information science, ICT, economics, business, e-commerce, environment, healthcare, life science are covered. The list of topics spans all the areas of modern intelligent systems and computing such as: computational intelligence, soft computing including neural networks, fuzzy systems, evolutionary computing and the fusion of these paradigms, social intelligence, ambient intelligence, computational neuroscience, artificial life, virtual worlds and society, cognitive science and systems, Perception and Vision, DNA and immune based systems, self-organizing and adaptive systems, e-Learning and teaching, human-centered and human-centric computing, recommender systems, intelligent control, robotics and mechatronics including human-machine teaming, knowledge-based paradigms, learning paradigms, machine ethics, intelligent data analysis, knowledge management, intelligent agents, intelligent decision making and support, intelligent network security, trust management, interactive entertainment, Web intelligence and multimedia. The publications within “Advances in Intelligent Systems and Computing” are primarily proceedings of important conferences, symposia and congresses. They cover significant recent developments in the field, both of a foundational and applicable character. An important characteristic feature of the series is the short publication time and world-wide distribution. This permits a rapid and broad dissemination of research results. ** Indexing: The books of this series are submitted to ISI Proceedings, EI-Compendex, DBLP, SCOPUS, Google Scholar and Springerlink ** More information about this series at http://www.springer.com/series/11156

Sigeru Omatu Mohd Saberi Mohamad Paulo Novais Enrique Díaz-Plaza Sanz José Alberto García Coria •

•

• •

Editors

Distributed Computing and Artificial Intelligence, Special Sessions II, 15th International Conference

123

Editors Sigeru Omatu Hiroshima University East-Hiroshima, Japan Paulo Novais Departamento de Informatica Universidade do Minho Braga, Portugal José Alberto García Coria Department of Computer Science University of Salamanca Salamanca, Spain

Mohd Saberi Mohamad Department of Software Engineering Universiti Teknologi Malaysia Johor, Malaysia Enrique Díaz-Plaza Sanz Department of Computer Science, School of Science University of Salamanca Madrid, Madrid, Spain

ISSN 2194-5357 ISSN 2194-5365 (electronic) Advances in Intelligent Systems and Computing ISBN 978-3-030-00523-8 ISBN 978-3-030-00524-5 (eBook) https://doi.org/10.1007/978-3-030-00524-5 © Springer Nature Switzerland AG 2020 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, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

The International Conference on Distributed Computing and Artificial Intelligence (DCAI) is an annual forum that brings together ideas, projects, and lessons associated with distributed computing and artificial intelligence, and their application in different areas. Artificial intelligence is changing our society. Its application in distributed environments, such as Internet, electronic commerce, environment monitoring, mobile communications, wireless devices, distributed computing, to mention only a few, is continuously increasing, becoming an element of high added value with social and economic potential, in industry, quality of life and research. These technologies are changing constantly as a result of the large research and technical effort being undertaken in both universities and businesses. This conference is a stimulating and productive forum where the scientific community can work toward future cooperation in distributed computing and artificial intelligence areas. Nowadays, it is continuing to grow and prosper in its role as one of the premier conferences devoted to the quickly changing landscape of distributed computing, artificial intelligence and the application of AI to distributed systems. The last edition at its fifteenth DCAI conference held in Toledo, Spain, from June 20 to 22, 2018, involved the exchange of ideas and trends related to distributed computing, artificial intelligence and their application in order to provide efficient solutions to real problems. 15th International Conference on Distributed Computing and Artificial Intelligence’s technical program presented both high quality and diversity, with contributions in well-established and evolving areas of research. More than 120 papers were submitted to main and special sessions tracks from over 20 different countries (Algeria, Angola, Austria, Brazil, Colombia, France, Germany, India, Italy, Japan, Netherland, Oman, Poland, Portugal, South Korea, Spain, Thailand, Tunisia, UK, and USA), representing a truly “wide area network” of research activity. Moreover, DCAI’18 Special Sessions were a very useful tool in order to complement the regular program with new or emerging topics of particular interest to the participating community. Special sessions that emphasize on multidisciplinary and transversal aspects, such as Advances on Demand Response and Renewable v

vi

Preface

Energy Sources in Smart Grids (ADRESS), AI–driven methods for Multimodal Networks and Processes Modeling (AIMPM), Social Modelling of Ambient Intelligence in Large Facilities (SMAILF), Communications, Electronics and Signal Processing (CESP), Complexity in Natural and Formal Languages (CNFL), Web and Social Media Mining (WASMM), were especially encouraged. A specific session was also organized with the winning works of the IBM Hackathon Cogs for Good held at the headquarters of the Polytechnic University of Valencia, the University of Comillas and the University of Salamanca, organized by IBM. We thank the sponsors (IBM, Indra, Viewnext, IEEE Systems Man and Cybernetics Society Spain) and the funding supporting of the project “IOTEC: Development of Technological Capacities around the Industrial Application of Internet of Things (IoT)”. 0123_IOTEC_3_E. Project financed with FEDER funds, Interreg Spain-Portugal (PocTep).” Sigeru Omatu Mohd Saberi Mohamad Paulo Novais Enrique Díaz-Plaza Jose Alberto García Coria

Contents

Image Analysis for Privacy Assessment in Social Networks . . . . . . . . . . Joaquin Taverner, Ramon Ruiz, Elena del Val, Carlos Diez, and Jose Alemany

1

Rassel: Robot Assistant for the Elderly . . . . . . . . . . . . . . . . . . . . . . . . . Maite Giménez, Jaume Jordán, Javier Palanca, and Jaime Rincon

5

Domestic Violence Prevention System . . . . . . . . . . . . . . . . . . . . . . . . . . Samuel Gallego Chimeno, Joaquín Delgado Fernández, Sergio Márquez Sánchez, Pablo Pueyo Ramón, Óscar Mauricio Salazar Ospina, Marcel Vicente Muñoz, and Aarón González Hernández

10

LOWG – Intelligent Monitorization System with Custom Alerts to Avoid the Home Basics Services Related Risk . . . . . . . . . . . . . . . . . . Carlos Peiró González, Jose Eduardo Reinoso Andrade, Alejandro Fuster Baggetro, and Araceli Teruel Domenech Design Thinking for Social Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . Ana Gutiérrez Sanchis SiloMAS: A MAS for Smart Silos to Optimize Food and Water Consumption on Livestock Holdings . . . . . . . . . . . . . . . . . . . . . . . . . . . Sergio Marquez, Roberto Casado-Vara, Alfonso González-Briones, Javier Prieto, and Juan M. Corchado Intelligent Livestock Feeding System by Means of Silos with IoT Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alfonso González-Briones, Roberto Casado-Vara, Sergio Márquez, Javier Prieto, and Juan M. Corchado

15

20

27

38

vii

viii

Contents

Cooperative Algorithm to Improve Temperature Control in Recovery Unit of Healthcare Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Roberto Casado-Vara, Fernando De la Prieta, Sara Rodriguez, Javier Prieto, and Juan M. Corchado Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

49

63

Image Analysis for Privacy Assessment in Social Networks Joaquin Taverner(B) , Ramon Ruiz, Elena del Val, Carlos Diez, and Jose Alemany Universitat Polit`ecnica de Val`encia, Camino de Vera s/n, Valencia, Spain {joataap,raruidol,edelval,cardieal,jalemany1}@dsic.upv.es

Abstract. Nowadays, the concern about privacy in online social networks has increased. However, the definition of an appropriate privacy policy might be a complex task, especially when several users are involved and have different privacy preferences. This problem usually appears when a user publishes a photo. In this paper, we propose a tool to automatically define the audience of a photo based on a trust metric. This metric uses a set of features (i.e., distance between users, number of people, emotions, etc.) obtained by the image analysis provided by IBM Cloud Visual Recognition Service. In a preliminary experiment considering 40 photos of 4 users, the results show that the proposed trust metric approximates the real trust relationships between users. We plan to integrate the tool into a real online social network. Keywords: Image analysis Trust

1

· Privacy negotiation · Social networks ·

Introduction

One of the problems that arises when sharing content (e.g., photos) in an online social network is the definition of the privacy policy [1,7,10]. This problem becomes more complex if there are more than one user involved in the shared content (e.g., various users appearing in the same photo) [4,11]. Consider the next scenario where a user A decides to publish a photo where (besides him) other people appear (users B and C). In that moment, a dilemma may arise to the user A: should I publish the photo using my privacy policy? or it would be better if I publish according to the privacy concerns of the users involved? which is the most suitable privacy policy? As each user has his own concern about privacy, it is necessary to reach an agreement. Taking this problem into consideration, this proposal aims to provide an automatic privacy policy assessment for photo sharing in social networks. In order to achieve this goal, we propose to use a trust model to define the relationships between the users based on feature extraction from published images. The proposed model consists of the following modules: (i) Image Feature Extraction, (ii) Trust Estimation, (iii) Privacy Policy Recommendation. c Springer Nature Switzerland AG 2020  S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 1–4, 2020. https://doi.org/10.1007/978-3-030-00524-5_1

2

J. Taverner et al.

Fig. 1. Examples of the results obtained by the Image Feature Extraction process.

The Image Feature Extraction module analyzes the image and detects the faces that appear in the photo. This module uses the IBM Cloud Visual Recognition service [2] to identify the users associated with the faces, and estimates the degree of trust between them. To calculate their degree of trust, the module analyzes the following features: the distance between the users that appear in the photo, the sentiment of each user, the number of people, the type of photo (e.g., closeup photography, portrait, etc.) (see Figs. 1a and b), and finally, the number of times users appear together in a photo. Based on these features extracted from the image, the Trust Estimation module estimates the degree of trust for each pair of users that appear in the photo. Each time a new photo is added to the online social network, the degree of trust of the users identified in the photo is updated. The degree of trust is used by the Privacy Policy Recommendation module to assist in the decision-making process of which is the most suitable privacy policy for publishing a photo. The module creates a personalized list of users that could see the photo based on their degree of trust. Considering the previous scenario where user A publishes a photo, the audience list that is going to see the photo is automatically created. The members of this list are a subset of the user’s A friends, user’s B friends, and user’s C friends that have a trust value with the co-owners of the photo (A, B, and C) over a trust threshold. The value of the trust threshold is established considering the most restrictive value of trust of the users involved. To test the Image Feature Extraction and the Trust Estimation modules, we design an experiment. For this experiment, we considered forty photos from four users (ten for each one). These photos were analyzed to calculate the trust values between users. For example, in Fig. 1 we can see the results of this process. The top label represents the emotion detected in the face, followed by the user

Image Analysis for Privacy Assessment

3

identifier (in this case we use A, B, C, and D). The bottom label is the total percentage of the face in the image to estimate the distance to the camera. And the label on the line indicates the distance between the users. We can observe that users A and B (Fig. 1a) have a neutral emotion and the distance between both is higher than the users C and D, that are expressing happiness. Therefore, if we only take into account this two images we can deduce that there is a higher level of confidence between users C and D than between users A and B. Then, we compare the calculated trust with the real trust values between users. The real trust values were obtained using a questionnaire previous to the experiment. The results are shown in Table 1. As can be noted from the table, the trust relationship between users is not symmetric. Asymmetry occurs because of the nature of the human relationships and differences in peoples’ perceptions, opinions, beliefs, and expectations [6,8]. In our case, the asymmetry is due to the users’ emotions shown in the photos. Table 1. Comparison between the real trust values (above) and the trust values obtained by the proposed model (below). Users A

B

A

-

0.6 0.8 0.6 0.66 0.67 0.71

C

D

B

0.6 0.7

-

C

0.6 0.6 0.68 0.65 -

D

0.6 0.0 0.72 0.0

0.4 0.0 0.23 0.0 0.8 0.71

1.0 0.84 -

Finally, to evaluate the Privacy Policy Recommendation module, we plan to integrate it in the PESEDIA social network designing an experiment with real users. With this experiment, we want to test if the audience list associated with the photo by our proposal corresponds to the users’ expected audience obtained by an initial questionnaire.

2

Conclusions

In this work, we propose a tool to assist users in the privacy decision making process when sharing a photo on a social network. This tool consists of three modules: Image Feature Extraction, Trust Estimation, and Privacy Policy Recommendation. The majority of direct social trust metrics are based on the activity in social networks [3] such as the number of comments, number of likes, or number of tags [8,9].

4

J. Taverner et al.

In this paper, we propose a metric based on image features. This metric could be complementary to other existing approaches [5] to estimate the trust value in a more informed way. We have created a recommendation module based on the proposed trust metric. This module allows the automatic definition of the audience of a publication where more than one user appears. Acknowledgements. This work is partially supported by the Spanish Government project TIN2017-89156-R, by the FPI grants BES-2015-074498 and ACIF/2017/085, and the Post-Doc scholarship with the Ref. SP20170057.

References 1. Alemany, J., del Val, E., Alberola, J., Garc´ıa-Fornes, A.: Estimation of privacy risk through centrality metrics. Future Gener. Comput. Syst. 82, 63–76 (2017) 2. Bhattacharjee, B., Boag, S., Doshi, C., Dube, P., Herta, B., Ishakian, V., Jayaram, K., Khalaf, R., Krishna, A., Li, Y.B., et al.: IBM deep learning service. IBM J. Res. Dev. 61(4), 10–11 (2017) 3. Marsh, S.P.: Formalising trust as a computational concept (1994) 4. Mester, Y., K¨ okciyan, N., Yolum, P.: Negotiating privacy constraints in online social networks. In: Koch, F., Guttmann, C., Busquets, D. (eds.) Advances in Social Computing and Multiagent Systems, pp. 112–129. Springer, Cham (2015) 5. Nepal, S., Sherchan, W., Paris, C.: STrust: a trust model for social networks. In: 2011 IEEE 10th International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom), pp. 841–846. IEEE (2011) 6. Sabater, J., Sierra, C.: Reputation and social network analysis in multi-agent systems. In: Proceedings of the First International Joint Conference on Autonomous Agents and Multiagent Systems: Part 1, pp. 475–482. ACM (2002) 7. Shehab, M., Touati, H.: Semi-supervised policy recommendation for online social networks. In: 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), pp. 360–367. IEEE (2012) 8. Sherchan, W., Nepal, S., Paris, C.: A survey of trust in social networks. ACM Comput. Surv. (CSUR) 45(4), 47 (2013) ˇ 9. Situm, M.: Analysis of algorithms for determining trust among friends on social networks. Vienna, June 2014 10. Squicciarini, A.C., Paci, F., Sundareswaran, S.: PriMa: a comprehensive approach to privacy protection in social network sites. Ann. Telecommun. - annales des t´el´ecommunications 69(1–2), 21–36 (2014) 11. Such, J.M., Porter, J., Preibusch, S., Joinson, A.: Photo privacy conflicts in social media: a large-scale empirical study. In: Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems, pp. 3821–3832. ACM (2017)

Rassel: Robot Assistant for the Elderly Maite Gim´enez(B) , Jaume Jord´ an(B) , Javier Palanca(B) , and Jaime Rincon(B) Universitat Polit`ecnica de Val`encia, Camino de Vera s/n, Valencia, Spain {mgimenez,jjordan,jpalanca,jrincon}@dsic.upv.es

Abstract. Assisting the elderly is today a necessity and a responsibility of our society. As they reach old age they need more assistance and companionship when they eventually live alone. In this context, technology, and cognitive assistants in particular, can help in the task of caring for our elderly. In this work, we have developed an assistant robot for the elderly, called Rassel, that helps to take care of senior people who is living alone, helps them to keep active by recommending custom activities and also helps with their health assistance by monitoring their vital signs and maintaining contact with relatives or emergency contacts if a medical urgency occurs. This robot has been developed and presented in the IBM Country Project ‘Cogs for Good’. Keywords: Internet of Things

1

· IBM cloud · Assistant robot · Elderly

Introduction

Nowadays, there is a significant increase in the population of elderly and dependent people that need for support and assistance. For this reason, it is necessary to improve the self-sufficiency and independence of the assisted person while guaranteeing its safety and wellness at home. The main goal of this work is the design and development of an assistant robot for the elderly. The robot will be a conversational assistant that provides recommendations for an active and satisfactory life, fall and risk situations detection, system of alerts to relatives and vital signs monitoring by means of a smart wristband (Fig. 1).

2

Rassel Services

Rassel comes from Robot ASSistant for the ELderly, and it is designed to be a conversational robot that can give company to the person who owns it. Some of the services that Rassel provides are: – – – – –

Conversational Agent Activity Recommendation Fall Detection Vital Signs Monitoring Alerts system Next, we are going to present each of the services with more detail.

c Springer Nature Switzerland AG 2020  S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 5–9, 2020. https://doi.org/10.1007/978-3-030-00524-5_2

6

M. Gim´enez et al.

Fig. 1. General environment for Rassel

2.1

Conversational Agent

An assistant robot needs to communicate with its owner in a simple way. This is even more important if the user is not used to the technology. This is why touch screens or command-line interfaces were discarded. Instead, Rassel provides a voice-managed conversational interface that is easier for older people to understand and use. To build the conversational agent we used three services from the IBM Watson Cognitive Services1 : Text To Speech. To synthesize what we want to communicate to the user in a way that he can understand (we only have to be careful with the audio level when the user is hearing impaired). This feature is performed using the TextToSpeech service of IBM Watson. Speech To Text. To introduce commands into the robot we have chosen also a voice interface. Using the IBM Watson SpeechToText service, we can translate what the user says and understand what is he asking to Rassel (e.g. a new recommendation, to call some relative, etc.) Conversational Bot. In order to make a fluid conversation with the user we have developed a full conversational context with all the topics that Rassel is able to talk about. We have used the Watson Assistant from IBM Watson to build and train a conversational bot, and finally, we linked all the conversational steps with the required service of those that are presented below.

1

http://console.bluemix.net.

Rassel: Robot Assistant for the Elderly

2.2

7

Activity Recommendation

Rassel provides a recommender system which is able to learn from the user preferences and that has a database of activities which are suitable for old people. By combining this information from the user and the activities database, the activity recommendation service runs a collaborative-filtering algorithm [1,2] that selects the activity that best suits for the user taking into account how that activity liked to other users with a similar profile. This information is stored in the cloud, since the recommender anonymously merges the preferences of all their users to improve the recommendation process and to promote serendipity. To keep an online database in the cloud Rassel uses the IBM Cloudant database service, which allowed us to move the application data closer to all the places it needs to be used. 2.3

Fall Detection

One of the most common dangers for the elderly is falling to the ground and hurting themselves. Rassel can not prevent a person from falling, but it can monitor its owner using different sensorization. In this module, we use a computer vision classifier to monitor Rassel’s owner and analyze if she/he has fallen and could not get up off the floor (in next subsection we also present an inertial monitoring unit to improve fall detection). This fall detection uses the Visual Recognition component from IBM Watson to train a model that is able to classify images to detect whether or not there is a person lying on the ground. It has been trained with a collection of images of standing people and laying people and it raises an alert when it detects a laying person. 2.4

Vital Signs Monitoring

Rassel provides communication with a wearable device as a backup element to improve its service package. This portable device incorporates an inertial measurement unit (IMU), consisting of a triaxial accelerometer and a triaxial gyroscope. The IMU is used as a fall detection unit for people by linking the data acquired with the fall detection service through image processing. This reduces the false positives provided by image processing. Simultaneously, the wearable device can acquire biosignals, such as photoplethysmography (PPG) [3], electrocardiography (ECG) [4] and skin resistance (GSR) [1]. This biosignals are sent through the Internet of Things (IoT) service offered by IBM Watson, so that they can be used for telemedicine monitoring or to predict some pathological event in the patient. Figure 2 shows the sensors used to capture the PPG, ECG and GSR signals. The wearable device has been manufactured using an M5Stack2 (Fig. 3) development board that includes an ESP-32 chip, LCD display and IMU. The ESP-32 chip has Wi-Fi and Bluetooth communication protocols that facilitates the transmission of the data obtained to IBM Watson IoT services. 2

http://www.m5stack.com/.

8

M. Gim´enez et al.

Fig. 2. Sensors

2.5

Alerts System

This module has a simple function: to send alerts to the emergency contacts when there is some event to be notified (a fall detection, an anomalous measurement on the wristband, etc.). It uses instant messaging (the Telegram service) since Rassel does not currently support LTE connectivity. 2.6

Discussion

Rassel will belong to a new generation of assistant robots for the elderly which can be deployed in the homes of the dependent elderly people in order to increase their quality of life as well as their safety and the tranquility of their relatives. The main advantages of Rassel will be the constant monitoring of vital signs of the elderly person, the conversational assistance including recommendation of activities, as well as the increased safety by the fall and risk detection that will trigger an alarm to the relatives. Furthermore, the proposed technologies to use are totally feasible and applicable currently with an acceptable cost as we show in the prototype of Rassel that we built. Finally, it is important to note that the

Fig. 3. M5Stack development board

Rassel: Robot Assistant for the Elderly

9

diffusion of Rassel as a product could be potentially high due to the increase of elderly people in our society, and it could be affordable by average private users or even subsidized by the governments or health insurance companies. In addition, the expected acceptance is also high due to the advantages of Rassel user-friendly design.

3

Conclusions

The developed assistant robot for the elderly people called Rassel has several advantages at different levels. From the social point of view, Rassel prevents loneliness of the elderly person, which is an important psychological problem, as well as it promotes communication and different activities for an active and healthy life. Regarding safety, Rassel detects falls and any risk situation, which will be communicated immediately through alerts to the relatives of the elderly person. Finally, from the health point of view, Rassel provides a continuous noninvasive vital signs monitoring of the elderly person to prevent any problem. Integrating all these features into a single assistant robot is a new approach in this field. It benefits from all the cognitive services offered by IBM Watson, which allowed us to build a robot without high computational requirements, since all the hard work is done in the cloud. As a future work, Rassel will include medical appointments follow-up, online medical services, nutrition patterns and recommendations, and medical history. Acknowledgments. This work was partially supported by MINECO/FEDER TIN2015-65515-C4-1-R project of the Spanish government.

References 1. Linden, G., Smith, B., York, J.: Amazon.com recommendations: item-to-item collaborative filtering. IEEE Internet Comput. 7(1), 76–80 (2003) 2. Sarwar, B., Karypis, G., Konstan, J., Riedl, J.: Item-based collaborative filtering recommendation algorithms. In: Proceedings of the 10th International Conference on World Wide Web, pp. 285–295. ACM (2001) 3. Lee, Y.H., Tseng, H.W., Liao, Y.D., Lin, T.W., Chen, Y.L.: Sleeping detect using wearable device by PPG, pp. 259–262 (2016) 4. Led, S., Fernandez, J., Serrano, L.: Design of a wearable device for ECG continuous monitoring using wireless technology. In: The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, vol. 4, pp. 3318–3321. IEEE (2004)

Domestic Violence Prevention System Samuel Gallego Chimeno1, Joaquín Delgado Fernández1, Sergio Márquez Sánchez1(&), Pablo Pueyo Ramón1, Óscar Mauricio Salazar Ospina2, Marcel Vicente Muñoz1, and Aarón González Hernández1 1

BISITE Digital Innovation Hub, University of Salamanca, Edificio I+D+i Universidad de Salamanca, c/Espejo S/N, 37007 Salamanca, Spain {samuelgch,joaquindf,smarquez}@usal.es, [email protected], [email protected], [email protected] 2 Departamento de Ciencias de la Computación y la Decisión, Universidad Nacional de Colombia – Sede Medellín, Medellín, Colombia [email protected] Abstract. Domestic violence is a common problem in society. This type of violence can be understood as a behaviour pattern in the form of physical and/or sexual abuse, threats, coercion, intimidation, isolation, emotional or economic abuse exercised in the field of family life against any member who forms its nucleus. Currently, numerous efforts have been made to mitigate this type of violence, on a social, legal, technological or any other level. However, this is a problem that is difficult to control due to the diversity of ways in which this pattern of behavior can be expressed and the large number of repeat offenders. In this context, it is necessary to take advantage of the benefits that technology brings to detect this type of problem early and take corrective action in time. Based on the above, this work proposes the development of a system supported by intelligent services to detect cases of violence in homes with a history of violence. The experimental results obtained from the implementation of the case study show that the incorporation of intelligent services into early domestic violence prevention systems can help to control cases of recidivism and take corrective action in advance, thus mitigating the consequences and in many cases helping to save lives. Keywords: Domestic violence
Watson
Violence prevention systems Threats
Violent action
Domestic problems




1 Introduction Domestic violence is a concept that refers as such to any type of violent action (see physical force, harassment, and/or intimidation), produced within a household by a member of the family against any other component of the household. In this regard, the system will focus on prevention, i.e. anticipation before any violence occurs, focusing on the verbal analysis of each family member’s dialogue against normal patterns of behaviour within a standard spectrum. © Springer Nature Switzerland AG 2020 S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 10–14, 2020. https://doi.org/10.1007/978-3-030-00524-5_3

Domestic Violence Prevention System

11

2 Objectives Based on the above, the aim of this paper is to offer a system of prevention of domestic violence based on the cognitive services offered by Watson [1, 2]. This system allows the generation of alarms that anticipate in case of detecting patterns of violence in daily conversations in the interior of a home, allowing the taking of decisions that correct and allow to solve such situations. In order to capture the information related to the conversations, it is proposed that a series of microphone sensors be installed in strategic places in the home or that any devices with a microphone be used (such as mobile phone, tablet, computer, etc.). In itself, the proposed system relies on Wat-son’s cognitive services [2] to develop a model for early detection of cases of domestic violence. Figure 1 presents the architecture of the proposed system, designed taking into account the reference readings included in the article [4–26], which is composed of five modules described below:

Fig. 1. Violence prevention system architecture

1. Data capture module (A). This module is in charge of constantly capturing conversations inside the home, then the conversations are encrypted and sent to the server through an API. 2. Data transformation module (B). The API receives the audio from the conversation, decrypts it and sends it to Watson’s speech to text service, which returns the text of the conversation. The text is then translated by the second Watson service used, the language translator; this process is done because the modules in the next module only process the information in English. 3. Evidence analysis module (C). This phase involves the integration of three intelligent services: (1) tone analyzer, responsible for analyzing the emotions and tones used in the conversation in order to identify negative feelings. (2) the personality insights service applies linguistic analysis and personality theory to the text obtained to infer attributes such as kindness, extroversion, emotional rank, among

12

S. G. Chimeno et al.

others. Finally, (3) the natural language understanding service analyses concepts, entities and context of the conversation, helping to identify whether the conversation is framed in domestic violence scenarios or whether it is related to other everyday aspects. 4. Pattern of violence detection module (D). This machine learning module analyzes the results obtained from the three previously defined services in order to identify their relationship to cases of domestic violence. For this purpose, the algorithm is trained with previous cases of violence, looking for the learning of patterns present in the words used during the conversation. 5. Alarm module (E). This last module is deployed when patterns of violence are identified within the analyzed conversation, alerting the people or organisms previously configured (relatives, police, neighbors, etc.) so that they can take the necessary corrective actions to solve the violence situation. Taking into consideration those people who have received, or are believed to suffer physical, mental or sexual aggressions within their family context, the system will determine if it is a real threat. In this case, the police, social services or a neighbor will be warned. With this system, the number cases of domestic violence is expected to be lower, taking into account the efficiency of the detection on the test that have already been done. The abusers will also be afraid of it due to its existence. Moreover, this system could be used as a proof to demonstrate if a person is suffering any kind of aggression if the judge supports it.

3 Conclusions In this article, a new system that uses cognitive modules using Watson technology and a series of microphonic sensors distributed in the home has been proposed, considering the existing problem of domestic violence. It offers a new perspective regarding the prevention and reduction of conflicts that are currently happening at home. The use of technology as an element of analysis and treatment of data for this problem is considered as an important advance. Usually, prevention actions are focused on social awareness.

References 1. Carrillo Calderón, M.E.: Agentes virtuales con capacidades cognitivas utilizando IBM Watson (Bachelor’s thesis) (2017) 2. Bluemix, IBM. https://console.bluemix.net/catalog/?cm_mc_uid=78357560092315264828722& cm_mc_sid_50200000=92669401526482872261&cm_mc_sid_52640000=9217590152648287 2266. Accessed 10 2018

Domestic Violence Prevention System

13

3. Fernandes, F., Gomes, L., Morais, H., Silva, M., Vale, Z., Corchado, J.M.: Dynamic energy management method with demand response interaction applied in an office building. In: de la Prieta, F., et al. (eds.) Trends in Practical Applications of Scalable Multi-Agent Systems, the PAAMS Collection. AISC, vol. 473, pp. 69–82. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-40159-1_6 4. Dang, N.C., De la Prieta, F., Corchado, J.M., Moreno, M.N.: Framework for retrieving relevant contents related to fashion from online social network data. In: Omatu, S., et al. (eds.) Trends in Practical Applications of Scalable Multi-Agent Systems, the PAAMS Collection. AISC, vol. 473, pp. 335–347. Springer, Cham (2016). https://doi.org/10.1007/ 978-3-319-40159-1_28 5. Chamoso, P., De la Prieta, F., De Paz, F., Corchado, J.M.: Swarm agent-based architecture suitable for internet of things and smartcities. In: Omatu, S., et al. (eds.) Distributed Computing and Artificial Intelligence, 12th International Conference. AISC, vol. 373, pp. 21–29. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-19638-1_3 6. Casado-Vara, R., Chamoso, P., De la Prieta, F., Prieto, J., Corchado, J.M.: Non-linear adaptive closed-loop control system for improved efficiency in IoT-blockchain management. Inf. Fusion 49, 227–239 (2019) 7. González-Briones, A., Chamoso, P., Yoe, H., Corchado, J.M.: GreenVMAS: virtual organization based platform for heating greenhouses using waste energy from power plants. Sensors 18(3), 861 (2018) 8. Casado-Vara, R., Prieto-Castrillo, F., Corchado, J.M.: A game theory approach for cooperative control to improve data quality and false data detection in WSN. Int. J. Robust Nonlinear Control 28(16), 5087–5102 (2018) 9. Morente-Molinera, J.A., Kou, G., González-Crespo, R., Corchado, J.M., Herrera-Viedma, E.: Solving multi-criteria group decision making problems under environments with a high number of alternatives using fuzzy ontologies and multi-granular linguistic modelling methods. Knowl. Based Syst. 137, 54–64 (2017) 10. Li, T., Sun, S., Bolić, M., Corchado, J.M.: Algorithm design for parallel implementation of the SMC-PHD filter. Signal Process. 119, 115–127 (2016). https://doi.org/10.1016/j.sigpro. 2015.07.013 11. Chamoso, P., Rodríguez, S., de la Prieta, F., Bajo, J.: Classification of retinal vessels using a collaborative agent-based architecture. AI Commun. 31(5), 427–444 (2018). Preprint 12. Chamoso, P., González-Briones, A., Rodríguez, S., Corchado, J.M.: Tendencies of technologies and platforms in smart cities: a state-of-the-art review. Wirel. Commun. Mob. Comput. 2018, 17 (2018) 13. Gonzalez-Briones, A., Prieto, J., De La Prieta, F., Herrera-Viedma, E., Corchado, J.M.: Energy optimization using a case-based reasoning strategy. Sensors (Basel) 18(3), 865 (2018). https://doi.org/10.3390/s18030865 14. Gonzalez-Briones, A., Chamoso, P., De La Prieta, F., Demazeau, Y., Corchado, J.M.: Agreement technologies for energy optimization at home. Sensors (Basel) 18(5), 1633 (2018). https://doi.org/10.3390/s18051633 15. Gazafroudi, A.S., Corchado, J.M., Kean, A., Soroudi, A.: Decentralized flexibility management for electric vehicles. IET Renew. Power Gener. (2019). http://ietdl.org/t/IBgIPb 16. Gazafroudi, A.S., Soares, J., Ghazvini, M.A.F., Pinto, T., Vale, Z., Corchado, J.M.: Stochastic interval-based optimal offering model for residential energy management systems by household owners. Int. J. Electr. Power Energy Syst. 105, 201–219 (2019) 17. Durik, B.O.: Organisational metamodel for large-scale multi-agent systems: first steps towards modelling organisation dynamics. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 6 (3), 17–27 (2017). ISSN: 2255-2863

14

S. G. Chimeno et al.

18. Bremer, J., Lehnhoff, S.: Decentralized coalition formation with agent-based combinatorial heuristics. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 6(3), 29–44 (2017). ISSN: 2255-2863 19. Munera, E., Poza-Lujan, J.-L., Posadas-Yagüe, J.-L., Simó-Ten, J.-E., Blanes, F.: Integrating smart resources in ROS-based systems to distribute services. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 6(1), 13–19 (2017). ISSN: 2255-2863 20. Omatu, S., Wada, T., Rodríguez, S., Chamoso, P., Corchado, J.M.: Multi-agent technology to perform odor classification. In: Ramos, C., Novais, P., Nihan, C.E., Corchado Rodríguez, J.M. (eds.) Ambient Intelligence - Software and Applications. AISC, vol. 291, pp. 241–252. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-07596-9_27 21. Román, J.A., Rodríguez, S., Corchado, J.M.: Improving intelligent systems: specialization. In: Corchado, J.M., et al. (eds.) PAAMS 2014. CCIS, vol. 430, pp. 378–385. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-07767-3_34 22. Oliver, M., Molina, J.P., Fernández-Caballero, A., González, P.: Collaborative computerassisted cognitive rehabilitation system. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 6(3), 57–74 (2017) 23. Griol, D., Molina, J.M.: Simulating heterogeneous user behaviors to interact with conversational interfaces. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 5(4), 59–69 (2016) 24. Desquesnes, G., Lozenguez, G., Doniec, A., Duviella, E.: Planning large systems with MDPs: case study of inland waterways supervision. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 5(4), 71–84 (2016) 25. Griol, D., Molina, K.: Measuring the differences between human-human and humanmachine dialogs. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J.4(2), 99–112 (2015) 26. Alvarado-Pérez, J.C., Peluffo-Ordóñez, D.H., Therón, R.: Bridging the gap between human knowledge and machine learning. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J.4(1), 54–64 (2015)

LOWG – Intelligent Monitorization System with Custom Alerts to Avoid the Home Basics Services Related Risk Carlos Peiró González(&), Jose Eduardo Reinoso Andrade(&), Alejandro Fuster Baggetro(&), and Araceli Teruel Domenech(&) Universidad Politécnica de Valencia, Valencia, Spain [email protected], [email protected], [email protected], [email protected]

Abstract. The monitoring of basic household services is increasingly accessible to everyone, however, accidents in the home and unnecessary expenditure of limited resources such as electricity, gas and water continue to occur in our society and are not being treated from a preventive and real monitoring perspective. Thanks to the large amount of IoT device available on the market today and the software tools provided by the IBM Cloud ecosystem we are able to deploy real and intelligent monitoring in a home at low cost. Keywords: IBM Cloud Data analytics


Security
Basic services
Monitoring
IoT


1 Introduction With the growing need to control everything that concerns us, in recent years have been appearing systems of all kinds to control and monitor aspects of our lives, such as expenses, meals, physical activity and tasks. So a home monitoring system should be a natural step in our society. The technology is implanted in our day to day, however in areas as daily as the saving in basic services of the home and the prevention of domestic accidents still is not, and although we do not perceive many accidents are mainly related to water damage that occur every day in a country [1]. A high percentage of current homes in Spain are not prepared for the modern and expensive existing home automation systems. Furthermore, although some companies offer us partial solutions to the aforementioned problems, in no case do they form an ecosystem that unites the three basic services (water, electricity and gas).

2 Description of Objectives and Motivation LOWG’s main motivation is to bring the latest technologies to any home in a simple and economical way. One of the realities that made this project very interesting for us is that we are one of the millions of people living in this country who have uncertainty © Springer Nature Switzerland AG 2020 S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 15–19, 2020. https://doi.org/10.1007/978-3-030-00524-5_4

16

C. P. González et al.

when it comes to receiving electricity, gas and water bills, and it is one of the main reasons why this project should become a reality, not for us, but for everyone. The objectives are clear, to create a close application with the following functionalities: • Intelligent alert system: A system that, from the self-learning of the same service, detects unusual uses of electricity, water and gas consumption, sending a notification to the user’s device, alerting of the possible escape. • Real-time control of consumption: Through the average consumption per hour of previous invoices, we calculate the average consumption of the invoice that adds up to date. • Prediction system: Forecast of the total of the invoice through the data of previous invoices. • Consumption history: Consumption history of previous invoices. • Chat for the members of the dwelling: Chat in which the inhabitants of the dwelling can be in constant communication. • Recommending system to obtain the best advice: From the learning of the user and the rest of LOWG users, advice can be issued to try to save on the consumption of your next bills and prevent accidents or malpractice. • Comparison with other users: Ranking of LOWG users based on the similarity between user invoices. As a starting point for LOWG we believe that these would be the most important functionalities to cover. All this is possible thanks to the combination of the services provided by IBM Cloud with the Big Data technologies that we will detail next.

3 Description of the Proposed Model The first step in the implementation of LOWG involves the implementation of an IoT device whose sensors are capable of measuring the three magnitudes we are concerned with: • The amount of energy consumed per second. • The flow of water circulating per second. • The volume of gas emitted per second. The device should be simple and inexpensive, as its only function will be to collect data and send it to the cloud. In other words, there is not going to be any type of transformation or analysis of data embedded in the device, but rather, as is logical, they are going to be carried out entirely in the cloud. This is due to the language flexibility, scalability, security and features offered by IBM Cloud PaaS services. On the other hand, it is necessary for the device to have access to the Internet (preferably via wifi) so that it can transmit its data. With these restrictions, we have thought that something like an arduino with a shield wifi, or simply a particle photon, fit perfectly to our case of use. The best way to transfer data from the device to the cloud is via the Internet of Things Platform service. This service allows us to easily register and connect our device. It also

LOWG – Intelligent Monitorization System with Custom Alerts

17

supports the MQTT message passing protocol (based on publish/subscribe), which is the one that is going to be used in this case to take the readings of our sensors to the cloud. The Internet of Things Platform is not only a data entry gateway, but also allows us to do a real-time analysis of the data [3–12]. This analysis is mainly used for issuing alerts. By defining rules, the Internet of Things Platform service itself is able to detect the risks and produce the alerts that LOWG users will receive on their mobile phones. Of course, the rules are not static. They are initialized to common default values, but they change over time. The goal is for LOWG to learn from the user and offer personalized alerts. Thanks to the batch analysis that is discussed later, the system will be able to detect the user’s consumption patterns. By means of these patterns, the rules are defined/modified in order to detect anomalies with greater precision. As an example of the usefulness of this functionality, we can imagine that for a user who usually showers at 5 AM it is not abnormal to detect an important flow of water at that time, while it could be for another who gets up at 7 AM. The captured data is continuously transferred to a cluster for a batch analysis much more exhaustive and deep than the one carried out in real time. For this we use the Analytics Engine service, a powerful platform for data analysis. Analytics Engine deploys a cluster on which to use the well-known big data tools open source Hadoop and Spark. The cluster is very customizable: we can scale it horizontally and install the libraries and open source packages we need. Another point for Analytics Engine is that it separates the computation from the storage, so that a failure in the first one will never corrupt the second one. In short, we have chosen analytics engine because it is a simple, flexible, efficient and safe way to perform the analyses required by LOWG. These analyses, as already mentioned, consist fundamentally in the detection of consumption patterns of basic resources. Our objective is to know the user’s routine to the point of being able to predict with good precision his consumption of light, water and gas for each hour of the week. To do this, we can train regression models (such as linear regression, SVMs, etc.) with the historical data collected. One of the key points here is that the models generated are dynamic, that is, LOWG never stops learning from the user. As new data arrives, LOWG incorporates them to its models, being more and more precise in its estimations (and therefore in its alerts) and having tolerance to the routine changes that take place (vacations, illness, etc.). As already mentioned, one of the clear uses of batch analysis is the production of personalized alerts, however, it is not the only one. The patterns obtained can be easily harnessed to create personalized tips. Knowing the peak hours of consumption, recommendations can be made. Some examples would be: “Putting the washing machine on at night would be cheaper” or “Turning on the lights in the morning is an unnecessary expense”. On the other hand, our data can also be used to make comparisons between similar users in terms of average consumption of the three resources. For example: “The 20 users most similar to you consume on average 10% less electricity than you”. Something like this may encourage you to try to save more. Of course the frequency of tips and comparisons will be fully editable by the user. The aim is not to annoy, but to help. A differentiating factor in our application is the social one. LOWG takes into account that many homes are inhabited by more than one individual and that is why it

18

C. P. González et al.

has the chat. For the same LOWG unit you can register as many users as you want. In this way, we create a chat accessible through the mobile app that includes all members of the household. It is through this chat that LOWG sends the alerts (a gas leak is something that concerns all members of a house). To include this functionality we use SDK for NodeJS, which allows us to create our chat service with all the advantages and guarantees of the cloud (scalability, security, etc.).

4 Results Since the implementation has not been possible due to lack of time, it is not possible to analyze the results obtained. However, the results will be expected to meet the needs of all users, i.e. that the calculation of the forecast of expenditure is as approximate as possible, that the alert notifications of unusual use are correctly read and not false positives, that the recommendations for the use of resources are in line with reality. In any case, the implementation of LOWG should not be excessively complex or problematic, especially using IBM Cloud tools [13].

5 Conclusions The size and scope of this project has increased as the LOWG team has come together. Thinking about the obstacles that the project could have, we are presented with real but assumable difficulties. We need an initial knowledge, on the part of an expert, to be able to detect and prevent accidents in a standard home, also there are impediments in the installation of our sensors since not all the homes have a standard installation of water, light and gas, we need to know the casuistry that exists in this subject. In addition, for a real large-scale implementation, we would have to carefully study the costs of the physical devices and IBM cloud platforms used. The possibilities of monetizing this project are very wide, charge for use, charge energy companies for personalized recommendations to their customers of offers and tariffs, data to generate population statistics regarding energy consumption of a population, charge insurers for subscribing to this application (less accidents in the home equals less compensation to pay or repair). The possibilities are wide but the sector is complicated, this is something we need to study more thoroughly. An important aspect to highlight is the possibility of deploying a complex application, as in this case, with a low investment since we have IaaS and PaaS services on the IBM cloud platform. The idea of deploying our services on AWS was also raised, but the amount of time we would have to invest to develop modules similar to those developed on IBM cloud would be much greater [14]. The technological evolution is a fact that advances so fast that we are not even able, in most cases, to stop to think of new ideas, since the reality is that everything is implemented. However, we see a very big void, and a lot of lack of information about the advances in home economics and security that, by directly affecting us, is very interesting, and leads us to try to investigate and develop a service that solves these problems.

LOWG – Intelligent Monitorization System with Custom Alerts

19

References 1. Report on the Spanish home insurance market. http://asiturfocus.es/estudios-de-mercado/ informe-sobre-el-mercado-espanol-de-seguros-de-hogar/ 2. Corchado, J.M., Bajo, J., de Paz, Y., Tapia, D.: Intelligent environment for monitoring Alzheimer patients, agent technology for health care. Decis. Support Syst. 34(2), 382–396 (2008). ISSN: 0167-9236 3. González-Briones, A., Chamoso, P., Yoe, H., Corchado, J.M.: GreenVMAS: virtual organization based platform for heating greenhouses using waste energy from power plants. Sensors 18(3), 861 (2018) 4. Chamoso, P., González-Briones, A., Rodríguez, S., Corchado, J.M.: Tendencies of technologies and platforms in smart cities: a state-of-the-art review. Wirel. Commun. Mob. Comput. 2018, 17 (2018) 5. Gonzalez-Briones, A., Prieto, J., De La Prieta, F., Herrera-Viedma, E., Corchado, J.M.: Energy optimization using a case-based reasoning strategy. Sensors (Basel) 18(3), 865 (2018). https://doi.org/10.3390/s18030865 6. Khan, M.A., Freitag, F.: Sparks in the fog: social and economic mechanisms as enablers for community network clouds. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 3(1), 1–12 (2014). ISSN: 2255-2863 7. Gonzalez-Briones, A., Chamoso, P., De La Prieta, F., Demazeau, Y., Corchado, J.M.: Agreement technologies for energy optimization at home. Sensors (Basel) 18(5), 1633 (2018). https://doi.org/10.3390/s18051633 8. Gazafroudi, A.S., Corchado, J.M., Kean, A., Soroudi, A.: Decentralized flexibility management for electric vehicles. IET Renew. Power Gener. (2019). http://ietdl.org/t/IBgIPb 9. Gazafroudi, A.S., Prieto-Castrillo, F., Pinto, T., Corchado, J.M.: Organization-based multiagent system of local electricity market: bottom-up approach. In: De la Prieta, F., et al. (eds.) PAAMS 2017. AISC, vol. 619, pp. 281–283. Springer, Cham (2018). https://doi.org/10. 1007/978-3-319-61578-3_38 10. Omatu, S., Wada, T., Rodríguez, S., Chamoso, P., Corchado, J.M.: Multi-agent technology to perform odor classification. In: Ramos, C., Novais, P., Nihan, C.E., Corchado Rodríguez, J.M. (eds.) Ambient Intelligence - Software and Applications. AISC, vol. 291, pp. 241–252. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-07596-9_27 11. Román, J.A., Rodríguez, S., Corchado, J.M.: Improving intelligent systems: specialization. In: Corchado, J.M., et al. (eds.) PAAMS 2014. CCIS, vol. 430, pp. 378–385. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-07767-3_34 12. Tapia, D.I., et al.: Evaluating the n-Core Polaris real-time locating system in an indoor environment. In: Rodríguez, J., Pérez, J., Golinska, P., Giroux, S., Corchuelo, R. (eds.) Trends in Practical Applications of Agents and Multiagent Systems. AISC, vol. 157. Springer, Heidelberg (2012). https://doi.org/10.1007/978-3-642-28795-4_4 13. IBM Cloud Documentation, IBM. https://console.bluemix.net/developer/watson/documentation 14. Difference Between IBM Cloud, Amazon AWS & Microsoft Azure. https://developer.ibm. com/answers/questions/427732/difference-between-ibm-cloud-amazon-aws-microsoft/

Design Thinking for Social Challenges Ana Gutiérrez Sanchis(&) Comillas Pontifical University, Calle Alberto Aguilera, 28015 Madrid, Spain [email protected]

Abstract. This research detected some social needs in the current society from three universities in Spain to help our population with the use of technology. This project had three phases: The Design Thinking, the Hackathons and the Technical implementation for each selected project. When our market research and the specific social needs are found, the IT systems can be created focus on the final customers. Keywords: Design Thinking


Social challenges
Assistant
Watson

1 Introduction The COGS FOR GOOD project by IBM Country Project pretended to identify some social problems which can be solved using technology. It was pretended to make a collaborative work between volunteer students from Comillas Pontifical University, University of Salamanca and Polytechnic University of Valencia. The challenge was to connect the academic knowledge with the real life of our society to apply the technological progresses with Watson to solve social problems for giving sense to this type of developments. The Cathedra in Economic and Business Ethics from Comillas Pontifical University had the responsibility to define which social challenges would be propose for the implementation and which requirements should be developed by our partners. The companies and departments of Universities can be looked at the Table 1. Finally, we got some ideas to show our partners in the hackathons to get a technological improvement.

2 Objectives and Methodological Process of Design Thinking IBM and Comillas Pontifical University had the next objectives and goals: 1. To identify some social problems that can be solved with technology. 2. To get a collaborative work between the companies, universities and society. 3. To apply the technics for getting solutions to social problems and give sense to the current technological work.

© Springer Nature Switzerland AG 2020 S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 20–26, 2020. https://doi.org/10.1007/978-3-030-00524-5_5

Design Thinking for Social Challenges

21

Table 1. Companies and Universities partners for IBM Country Project “Cogs for Good” Companies/Universities IBM Spain VIEWNEXT (An IBM Subsidiary) Comillas Pontifical University University of Salamanca Polytechnic University of Valencia

Project and Departments IBM Country Project “Cogs for Good”

Cathedra in Economics and Business Ethics BISITE Research Group Computer Technology-Artificial Intelligence Group (GTI-IA)

The first steps about the organization and the methodological process about the design thinking was implemented by the Cathedra in Economics and Business Ethic (Comillas Pontifical University). They organized the first Design Thinking session working in collaboration with IBM. Initially, it was thought just one session to decide the best ideas before the hackathons, but we got many ideas that connected with others, so it was mandatory to call for the second and third sessions before showing it to our partners for implementation (look at Fig. 1).

1st Session: Business Model CANVAS

2nd Session: Challenges’ delimitaƟon

3rd Session: Preparing the hackathons

Hackathon in USAL (Salamanca) Hackathon in UPV (Valencia)

Fig. 1. Methodological process of Design Thinking before the Hackathons

The 1st Session of Design Thinking was opened with a call for any student, worker or professor from Comillas Pontifical University. IBM Spain made a presentation of Watson System with the objective to apply it to our social challenges. After this short exposition, the participants were divided into 8 groups of 5–6 people. We used the brainstorming to collect the more ideas possible in each group. We grouped the more similar ideas and we reject those that had less common points working in teams within each group. Finally, we got one idea for each group using the Business Model CANVAS in which was shown: 1. 2. 3. 4. 5.

Market segments Value proposals Channels Relationships with clients Sources of income

22

6. 7. 8. 9.

A. G. Sanchis

Key Activities Key Resources Key Alliances Cost structure

Each group decided a speaker to present their business idea for 5 min and, at the end of the session, every participant voted honestly with 3 points the best ideas to developed as social challenges. After the 1st session of Design Thinking, the jury of Comillas Pontifical University decided to call for a second session to some groups, because some challenges were compatible with each other. During the Second Session of Design Thinking “Challenges’ delimitation”, we were talking to match up the ideas until getting the three final challenges proposed. Look at the next table (Table 2): Table 2. Final challenges proposed Challenge A Challenge B Challenge C Virtual Healthcare Assistant “Dr. Watson” Home Donations Solitude Waste Bracelet

Initially, it was planned to show just one challenge per University. The challenge C was discarded for two reasons: the students involved with this challenge didn’t come to the second session and The Next IBS gave up its collaboration. At the end, we decided that the two Social Challenges (A and B) would be exposed to both universities (USAL and UPV) to offer flexibility to the candidates during their implementation. During the Third Session of Design Thinking “Preparing the hackathons”, we continue working in-depth with our IBM colleague-professor in how to show the technical requirements to be asked for Hackathons.

3 The Social Challenges Exposition The Challenge A was the Virtual Healthcare Assistant (for Refugees or people at risk of exclusion). It included a request to create a resource management with donations and volunteers, like a setting-up of a donation market place between the supply and the demand. We were thinking about to create a Virtual ID for people received in refugees’ camps who doesn’t have anything at the beginning (i.e. identity, medical historial, education certificates). It could be a database managed by a third party to be legal with this vulnerable collective and data. Other idea to implement was to include virtual interaction (with friends, family, etc.) to the medical historical registers for analyzing the psychological effects and using them by health professionals to their specific treatments (i.e. bullying, domestic or gender violence, moving). We are increasing the

Design Thinking for Social Challenges

23

use of technology in our relations and it is very important to take into account for health aspects. Thinking about medicine, we asked for a virtual consultation with a robot that could give back a first diagnosis through the analysis of symptoms with basic elements of communication. Also, we thought about a Biomedicine Portable, like a medical laboratory in connection to Watson. It should be very useful to detect specific diseases that could be highly contagious in territories with a high concentration of population and to prevent the contagion. This virtual healthcare assistant could have a connexion with the social resources management. We asked for apps to help people to be aware which are the sources they have connecting all the private contracts and public services, delimitating to each demand personalized. Also, to create an internal management about activities or social resources that the person needs in each specific case (Fig. 2).

VIRTUAL HEALTHCARE ASSISTANT BIOMEDICINE Laboratory Portable Fix

ONLINE MEDICAL DOCTOR Advisor First Diagnostic

VIRTUAL ID

MANAGMENT OF SOCIAL RESOURCES

Medical History Rights

SOCIAL ASSISTANT ONLINE

Apps Internal managment

SOCIAL MARKETPLACE Customers Providers

Education background Other characteristics

Fig. 2. Virtual healthcare assistant

The Challenge B proposed was a Home Assistant, that contains five main services: security, economy, social, education and health (look at Fig. 3). About security was mentioned to ask an apparatus to have a personal and personalized emergency system outdoor (i.e. Household alarm in cases of intrusion, fire, flood, gas) and a security system with detection and control (i.e. children, elderly people, dependents, generation of asynchronous alerts in violations, robberies or emergencies). It could be automatically or manually. The notifications could be communicated to a private, public or pseudo-public community and the stakeholders for alerts would be the emergency services, community, users, insurers or security companies. Finally, this home assistant could prepare a quick analysis of possible intrusions in your social networks’ information.

24

A. G. Sanchis

Security

Health

Economy HOME ASSISTANT

Educatio n

Social

Fig. 3. Home assistant

The second main service that this home assistant should offer is about economy. It was asked for something to get an efficient use of basic resources (water, electricity and gas), so an efficiency of domotics at home. Also, it was required a platform with a time optimization to use the home machines (i.e. variation in the prices of electricity throughout the day, assistance in the use of light) to get an efficient world in terms of environment and to save money in terms of microeconomies at home. As well, a food management and a counselor for domestic economy with stakeholders (banks, department stores, shopping centers, schools, community, taxes) to pro-pose you some options for improving your lifestyle, economy or whatever you are interested. And, also, the home assistant could remind you when you are able for economic donations, food, clothing or medicines that you are not using. The third service that this home assistant could offer is to control people in need of social assistance to make intervention on loneliness, to communicate with these people, to control that Social Services have done and to offer an online Social Assistant. The fourth service is about health, medication control alarm, medical supplies, appointment tracking, constant control, nutrition, online medical assistance, first diagnosis and records. Finally, this home assistant could help with education of people in need using it for MOOCS, courses, social insertion and learning new technologies required in the society. The selection criterial for the Hackathons previously established by the jury to evaluate the proposal challenges were (Fig. 4): 1. It conforms to the requirements requested by IBM Spain and Comillas Pontifical University. 2. It involves the use of the IBM Cloud platform. 3. It was realistic and achievable.

Fig. 4. Home assistant detailed

Design Thinking for Social Challenges

25

26

A. G. Sanchis

4 Conclusions Four proposals were selected from the Hackathons to present during the session IBM Hackathon Cogs for Good at 15th International Conference on Distributed Computing and Artificial Intelligence (DCAI 2018) celebrated in Toledo. The Hackathon taken at Polytechnical University of Valencia (UPV) had more students that the Hackathon at BISITE Research Group (University of Salamanca), so just one proposal was chosen from USAL and tree from UPV. These selected projects use some systems as IBM Cloud, Machine learning module, Natural Language understanding or Speech to text. You are invited to read in-depth how they work along their following articles. 1. “Image Analysis for Privacy Assessment in Social Networks” by Joaquin Taverner, Ramon Ruiz, Elena del Val, Carlos Diez, and Jose Alemany. 2. “Rassel: Robot Assistant for the Elderly” by Maite Giménez, Jaume Jordánn, Javier Palanca, and Jaime Rincón 3. “LOWG – Intelligent monitorization system with custom alerts to avoid the home basics services related risks” by Carlos Peiro Gonzalez, Jose Eduardo Reinoso Andrade, Alejandro Fuster Baggetro, Araceli Teruel Domenech 4. “Domestic violence prevention system” by Samuel Gallego Chimeno, Joaquín Delgado Fernández, Sergio Márquez Sánchez, Pablo Pueyo Ramón, Óscar Mauricio Salazar Ospina, Marcel Vicente Muñoz, Aarón González Hernández. It is important to conclude that this research was an enriching experience for everybody because we work with a multidisciplinary view. The Design thinking involved students from MBA’s, Mathematics, Engineers and Social Sciences. It influences to catch some new ideas with different views from universities (using companies resources as IBM) towards IT sector related to a future market business to help with social needs.

SiloMAS: A MAS for Smart Silos to Optimize Food and Water Consumption on Livestock Holdings Sergio Marquez1 , Roberto Casado-Vara1 , Alfonso Gonz´ alez-Briones1,2(B) , Javier Prieto1 , and Juan M. Corchado1,2,3,4 1

BISITE Research Group, University of Salamanca, Edificio I+D+i, Calle Espejo 2, 37007 Salamanca, Spain [email protected] 2 Air Institute, IoT Digital Innovation Hub, Carbajosa de la Sagrada, 37188 Salamanca, Spain 3 Department of Electronics, Information and Communication, Faculty of Engineering, Osaka Institute of Technology, Osaka 535-8585, Japan 4 Pusat Komputeran dan Informatik, Universiti Malaysia Kelantan, Karung Berkunci 36, Pengkaan Chepa, 16100 Kota Bharu, Kelantan, Malaysia

Abstract. A few years ago, hitting the silo container from the outside was the only way of knowing whether it had to be refilled with feed or water. However, current advances make it possible to develop more evolved mechanisms that not only allow the farmer to know if it is necessary to fill the silo with feed but give a precise estimate of the quantity of feed or water remaining in the silo and information on other parameters that help control the quality of the feed. To this end, it is necessary to design a device that will be placed on the inside of the silo and will detect if there is feed and how much of it by means of a sensor with ultrasonic technology. The prototype includes several motion engines which perform a complete sweep for the calculation of volume; this is important as each type of feed has a different density. In addition, the development of such a system will make it possible to optimize the delivery of feed to livestock holdings through route planning for the truck, for example, in cases where two nearby farms are short of supply. For this purpose, we have developed a system that incorporates an IoT device with a laser for calculating the volume of feed inside a silo. In addition, this system includes a series of sensors that can monitor temperature and humidity. Thus, the owners obtain more information from which they can draw conclusions about the conservation of the feed and about its general exploitation. Furthermore, it is possible to understand to what extent the cold and humidity affect animals and their consumption of the feed. This research work describes the evaluation of the developed prototype in several independent silos on the Hermi Group’s farm (Salamanca) and outlines the obtained results. Keywords: Sensor-based monitoring Ambiental intelligent · Agents

· Smart silo · IoT ·

c Springer Nature Switzerland AG 2020  S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 27–37, 2020. https://doi.org/10.1007/978-3-030-00524-5_6

28

1

S. Marquez et al.

Introduction

Current rabbit farms are completely automated. There are several warehouses where the rabbits are bred and there is usually a shed where first-time mothers are kept. These farms have many fattening stations for rabbits, and extra stations for fattening if necessary. Currently, rabbit farms look after the welfare of the animals and for this reason, do not usually place more than 7 rabbits per cage. The nests of the rabbits are made of straw and wood shavings. The industrial building, where the rabbits are bred, is ventilated naturally but also has ventilation systems which regulate the temperature of the environment during the calving season. As far as feeding is concerned, different types of feed are usually used depending on the rabbit’s growth stage. The silos on rabbit farms are made of fiberglass or aluminum and they normally experience humidity and temperature problems. Moreover, it is not possible to measure, with precision, the amount of feed left in the silo and as a result, it is difficult to tell when the silo should be refilled. Since the automatic coordination of these tasks is a very complex problem, this work proposes the use of a MAS for this purpose [9]. The literature in the area of MASs demonstrates that these systems are capable of solving complex problems in different fields. In this case, the following tasks must be coordinated: (1) Calculation of the amount of grain in the silo. It is important to know the amount of grain in the silo with a high degree of precision. Otherwise, the farm is at risk of running out of grain and leaving the animals hungry. (2) Temperature and humidity. If these two parameters are not controlled inside the silo, the grain can get spoiled or lose quality. Poor grain quality can lead animals to illness and thus, loss of rabbit production on the farm. The proposed system measures the level of feed in the silo. In addition, the measurements can be viewed on an LED display, installed on the outside of the silo. Optionally, the measurements can be transmitted to a Cloud for their visualization and management. Furthermore, data are collected through temperature and humidity sensors, keeping the farmer updated on the conservation conditions of the grain that is inside. All the tasks described above are coordinated with a MAS. The MAS has been developed using the JADE framework. The benefit of using a MAS is that it adapts to changes in the environment (e.g., the number of silos increases). In this paper, we address the problem of intelligent silo control. We focus on monitoring and controlling the amount of grain in the silo. We also monitor and control the humidity and temperature of the interior of the silo to maintain the quality of the grain. As a result, the quality of the grain given to the rabbits is good and a constant volume of feed is maintained in the silo. The main contribution of this paper is the increased efficiency in the monitoring and the control of the quantity and quality of the grain stored in the silos. The rest of the article is structured as follows: Sect. 2 describes related state of the art proposals. Section 3 provides a full description of the proposed system proposed. Section 4 details the case study and discusses the results. Finally, Sect. 5 outlines the conclusions drawn from this research.

SiloMAS Platform

2

29

Related Work

Multi-Agent systems (MASs) are autonomous computational entities with the capacity to execute tasks and achieve their goals with minimum human supervision or none. Gonz´ alez-Briones et al. defines the characteristics that agents should have in order to work effectively: (a) reactivity [8]: they respond immediately to changes perceived in their environment; (b) distribution of tasks: each agent has well-defined functionalities and identifies the problems that must be solved; (c) proactivity: agents take the initiative to solve problems; (d) cooperation and coordination: they perform tasks by means of exchange of messages with other agents through a common language; (e) autonomy: agents do not require the direct intervention of human beings to operate; (f) deliberation: each agent has the capacity to carry out internal reasoning processes which allow them to make decisions; (g) mobility: they can move from one node to another through the network; (h) adaptation: to improve their performance, agents adapt their behaviour to the changes in the environment, and finally (i) parallelism: the system can optimize its performance if different tasks are executed simultaneously by the agents [15]. In this way, multi-agent systems (MAS) are formed from a set of intelligent agents that work together and interact in a coordinated computational environment to solve specific and highly complex problems. According to Gonz´ alez-Briones et al. MASs offer ease of information acquisition and highly distributed processing [10,14]. Thus, this paradigm offers new ways of analyzing, designing and implementing complex software systems [11]. Additionally, MASs have been used in the development of recommendation systems [12,13] whose search results take into account the particular cognitive needs and characteristics of users. In the case of e-commerce [15], these systems offer recommendations on financial resources according to the characteristics, tastes and needs of the customers [4,24–31]. Monitoring the approximate volume of grain in silos is an important issue for farmers and industries. In addition, monitoring and controlling the humidity and temperature of the silo is critical to maintaining the quality of the grain stored in the silo [27]. The technique of measuring the level of grain through sensors placed in the silos is proposed in [18] to be able to monitor the level of grain inside the silo. A sensor based fluid level measurement system has been applied in [21] where the sensor design is based on a passive circuit comprising the capacitance plates. [19] uses ultrasonic lamb waves to detect the presence of liquid as the wave characteristic changes into liquid contact. Applying technology of optical fibers in [20] which makes use of the measurement in the variations in amplitude of the distance of the liquid. Instead, [22] monitors the quality of the grain by installing sensors, at different levels on the inside of the silo [22]. Monitoring the amount and quality of grain is important in the farming industry. In this paper, we propose the use of a MAS to coordinate those tasks. The state of the art demonstrates that MASs are very effective in coordinating and solving complex problems. Therefore, in this case study a MAS will be used to coordinate the sensor that measures the amount of grain in the silo with those that measure the temperature and humidity level [32–37].

30

3

S. Marquez et al.

SiloMAS Platform

This section describes the technical aspects of the multi-agent system developed in this work. It is responsible for calculating, monitoring, transmitting data and of notifying the farmer of the volume of grain in the silo. 3.1

Silo Prototype

The system is designed to measure the surface of granular solid materials in silos and other storage tanks. It is capable of performing this task automatically and thanks to the use of ultrasonic technology, specifically SONAR technology, there is no need for direct contact with the grain, we can take a point in a vertical silo with which internal volume can be calculated. The importance of this prototype lies in its ability to measure the volume of grain with great precision, an essential characteristic for effective feed management. The lack of such a system on the farm makes it very difficult to manage grain supplies efficiently. Inefficiency puts farms at risk of grain shortages causing animals to suffer from hunger. Poor management may also result in food wastage due to surplus stock, leading to economic loss for farmers and industries. It is therefore necessary to predict factors that are related to stock management, including grain consumption, temperature, humidity, animal age, animal weight, weight evolution and the probability of disease. In the wake of this problem, we propose an IoT (Internet of Things) device that can timely refill the silo with grain. The system allows take the measure of a point through SONAR technology using as model XL-MaxSonar-WRL MB7066, which is capable of taking highly precise measurements, up to 10 m deep. In addition, a temperature and humidity sensor SHT20 I2C Temperature & Humidity Sensor (Waterproof Probe) collects data to predict the quality of the grain found inside the silo (see Fig. 1). Once the measures have been transmitted to the cloud, calculations are performed to determine the internal volume and place orders in a much more optimal way. The system is placed inside the silo to the top cover of it by two screws leaving a watertight box IP67 (see Figs. 2 and 3). It is considered the possibility of integrating the system into elements other than the cover itself where the position is adjusted by means of transversal screws that pass through holes in the silo, and self-locking nuts, so that the angle can be suitable, and therefore

Fig. 1. Diagram of elements of the device. Modules and communications.

SiloMAS Platform

(a) Perspective of the IoT device prototype

31

(b) Profile of the IoT device prototype

Fig. 2. View of the IoT device protototype in which the sensors, casing and adapter are appreciated.

the position of the sensor can be completely fixed. Thanks to SONAR’s long measuring range it is possible to reach the bottom of high silos or measure through narrow openings in feed chutes and hoppers as long as the required accuracy is around 5%.

Fig. 3. Features and services offered through the website.

32

S. Marquez et al.

The Wi-fi module is connected to the router and this generates a Wi-fi network to which the microcontroller is connected. In this way, the system is connected to the Internet and can be controlled by means of an instruction protocol designed specifically for it, sent by the MQTT protocol through the Internet. In addition, the router is connected to the network. The system can be configured to take measurements periodically or instantaneously, depending on the preferences of the user (see Fig. 3). The novelty of this system lies in its characteristics. Apart of task automation and coordination and highly precise volume calculation, the elements used in the design of the prototype are dust-repellent. The shape of SONAR, for example, prevents dirt from adhering and the casing of the humidity and temperature prevents it from getting covered with dust. These features ensure that dirt does not affect the readings taken by the sensors, and little maintenance is required. 3.2

Software Agent Platform

It is necessary that the cloud to provide a highly dynamic SiloMAS platform, with the ability to adapt itself to changes at the time of execution. This makes it possible to determine the behaviour of the Sonar agent according to the objectives it must achieve (quantity of waste to be sold, participation percentage, etc.), taking into account the objectives of other agents and the changes that may arise in the environment. The core of the architecture is a group of deliberative agents that act as controllers and administrators of all applications and services. The functionalities of the agents are not within their structure, instead they are modeled as services. This approach offers greater error resilience and greater flexibility in the behavior of agents at execution time. The use of a multi-agent system, as discussed in the state of the art, allows us to adapt to a changing context, such as the increase or change in the location of silos. A MAS allows to model a highly dynamic platform that uses self-adaptation capabilities at execution time [1,7]. The JADE framework has been chosen for the development of SiloMAS. This Java framework provides mechanisms for the incorporation of agents that manage security aspects, ensuring that the collected data are really those that are being transmitted and analysed [1–3]. The architecture has been developed for the incorporation of multiple heterogeneous sensors, so that the functionalities can be extended in the future through the incorporation of new services [4–6]. The system is specifically designed to analyse user behaviour, sensor data, third party data and decision making. The main goal of this project allow us to use a multi-agent approach and this architecture will manage information from wireless sensor networks (e.g., Wi-Fi, Bluetooth, ZigBee, GPS, GPRS, etc.) for knowledge discovery and decision making (Big Data Analytics) on farms. Sensor networks belong to the category of complex, distributed, interconnected and rapidly evolving systems. Multi-agent systems have been identified as one of the most appropriate technologies for the deployment of sensor networks, due to their robustness, autonomy and ability to provide formalisms, algorithms and methodologies that meet the challenging needs

SiloMAS Platform

33

of sensor networks, specifically: adaptability, decentralized control, large scale, information uncertainty, resource limitation and physical distribution [16,17]. Being the adopted technologies the adequate ones for the development of the prototype raised in this work of investigation.

4

Case Study

To verify the effectiveness of the proposed system in optimizing feed and water consumption through calculation of silo content, a monitoring and evaluation case study has been designed and conducted. A series of silos have been used for this purpose. The analysis of results shows the impact of this system on the agro-livestock farm. 4.1

Experimental Set-Up

The IoT device was fitted at the lib of the silo at the top such that the IoT device faces the grains. The model of the silo Aviporc 230/5, with diameter 2.3 m, height 8.15 and capacity 21.07 m3 (approximately 12 tons of grain). The device was tested over 18 days where the silo was filled with 6 tons of corn (approx. 10,5 m3 )(see Fig. 4). The values were collected through XL-MaxSonar-WRL MB7066 and SHT20 I2C Temperature & Humidity Sensor (Waterproof Probe). Finally the server application connects to the MySQL database and stores all the information collected by the sensors: Temperature, Humidity and volume of the grain in the silo.

Fig. 4. Shows the location of the farm on which the device was installed.

34

4.2

S. Marquez et al.

Results

Figure (see Fig. 5) reflects the suitability of the proposed system since the error in weight estimation is below 6%. Leading to the conclusion that the proposed smart feeding system reduces waste and animal mortality.

Fig. 5. The figure shows the measurements obtained from the silo over the 18-day interval, showing grain, temperature and humidity in descending order.

5

Conclusions and Future Work

In this work, a new MAS approach has been presented for smart silo monitoring and management on farms by means of sensorization and IoT devices. The prototype design consists in several sensors and actuators. Among its functionalities, its main features is the ability to calculate the volume of food and obtain temperature and humidity data. This, in turn, makes it possible to estimate the quality of the feed stored in the silo, send alerts and make efficient decision in relation to livestock care and stock management. The system can measure the approximate amount of feed by means of the approximate calculation regarding the level in which the feed is located with respect to the upper SONAR. The design and components allow its installation to be easy to perform and also adapted to contact with dust particles and dirt. The foreseeable placement in the loading lid makes it easily accessible and using watertight boxes as packaging is considered as a robust system compared to other products that we find in the market. It is customizable in terms of the specific characteristics of each measurement and each silo in particular and it is cheaper with respect to other products with similar characteristics. As future work will be evaluated the possibility of integrating blockchain technology to add the benefits of this technology to this development [23,24].

SiloMAS Platform

35

Acknowledgements. This work has been partially supported by the Agreement between the Agricultural Technology Institute of Castile and Le´ on, Hermi Gesti´ on, S.L., and the University of Salamanca to conduct research activities for the development of a farm 4.0 model in the rabbit meat production sector.

References 1. Gonz´ alez-Briones, A., Prieto, J., De La Prieta, F., Herrera-Viedma, E., Corchado, J.M.: Energy optimization using a case-based reasoning strategy. Sensors 18(3), 865 (2018) 2. Gonz´ alez-Briones, A., Chamoso, P., De La Prieta, F., Demazeau, Y., Corchado, J.M.: Agreement technologies for energy optimization at home. Sensors 18(5), 1633 (2018) 3. Gonz´ alez-Briones, A., Castellanos-Garz´ on, J.A., Mezquita Mart´ın, Y., Prieto, J., Corchado, J.M.: A framework for knowledge discovery from wireless sensor networks in rural environments: a crop irrigation systems case study. Wirel. Commun. Mob. Comput. 2018 (2018) 4. Gonz´ alez-Briones, A., Chamoso, P., Yoe, H., Corchado, J.M.: GreenVMAS: virtual organization based platform for heating greenhouses using waste energy from power plants. Sensors 18(3), 861 (2018) 5. Gonz´ alez-Briones, A., Prieto, J., Corchado, J.M., Demazeau, Y.: EnerVMAS: virtual agent organizations to optimize energy consumption using intelligent temperature calibration. In: International Conference on Hybrid Artificial Intelligence Systems, pp. 387–398. Springer, Cham (2018) 6. Chamoso, P., Gonz´ alez-Briones, A., Rodr´ıguez, S., Corchado, J.M.: Tendencies of technologies and platforms in smart cities: a state-of-the-art review. Wirel. Commun. Mob. Comput. 2018 (2018) 7. Gonz´ alez-Briones, A., Chamoso, P., Prieto, J., Corchado, J.M., Yoe, H.: Reuse of wasted thermal energy in power plants for agricultural crops by means of multiagent approach. In: 2018 International Conference on Smart Energy Systems and Technologies (SEST), Sevilla, Spain, 2018, pp. 1–6 (2018) 8. Gonz´ alez-Briones, A., De La Prieta, F., Mohamad, M., Omatu, S., Corchado, J.: Multi-agent systems applications in energy optimization problems: a state-of-theart review. Energies 11(8), 1928 (2018) 9. Casado-Vara, R., Prieto, J., De la Prieta, F., Corchado, J.M.: How blockchain improves the supply chain: case study alimentary supply chain. Procedia Comput. Sci. 134, 393–398 (2018) 10. Gonz´ alez-Briones, A., Villarrubia, G., De Paz, J.F., Corchado, J.M.: A multi-agent system for the classification of gender and age from images. Comput. Vis. Image Underst. 172, 98–106 (2018) 11. Corchado, J.M., Ramos, J., De Paz, J.F., Gonz´ alez-Briones, A.: Multi-agent system for obtaining relevant genes in expression analysis between young and older women with triple negative breast cancer. J. Integr. Bioinform. 12(4), 1–14 (2015) 12. Casado-Vara, R., Gonz´ alez-Briones, A., Prieto, J., Corchado, J.M.: Smart contract for monitoring and control of logistics activities: pharmaceutical utilities case study. In: The 13th International Conference on Soft Computing Models in Industrial and Environmental Applications, pp. 509–517. Springer, Cham (2018)

36

S. Marquez et al.

13. Casado-Vara, R., Chamoso, P., Coria, J.A.G., Herrera-Viedma, E., Corchado, J.M.: GarbMAS: simulation of the application of gamification techniques to increase the amount of recycled waste through a multi-agent system. In: Distributed Computing and Artificial Intelligence, 15th International Conference, vol. 800, p. 332. Springer, Cham (2018) 14. Casado-Vara, R., de la Prieta, F., Prieto, J., Corchado, J.M.: Blockchain framework for IoT data quality via edge computing. In: Proceedings of the 1st Workshop on Blockchain-enabled Networked Sensor Systems, pp. 19–24. ACM, November 2018 15. Briones, A.G., Chamoso, P., Barriuso, A.: Review of the main security problems with multi-agent systems used in e-commerce applications. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 5(3), 55–61 (2016) 16. Briones, A.G., Chamoso, P., Rivas, A., Rodr´ıguez, S., Yoe, H., Corchado, J.M.: A MAS Based Architecture to Reuse Waste Energy From Power Plants in Indoor Peppers Cultivation (2018) 17. Briones, A.G., Chamoso, P., Rivas, A., Rodr´ıguez, S., De La Prieta, F., Prieto, J., Corchado, J.M.: Use of gamification techniques to encourage garbage recycling. a smart city approach. In: International Conference on Knowledge Management in Organizations, pp. 674–685. Springer, Cham (2018) ˙ siker, H., Canbolat, H.: Concept for a novel grain level measurement method in 18. I¸ silos. Comput. Electron. Agric. 65(2), 258–267 (2009) 19. Sakharov, V.E., Kuznetsov, S.A., Zaitsev, B.D., Kuznetsova, I.E., Joshi, S.G.: Liquid level sensor using ultrasonic lamb waves. Ultrasonics 41(4), 319–322 (2003) 20. V´ azquez, C., Gonzalo, A.B., Vargas, S., Montalvo, J.: Multi-sensor system using plastic optical fibers for intrinsically safe level measurements. Sens. Actuators, A 116(1), 22–32 (2004) 21. Woodard, S.E., Taylor, B.D.: A wireless fluid-level measurement technique. Sens. Actuators, A 137(2), 268–278 (2007) 22. Jian, F., Jayas, D.S., White, N.D.: Temperature fluctuations and moisture migration in wheat stored for 15 months in a metal silo in Canada. J. Stored Prod. Res. 45(2), 82–90 (2009) 23. Casado-Vara, R., Corchado, J.M.: Blockchain for democratic voting: how blockchain could cast off voter fraud. Orient. J. Comp. Sci. Technol. 11(1) (2018) 24. Casado-Vara, R., Prieto, J., Corchado, J.M.: How blockchain could improve fraud detection in power distribution grid. In: The 13th International Conference on Soft Computing Models in Industrial and Environmental Applications, pp. 67–76. Springer, Cham (2018) 25. Casado-Vara, R., Novais, P., Gil, A.B., Prieto, J., Corchado, J.M.: Distributed continuous-time fault estimation control for multiple devices in IoT networks. IEEE Access 7, 11972–11984 (2019) 26. Chamoso, P., Gonz´ alez-Briones, A., Rivas, A., De La Prieta, F., Corchado, J.M.: Social computing in currency exchange. Knowl. Inf. Syst. 1–21 (2019) 27. Casado-Vara, R., Prieto-Castrillo, F., Corchado, J.M.: A game theory approach for cooperative control to improve data quality and false data detection in WSN. Int. J. Robust Nonlinear Control 28(16), 5087–5102 (2018) 28. Morente-Molinera, J.A., Kou, G., Gonz´ alez-Crespo, R., Corchado, J.M., HerreraViedma, E.: Solving multi-criteria group decision making problems under environments with a high number of alternatives using fuzzy ontologies and multi-granular linguistic modelling methods. Knowl. Based Syst. 137, 54–64 (2017) 29. Li, T., Sun, S., Boli´c, M., Corchado, J.M.: Algorithm design for parallel implementation of the SMC-PHD filter. Sig. Process. 119, 115–127 (2016). https://doi.org/ 10.1016/j.sigpro.2015.07.013

SiloMAS Platform

37

30. Chamoso, P., Rodr´ıguez, S., de la Prieta, F., Bajo, J.: Classification of retinal vessels using a collaborative agent-based architecture. AI Commun. (Preprint), 1–18 (2018) 31. Gonzalez-Briones, A., Prieto, J., De La Prieta, F., Herrera-Viedma, E., Corchado, J.M.: Energy optimization using a case-based reasoning strategy. Sensors (Basel) 18(3), 865–865 (2018). https://doi.org/10.3390/s18030865 32. Gonzalez-Briones, A., Chamoso, P., De La Prieta, F., Demazeau, Y., Corchado, J.M.: Agreement technologies for energy optimization at home. Sensors (Basel) 18(5), 1633–1633 (2018). https://doi.org/10.3390/s18051633 33. Gazafroudi, A.S., Corchado, J.M., Kean, A., Soroudi, A.: Decentralized Flexibility Management for Electric Vehicles. IET Renewable Power Generation (2019). http://ietdl.org/t/IBgIPb 34. Gazafroudi, A.S., Soares, J., Ghazvini, M.A.F., Pinto, T., Vale, Z., Corchado, J.M.: Stochastic interval-based optimal offering model for residential energy management systems by household owners. Int. J. Electr. Power Energy Syst. 105, 201–219 (2019) 35. Dang, N.C., de la Prieta, F., Corchado, J.M., Moreno, M.N.: Framework for retrieving relevant contents related to fashion from online social network data. In: PAAMS (Special Sessions), pp. 335–347 (2016) 36. Omatu, S., Wada, T., Rodr´ıguez, S., Chamoso, P., Corchado, J.M.: Multi-agent technology to perform odor classification. ISAmI 291, 241–252 (2014) ´ Rodr´ıguez, S., Corchado, S.: Improving intelligent systems: special37. Rom´ an, J.A., ization. In: PAAMS (Workshops), pp. 378–385 (2014)

Intelligent Livestock Feeding System by Means of Silos with IoT Technology Alfonso Gonz´ alez-Briones1(B) , Roberto Casado-Vara1 , Sergio M´ arquez1 , 1 1,2,3 Javier Prieto , and Juan M. Corchado 1

BISITE Research Group, University of Salamanca, Edificio I+D+i, Calle Espejo 2, 37007 Salamanca, Spain [email protected] 2 Department of Electronics, Information and Communication, Faculty of Engineering, Osaka Institute of Technology, Osaka 535-8585, Japan 3 Pusat Komputeran dan Informatik, Universiti Malaysia Kelantan, Karung Berkunci 36, Pengkalan Chepa, 16100 Kota Bharu, Kelantan, Malaysia

Abstract. Intelligent agriculture has the potential of increasing sustainability and productivity in the field of agriculture and livestock, through efficient and precise use of resources. Thus, this technology gives the possibility of promoting growth in developing countries through automation and control of repetitive farming activities, such as monitoring the level of water and feed in the feeders, which allows farmers to save time. However, the implementation of an automatic feed and water level control system in a livestock enclosure requires a large investment in silo scales, which may be too expensive for an SME. Thanks to the evolution of IoT devices, it is possible to reduce the cost of this implementation while integrating new functionalities and interactions through the interconnection of devices with cloud solutions. This work presents a new system that allows to monitor the quantity and quality of food and water in a silo by estimating volume in real time. Moreover, it has an additional functionality; temperature and humidity estimation in a livestock enclosure. The hardware system will be managed by a multi-agent system in charge of the processes of managing the data, managing the quantity of food and water supplied to each feeder. The use of a multi-agent architecture allows for the development of a distributed solution that provides great possibilities for future analysis, for example through a massive data analysis. The case study results demonstrate the effectiveness of the system, it has provided the ideal amount of feed and water to the animals, controlling the quality of grain and water, reducing the number of colics caused by overfeeding. In addition, the time the farmer must spend on the farm reduces considerably.

Keywords: Sensor-based monitoring Smart silo · IoT · Multi-agent system

· Ambiental intelligent ·

c Springer Nature Switzerland AG 2020  S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 38–48, 2020. https://doi.org/10.1007/978-3-030-00524-5_7

Intelligent Livestock Feeding System

1

39

Introduction

Farming dates back to the Neolithic period, when people began to form settlements, cultivate the land and keep animals. It was not until the twentieth century, that the appearance of tractors revolutionized farming [1]. Over that long period only minor changes occurred in harvesting methods and livestock management. Thanks to the tractor, however, farming became more developed, maximizing the production and minimizing the costs, it can be affirmed that the arrival of the tractor to this field meant an evolution as the industrial revolution. Nowadays, however, the filed of agriculture is in a process of continuous evolution thanks to the constant progress of electronics, computing and robotics. Since the invention of a tractor we’ve come a long way and many farming tasks are done automatically. For example, farmers can now enjoy robotic milking systems or automatic irrigation systems. Furthermore, considerable improvements have been made in terms of process optimization, leading to increased energy savings on, for example, agricultural crops [3,4], where the energy emitted in the process of energy generation is reused to grow crops outside their natural growing season [4,5], or by channeling wasted energy from power plants to heat greenhouses [7]. Energy saving is not the only area in which progress has been made, researchers are working on a wide variety of aspects that contribute to more advanced farming. In the state of the art there are numerous works on water saving, including that of Gonz´ alez-Briones et al. [10]. Smart city management is closely related to the management of agricultural and livestock environment, as explained by Corchado et al. [2]. Also, artificial intelligence techniques have been employed in the analysis of information captured by drones for cattle counting in uninhabited environments [9]. Great benefits can be achieved by applying artificial intelligence techniques and small interconnected devices to silos. The irruption of the Internet of Things (IoT) has allowed for the development of high-performing functionalities in lowcost systems and prototypes [13,14]. The weighers that calculate the amount of food and water in silos are an expensive piece of equipment whose cost could be decreased thanks to IoT. It is possible to prevent that livestock run out of feed and water through correct stock control and planned use of reserves. A system that is capable of measuring the volume of feed and water in a silo facilitates optimal management of resources on a livestock farm while reducing the time the farmer has to dedicate to this task. In addition, it ensures feed and water availability and safety to prevent livestock loss due to starvation/dehydration or poor feed quality. This work presents an intelligent feeding system that uses IoT technology to measure the volume of feed and water in a silo. The acquired data make it possible for the system to efficiently manage the quantity and quality of feed and water provided to the animals and to avoid the exhaustion of those resources. The system acquires additional information in real-time, such as temperature and humidity level. The hardware part of the system will be in charge of sending the data via WiFi connection for the processes of motorization and silo management.

40

A. Gonz´ alez-Briones et al.

Decision making will be based on the multi-agent system (MAS), responsible for measuring volume and for sending poor quality or lack of feed alters. The system has a NoSQL persistence system. The rest of the article is structured as follows: Sect. 2 reviews previous state of the art proposals in this area. Section 3 provides a detailed description of the intelligent livestock feeding system proposed in this work. Section 4 describes the case study and outlines the results. Finally, Sect. 5 draws conclusions from this research.

2

Related Work

This section presents the related works that have been carried out using the two main technologies used for the development of the Intelligent livestock feeding system. This section looks at previous works that have used. 2.1

Intelligent IoT Silos

One of the shortcomings detected in current proposals is the lack of a comprehensive system that would not only be capable of calculating feed and water volume accurately but also of measuring the temperature and humidity in a silo. These are important parameters that affect the quality of the feed and therefore, it is crucial that this information is available to both livestock farmers and potential buyers of livestock products (milk and meat companies) who will know the quality of the product they are going to buy. The process of measuring volume in silos is not new, in the past numerous techniques have been used to do this, however, by measuring the level of the fluid present in the silo where this fluid could be solid as grains or liquid by using a suitable sensor [16]. One of the most commonly used methods for measuring grain in the silo is through the implementation of sensors that use capacitance techniques [11] or [12]. The simulations performed in these works show that the output data corresponded with the sensor readings on the state of the silo; full of grain, realtime measurement or empty [6]. This makes it possible to use a system that measures the contents of a silo by means of ultrasound, without making direct contact with the grain [15]. However, one of the weaknesses of this approach is that there is no direct contact with the contents in the interior of the silo. Thus, if the grain is not evenly distributed in the silo, the obtained measurements will be inaccurate. Unlike liquids, it is very difficult to ensure that the grain in the silo remain evenly distributed [7,16,30–32]. 2.2

Multi-agent Systems

Due to the importance of accurately calculating the amount of feed and water in a silo, especially in the calculation of solid substances which present a greater

Intelligent Livestock Feeding System

41

problem in the precise calculation of volume. It is crucial to understand how to design a system that efficiently achieves volume calculation in the silo. Since the system aims to monitor several silos on the same site, such as a farm with several silos, it is necessary that the system is based on a distributed system that allows the complete management of the silos of the installation. The distributed methodology that best adapts to the needs of the system to be developed is an agent-based system. Agent-based systems allow the agents that make up the system to interact with each other without user intervention. The autonomy with which the agent systems are equipped allows them to perceive changes of context and react to them being an ideal approach for obtaining data from the facilities and develop responses to changes that occur with decisions developed specifically for this purpose. Multi-agent systems (MAS) are agent-based systems that provide extensibility and flexibility features, which make it possible to add algorithms, calculation techniques or new functionalities [29]. MAS are used to develop complex systems in diverse contexts. Several agent systems have been created for the development of Ambient Intelligence projects in cities [20,21] for example, to optimize energy consumption and achieve energy efficiency [8,18,19,22,25]. Their advantages have led researchers like Agrawal et al. [17] to adopt this approach to the problem of volume calculation in silos. In this work, humidity is controlled by installing a temperature and humidity sensor at the top of the silo instead of placing several sensors on all the levels of the silo. When a large volume of food in the solid state is on the surface at a humidity relative greater and then a diffusion of moisture through the material is extremely slow under practical conditions [24]. Temperature and humidity variations are greater outside of the silo than inside. The mass of grain and also the change of moisture within the grain mass is lesser than the change of moisture in the grain of the surface [23]. Since most of the work related to the problem of quality monitoring has been done through studies on temperature and humidity variations in the grains inside the silo, placing the sensors at the different points inside the silo. In view of the shortcomings of previous works, it is necessary to design and develop an Intelligent livestock feeding system based on a MAS. The MAS will use an IoT based measurement system to calculate the volume of the feed and water and to determine their quality. It will even send notifications to the user [18,19,21,33–36].

3

System Prototype

The system can accurately and automatically measure the level of grain in the silo without directly touching the surface. Thanks to the LiDAR laser technology, we can generate a 3D points cloud in a vertical silo to create 3D representations. Efficient grain stock management is unattainable without an accurate measurement system that would constantly monitor and quantify the volume of the grain inside the silo. This ineffectiveness may lead to grain shortage, lack of supplies,

42

A. Gonz´ alez-Briones et al.

livestock starvation, or feed wastage due to stock surplus -leading farmers and industries to economical loss. In the wake of this problem, we propose the development of an IoT (Internet of Things) device that can timely update the volume of grains inside the silo at different time instants. The mechanism allows for the measurement of the volume creating a middle sphere with great precision. In addition, it enables to collect data using temperature and humidity sensors to know the quality of the grain located inside the silo. The pan and tilt mechanism is used, which allows for the horizontal and vertical movement of the sensor by means of the progressive movement of the motors and then perform through a mathematical algorithm, the calculation of the feed that we have inside the silo. The algorithm is the result of the conversion of the LIDAR pulses, followed by a secondary correction for the focusing angle. Figure 1 shows three measurements of the interior of the silo made with our device. The image on the left belongs to a completely empty silo, the middle image represents a half full silo and the image on the right a practically full silo.

Fig. 1. The volume of grain measured inside of a silo in three different case studies.

The novelty of this system lies in its characteristics, apart of its robust mechanism the elements that make it up, like the glass of LIDAR, have dust repelling surfaces. The prototype includes a communication system module based on WiFi and 3G technology, through a modem placed near the device we transfer the data to a Web server where historical data are stored or configure the characteristics of the silo to be measured. The IoT device comprises of the following parts (see Fig. 2). The device must be powered with electricity, there are two ways of providing electricity, through solar energy panels or from the electricity grid if available locally. Thanks to LIDAR’S ability to measure all types of visible surfaces, regardless of their texture, it is useful to many industries. It accurately measures minerals, feed, grains, fibrous materials, cements, fly ash, coal, biofuels, synthetic plastics, ingredients, liquids, food packaging, chemicals and many other materials that are stored in bulk, tablets or granular forms. Thanks to its long measuring range it can reach the bottom of tall silos and measure through narrow openings (see Fig. 3).

Intelligent Livestock Feeding System

43

Fig. 2. Diagram of elements of the device. Modules and communications.

Fig. 3. Example of installation of the IoT Silo Prototype in the pilot study case.

4

Case Study

This section describes a case study of smart feeding on a rabbit farm. First, a general description of the experiment is given and then the results are presented. 4.1

General Description of the Experiment

The IoT measuring device has been placed at the top of the silo so that it can measure the amount of remaining grain. The selected silo is the Aviporc 230/5 model with central fall, 8.15 high and with a capacity of 20.65 m3 and approximately 12 tons of grain (see Fig. 4). The device was tested over 12 h the test began with 10 tons of corn inside the silo (approx. 16.5 m3 ), 1 ton was pulled in average at intervals of 35 g/kg of rabbit every hour (we assume that

44

A. Gonz´ alez-Briones et al.

Fig. 4. Aviporc 230/5 silo model with central fall.

the average weight of a rabbit is 2.5 kg). The rabbit farm has 1000 rabbits and the rabbits are on fattening or breeding diets. The values of Hsensor1 , Hsensor2 , Hsensor3 are computed using the Ultrasonic GH-311 RT sensor fitted to the HS-645MG Servo motor. Both devices are connected by an Arduino One. The HS-645MG Servo motor can rotate from 0 to 25◦ and make angles of 0, 9.9 and 17.53◦ from the top of the silo. Thus, using the measurements Hsensor1 , Hsensor2 , and Hsensor3 , the volume of the grain in the silo is calculated at each time interval. The geometry of the silo is taken into account in those calculations. The chosen temperature and humidity sensor is DHT 11. This device is connected to the ultrasonic sensor via the Arduino one. This sensor collects temperature and humidity inside the silo at 10-min intervals. Finally, the server application connects to a MySQL database and the data collected by the sensors are stored in the following order: Timestamp, Temperature, Humidity, Hsensor1 , Hsensor2 , Hsensor3 and volume of the grain in the silo. 4.2

Experimental Results

Figure 5 shows the real-time and measured values of grain during the 12-h interval. In the figure we can find that the measurement is slightly lower than the current value. This is because a small weight estimation error has been found in the measurements, below 4% in average during the case study. In this way, if we predict the grain consumption, the measurement is more kind and therefore, you can fill the silo with enough anticipation.

Intelligent Livestock Feeding System

45

Fig. 5. Amount of grain over the 12 h of measurements.

Moreover, environmental data was collected by the temperature and humidity sensors inside of the silo. In this way, the quality of the grain was controlled and its damage due to poor storage conditions could be prevented. The temperature of the silo should be between 18◦ and 22 ◦ C so that bacteria do not begin to grow. Figure 6 shows the temperature values collected by the sensor. In addition, the grain should be stored between 10% and 20% humidity so that no mold grows.

Fig. 6. Temperature and humidity of the silo over the 12 h of measurements.

5

Conclusions

This work has presented a MAS for the monitoring of the volume and quality of feed and water in agricultural silos. The calculation of quantity and quality of feed and water is fundamental in agriculture; both the farmers and the buyers

46

A. Gonz´ alez-Briones et al.

must know the quality of the products they sell/purchase, respectively. The hardware prototype presents a novel model for volume measurement, based on the IoT concept and the use of ultrasonic sensors. This model also allows to measure environmental conditions. The agent based system is responsible for obtaining the measurements from all prototypes deployed in livestock facilities and sends the data, which, on the basis of the data, manages the supply of the feed to the livestock. The results obtained in the case study conducted on a rabbit farm, show that the proposed MAS system is a promising solution which can be used in a wide range of agricultural silos, benefiting the farming industry, the rabbit meat industry and facilitating the work of farmers and silo owners, etc. allowing them to correctly optimize the available resources and to take well-informed decisions, improving efficiency and economy. As future work will be evaluated the possibility of integrating blockchain technology to add the benefits of this technology to this development [26,27]. Acknowledgements. This work has been partially supported by the Agreement between the Agricultural Technology Institute of Castile and Le´ on, Hermi Gesti´ on, S.L., and the University of Salamanca to conduct research activities on the development of a farm 4.0 model in the rabbit meat production sector.

References 1. Olmstead, A.L., Rhode, P.W.: Reshaping the landscape: the impact and diffusion of the tractor in American agriculture, 1910–1960. J. Econ. Hist. 61(3), 663–698 (2001) 2. Corchado, J.M., Rodr´ıguez, S., Chamoso, P., Gonz´ alez, A.: Experimental communications network and applications within the Smart Village Project 3. Corchado, J.M., Rodr´ıguez, S., Chamoso, P., Gonz´ alez, A.: Virtual organization designed for recycling waste energy from power plants 4. Gonz´ alez-Briones, A., Chamoso, P., Rodr´ıguez, S., Yoe, H., Corchado, J.M.: Reuse of waste energy from power plants in greenhouses through MAS-based architecture. Wirel. Commun. Mob. Comput. (2018) 5. Briones, A.G., Chamoso, P., Rivas, A., Rodr´ıguez, S., Yoe, H., Corchado, J.M.: A MAS based architecture to reuse waste energy from power plants in indoor peppers cultivation 6. Casado-Vara, R., Gonz´ alez-Briones, A., Prieto, J., Corchado, J.M.: Smart contract for monitoring and control of logistics activities: pharmaceutical utilities case study. In: The 13th International Conference on Soft Computing Models in Industrial and Environmental Applications, pp. 509–517. Springer, Cham, June 2018 7. Gonz´ alez-Briones, A., Chamoso, P., Yoe, H., Corchado, J.M.: GreenVMAS: virtual organization based platform for heating greenhouses using waste energy from power plants. Sensors 18(3), 861 (2018) 8. Gonz´ alez-Briones, A., Prieto, J., Corchado, J.M., Demazeau, Y.: EnerVMAS: virtual agent organizations to optimize energy consumption using intelligent temperature calibration. In: International Conference on Hybrid Artificial Intelligence Systems, pp. 387–398. Springer, Cham, June 2018

Intelligent Livestock Feeding System

47

9. Rivas, A., Chamoso, P., Gonz´ alez-Briones, A., Corchado, J.: Detection of cattle using drones and convolutional neural networks. Sensors 18(7), 2048 (2018) 10. Gonz´ alez-Briones, A., Castellanos-Garz´ on, J.A., Mezquita Mart´ın, Y., Prieto, J., Corchado, J.M.: A framework for knowledge discovery from wireless sensor networks in rural environments: a crop irrigation systems case study. Wirel. Commun. Mob. Comput. (2018) 11. Ryu, M., Yun, J., Miao, T., Ahn, I.Y., Choi, S.C., Kim, J.: Design and implementation of a connected farm for smart farming system. In: 2015 IEEE SENSORS, pp. 1–4. IEEE, November 2015 12. Kulatunga, C., Shalloo, L., Donnelly, W., Robson, E., Ivanov, S.: Opportunistic wireless networking for smart dairy farming. IT Prof. 19(2), 16–23 (2017) 13. Casado-Vara, R., Prieto, J., De la Prieta, F., Corchado, J.M.: How blockchain improves the supply chain: case study alimentary supply chain. Procedia Comput. Sci. 134, 393–398 (2018) 14. Casado-Vara, R., de la Prieta, F., Prieto, J., Corchado, J.M.: Blockchain framework for IoT data quality via edge computing. In: Proceedings of the 1st Workshop on Blockchain-Enabled Networked Sensor Systems, pp. 19–24. ACM, November 2018 15. Awad, T.S., Moharram, H.A., Shaltout, O.E., Asker, D., Youssef, M.M.: Applications of ultrasound in analysis, processing and quality control of food: a review. Food Res. Int. 48(2), 410–427 (2012) 16. Casado-Vara, R., Prieto-Castrillo, F., Corchado, J.M.: A game theory approach for cooperative control to improve data quality and false data detection in WSN. Int. J. Robust Nonlinear Control 28(16), 5087–5102 (2018) 17. Agrawal, H., Prieto, J., Ramos, C., Corchado, J.M.: Smart feeding in farming through IoT in silos. In: The International Symposium on Intelligent Systems Technologies and Applications, pp. 355–366. Springer, Cham, September 2016 18. Gonz´ alez-Briones, A., Prieto, J., De La Prieta, F., Herrera-Viedma, E., Corchado, J.M.: Energy optimization using a case-based reasoning strategy. Sensors (Basel) 18(3), 865 (2018). https://doi.org/10.3390/s18030865 19. Gonz´ alez-Briones, A., Chamoso, P., De La Prieta, F., Demazeau, Y., Corchado, J.M.: Agreement technologies for energy optimization at home. Sensors (Basel) 18(5), 1633 (2018). https://doi.org/10.3390/s18051633 20. Gonz´ alez-Briones, A., Valdeolmillos, D., Casado-Vara, R., Chamoso, P., Coria, J.A.G., Herrera-Viedma, E., Corchado, J.M.: GarbMAS: simulation of the application of gamification techniques to increase the amount of recycled waste through a multi-agent system. In: International Symposium on Distributed Computing and Artificial Intelligence, pp. 332–343. Springer, Cham, June 2018 21. Chamoso, P., Gonz´ alez-Briones, A., Rodr´ıguez, S., Corchado, J.M.: Tendencies of technologies and platforms in smart cities: a state-of-the-art review. Wirel. Commun. Mob. Comput. (2018) 22. Gonz´ alez-Briones, A., De La Prieta, F., Mohamad, M., Omatu, S., Corchado, J.: Multi-agent systems applications in energy optimization problems: a state-of-theart review. Energies 11(8), 1928 (2018) 23. Jian, F., Jayas, D.S., White, N.D.: Temperature fluctuations and moisture migration in wheat stored for 15 months in a metal silo in Canada. J. Stored Prod. Res. 45(2), 82–90 (2009) 24. Pixton, S.W., Griffiths, H.J.: Diffusion of moisture through grain. J. Stored Prod. Res. 7(3), 133–152 (1971)

48

A. Gonz´ alez-Briones et al.

25. Gonz´ alez-Briones, A., Chamoso, P., Prieto, J., Corchado, J.M., Yoe, H.: Reuse of wasted thermal energy in power plants for agricultural crops by means of multiagent approach. In: 2018 International Conference on Smart Energy Systems and Technologies (SEST), Sevilla, Spain, pp. 1–6 (2018) 26. Casado-Vara, R., Corchado, J.M.: Blockchain for democratic voting: how blockchain could cast off voter fraud. Orient. J. Comp. Sci. Technol. 11(1) (2018) 27. Casado-Vara, R., Prieto, J., Corchado, J.M.: How blockchain could improve fraud detection in power distribution grid. In: The 13th International Conference on Soft Computing Models in Industrial and Environmental Applications, pp. 67–76. Springer, Cham, June 2018 28. Briones, A.G., Chamoso, P., Barriuso, A.: Review of the main security problems with multi-agent systems used in e-commerce applications. ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 5(3), 55–61 (2016) 29. Casado-Vara, R., Chamoso, P., De la Prieta, F., Prieto, J., Corchado, J.M.: Nonlinear adaptive closed-loop control system for improved efficiency in IoT-blockchain management. Inf. Fusion 49, 227–239 (2019) 30. Casado-Vara, R., Novais, P., Gil, A.B., Prieto, J., Corchado, J.M.: Distributed continuous-time fault estimation control for multiple devices in IoT networks. IEEE Access 7, 11972–11984 (2019) 31. Chamoso, P., Gonz´ alez-Briones, A., Rivas, A., De La Prieta, F., Corchado, J.M.: Social computing in currency exchange. Knowl. Inf. Syst. (2019) 32. Morente-Molinera, J.A., Kou, G., Gonz´ alez-Crespo, R., Corchado, J.M., HerreraViedma, E.: Solving multi-criteria group decision making problems under environments with a high number of alternatives using fuzzy ontologies and multi-granular linguistic modelling methods. Knowl.-Based Syst. 137, 54–64 (2017) 33. Li, T., Sun, S., Boli´c, M., Corchado, J.M.: Algorithm design for parallel implementation of the SMC-PHD filter. Sig. Process. 119, 115–127 (2016). https://doi.org/ 10.1016/j.sigpro.2015.07.013 34. Chamoso, P., Rodr´ıguez, S., de la Prieta, F., Bajo, J.: Classification of retinal vessels using a collaborative agent-based architecture. AI Commun. (Preprint) 31(5), 427–444 (2018) 35. Gazafroudi, A.S., Corchado, J.M., Kean, A., Soroudi, A.: Decentralized flexibility management for electric vehicles. IET Renew. Power Gener. (2019). http://ietdl. org/t/IBgIPb 36. Gazafroudi, A.S., Soares, J., Ghazvini, M.A.F., Pinto, T., Vale, Z., Corchado, J.M.: Stochastic interval-based optimal offering model for residential energy management systems by household owners. Int. J. Electr. Power Energy Syst. 105, 201–219 (2019)

Cooperative Algorithm to Improve Temperature Control in Recovery Unit of Healthcare Facilities Roberto Casado-Vara(B) , Fernando De la Prieta, Sara Rodriguez, Javier Prieto, and Juan M. Corchado BISITE Research Group, University of Salamanca, Calle Espejo 2, 37007 Salamanca, Spain {rober,fer,srg,javierp,corchado}@usal.es

Abstract. Healthcare facilities spend a lot of resources on taking care of patients while they recover from their illnesses. IoT (Internet of Things) devices are used to monitor and control the environment of healthcare facilities. According to Spanish standards of hygiene and safety in hospitals: the temperature must be between 18 ◦ and 24 ◦ C and relative humidity of 60%. In this paper, we present a cooperative control algorithm to increase data quality and false data detection via edge computing in healthcare facilities. Furthermore, it is demonstrated that blockchain can be used to store data in an immutable and secure way. In this work we present a new model for the efficient control and monitoring of indoor temperature in healthcare facilities, reducing energy consumption and storing data in a secure and immutable way via blockchain. Keywords: IoT · Algorithm design Blockchain · Cooperative control

1

· Game theory · e-health ·

Introduction

In the last few years, the term blockchain has been very frequently used among the scientific community. Gartner proposed in July 2016 that blockchain was a distributed database. First described by Nakamoto in 2008, he developed the concept of Bitcoin, an encryption-based virtual currency. Today, Bitcoin is the most advanced application of blockchain and the most advanced technological concept in academia. However, blockchain has many more applications in other fields and industries. Since Bitcoin’s blockchain only serves to store cryptocurrency transactions, other blockchain systems have been developed (e.g., Ethereum) for making smart contracts and storing a wide variety of data. Ethereum is a decentralized platform that runs smart contracts: applications that run exactly as programmed, without any possibility of downtime, censorship, fraud or thirdparty interference. This enables developers to store registries of data, move funds in accordance with instructions given long in the past (like a will or a future c Springer Nature Switzerland AG 2020  S. Omatu et al. (Eds.): DCAI 2018, AISC 802, pp. 49–62, 2020. https://doi.org/10.1007/978-3-030-00524-5_8

50

R. Casado-Vara et al.

contract) and many other things that have not been invented yet, all without a middleman or counterparty risk. Ethereum makes it possible to establish smart contracts between several parties. Several actions are involved in the process of making smart contracts in the blockchain. To build a block with the data collected by the smart nodes, the smart contract runs. Once the block is validated by the miners, it can be introduced into the blockchain. Furthermore, since network consensus is always necessary, altering records becomes very difficult and expensive. This prevents individuals or groups from changing a blockchain record with the effort of trying to make false data look precise and authentic [5]. One of the problems detected by the authors in monitoring and controlling the environment in which patients recover is that the temperature in the rooms is not constant. There are several reasons for this, one of the main causes is that the patients’ family members may change the temperature of their thermostats, open the windows of their rooms, etc. As a consequence of temperature fluctuations, patients have more difficulties in restoring their health. Furthermore, to ensure that patients are at the temperatures that are recommended by national regulations, many resources are wasted on heating and cooling. This results in low-quality healthcare as professionals are limited by the control systems available to them. Therefore, we consider it necessary to develop a new e-health model that will have all the functionalities required for monitoring and controlling temperature in healthcare facilities. There have been attempts to use multi-agent systems to optimize and improve the e-health system [2] however, the implementation of a WSN does not eliminate the key problems of centralized systems [6]. In this paper, a new temperature control system is proposed for use in the field of e-Health. This new system has an architecture with an edge computing layer in which data is transformed by applying an algorithm to improve data quality and false data detection. In this way, it is possible to improve efficiency in monitoring healthcare facilities. In the case study presented in this paper, IoT devices monitored the temperature of a healthcare facility. The cooperative algorithm that is executed in edge computing layer increases data quality by using game theory to reach temperature consensuses between IoT e-health devices and auto-correcting the inaccurate temperatures. Thus, the use of energy to heat or cold healthcare facilities can be optimized, offering a better-quality service to the patients. On the other hand, blockchain technology is used in this system to store temperatures. Thus, only authorized staff have access to this data. This paper is organized as follows: Sect. 2 reviews the literature related to blockchain. Section 3 proposes the architecture and the cooperative algorithm. Section 6 draws conclusions from the conducted research.

2

Related Work

Many researchers have done paramount studies on networking architectures. In the field of medicine the literature features examples of architectures for IoT in various fields such as military health services on the battlefield [22], ambient assisted living applications [17], patient health monitoring system [12] and

Cooperative Algorithm Healthcare

51

hospitalized patient monitoring [7]. Other researchers use the support of IoT architectures to present their algorithms for e-health [9]. Some of the latest work in this field is the use of algorithms to detect Alzheimer’s disease [27], algorithms for monitoring and alerting system [16], regular monitoring of arthritis disease [20] and early detection of heart diseases [15,31]. In this line our research focuses on presenting an algorithm to monitor and control temperature in patient recovery environments. To achieve this, we present an IoT architecture for smart Hospital in which the proposed algorithm can be used. Many researchers asssume that “WSNs are subsets of IoT”. WSNs are used to control and monitor a wide range of things [11,32]. Some works achieve good results in fields such as energy efficiency using WSNs [3], control of operations [8], optimal routing in WSNs [26,33], and some other applications such as social good [21]. Sensor networks can also be used with other technologies such as multi-agent systems to manage data [26], for data mining [23], multi-agent localization and WSN. These algorithms are also being used for data mining [1,34]. Moreover, WSNs and GT are the areas that are currently undergoing intense study, [10] is focused on finding innovative solutions to the challenges presented by next-generation WSNs. Since GT is an ideal tool for designing efficient and robust distributed algorithms, its use in the design and analysis of WSN information is attracting increased attention [19,35]. This survey looks at how GT is currently being used in WSNs. In [25,36] the authors make a general classification of the different uses of GT in WSNs; they are classified into the following groups: network management, communication, network security and applications. Our proposal can be included in the applications category, within the data collection subgroup. Existing works provide diverse frameworks for blockchain and IoT. Moreover, applications that merge both technologies are being developed. One of the earliest development was an application that authenticates and increases the reliability of WSNs using blockchain [4,37] [18,38]. Other applications also combine the Internet of Things and the global commerce [13,39]. Regarding the blockchain-based architectures, Q. Xia et al. [28] provide a framework for the exchange of health data based on a blockchain that addresses the access control challenges associated with sensitive data stored in the cloud. The system is based on a blockchain with permissions that allow access only to invited and therefore verified users. Yue et al. [29] present an architecture based on 3 layers: data usage layer, data management layer and data storage layer. This paper discusses the use of a private blockchain that acts as a cloud. Kuo et al. [14] make an in-depth revision of the latest biomedical/sanitary applications of blockchain technologies. The authors discuss the potential changes that these applications and architectures need and suggest solutions using the blockchain technology in the biomedical/sanitary field. In contrast to this work focused on the exchange of health data, [24,40] and [30] focus on other problems. Shae et al. [24] propose a blockchain-based architecture for clinical trials and precision medicine. Zhao et al. [30] use a WSN to design a lightweight backup and efficient recovery scheme for healthcare systems. This is a pioneering work in blockchain key man-

52

R. Casado-Vara et al.

agement, while its performance is heavily influenced by the state of the hardware and environment. The methods presented in the reviewed literature overcome different challenges concerning e-health. IoT architectures and blockchain-based architectures are presented in response to several issues detected by researchers. However, we observed a gap in the literature review since there are no algorithms that could automatically validate the temperature collected by a heterogeneous WSN. In our work, coalitions of neighbours are created by using clustering techniques. This distributed and self-organized (overall temperature equilibrium arises from local game interactions between sensors of an initially disordered temperature system) game is designed to provide reliability and robustness to the data collected by a WSN. It identifies defective sensors gathering inaccurate measurements and detects areas with similar temperatures. This article tackles the problem of reliability of WSN data from the point of view of game theory and probability, which is a novel approach in this field.

3

Proposed Architecture

The main goal of this work is to present a cooperative control algorithm that improves data quality of the temperature collected by the smart WSN nodes. To achieve our goal, we propose an architecture that allows to monitor and control the temperature of healthcare facilities. Through the literature review, we present a 3-layer architecture: data collection layer, data management layer and data storage and security layer. 3.1

Proposed System Architecture

This architecture has 3 layers: (1) Data collection layer. Sensors collect data from the environment or object under measurement and turn it into useful data. Data is at the core of an IoT architecture, and we have to decide between the immediacy and depth of knowledge when processing these data. The more immediate the need for information, the closer to the end devices your processing needs to be. The data from the sensors is started in analog form. The DAS connects to the WSN, adds inputs and carries out the transformation from analog to digital. The Internet gateway receives the digitized data and routes it to the data management layer for further processing. Here, temperature sensors monitor the temperature of healthcare facilities. (2) Data management layer. Once data has been digitized it is ready to enter the data management layer. However, data may require processing before entering the storage layer. This is where edge computing systems come into play, they perform more analyses. Usually, devices that are in the edge computing system sit in the facility or location where the sensors reside closer to the sensors (i.e., Smart controller). Digitized data is sent to the smart controllers. Then, the Raspberry Pi first runs a cooperative control algorithm to improve data quality and false data detection. Once data is transformed by the cooperative control algorithm, Raspberry Pi sends the data

Cooperative Algorithm Healthcare

53

to the sidechain (this sidechain is Rinkeby, a fork of Ethereum) that builds the block with the data sent to it by the IoT nodes. Once the block is built in the sidechain, the Raspberry Pi runs a smart contract in the Ethereum blockchain (i.e., main chain) to validate the sidechain of a new block. If the block is validated, the Raspberry Pi sends the block to the miners’ network for storage in the blockchain. (3) Storage and security layer. In blockchain network, there are two important entities: Miners and verifiers. Miners refer to the nodes who produce new blocks for the blockchain. Different application scenarios may define different nodes as the miners. New blocks are accepted only after validation by the verifiers, which are responsible for verifying the new blocks’ authenticity. Processes of generation, verification and inclusion of new blocks in the blockchain are called mining. To ensure the safety and reliability of mining processes, the consensus mechanism is critical in the blockchain network. In this work, transactions denote the temperature of the healthcare facilities records in the system. In relation to our work, Ethereum blockchain is introduced in the e-Health system to store and manage the temperature, which helps to improve the recovery time.

Fig. 1. High-level system architecture design diagram. Central to our research is an Ethereum blockchain, which provides a mechanism to execute logic and manage node interactions via smart contracts. Nodes share a blockchain database and communicate over the blockchain.

4

Cooperative Control Algorithm

The cooperative control algorithm requires the data to be in a matrix. Therefore, the first transformation that data has is to place them in an ordered mesh from point (1,1) to point (n,n) so that each of these points matches the position of the smart nodes. From the mesh it is easy to create a matrix in which the cooperative

54

R. Casado-Vara et al.

control algorithm is applied. Without loss of generality, if we have a mesh with n sensors ordered from (1,1) to (n,n), matrix shown in Eq. 1 is created. ⎞ ⎛ ts1,1 . . . ts1,n ⎟ ⎜ Tn,n = ⎝ ... . . . ... ⎠ (1) tsn,1 . . . tsn,n 4.1

Mathematical Description of the Algorithm

Let n ≥ 2 denote the number of players in the game, numbered from 1 to n, and let N = {1, 2,...,n} denote the set of players. A coalition, S, is defined to be a subset of N, S ⊆ N , and the set of all coalitions is denoted by S. A cooperative game in N is a function u (characteristic feature of the game) that assigns to each coalition Si ⊆ S a real number u(Si ). In addition one has the condition u(∅) = 0. In our case, the game will be non-negative (the values of the characteristic function are always positive), monotonous (if more players are added to the coalition the value of the expected characteristic function does not change), simple and 0-normalized (players are obliged to cooperate with each other since individually they will obtain zero benefit). In our case, the set of players is the set of ordered sensors S and the characteristic function u is defined as: u : 2n −→ {0, 1}

(2)

such that, for each coalition of sensors, u = 1 or 0 if that particular coalition can vote or not respectively (see Eq. (2, 3)). S  Si −→ u(Si ) = {0, 1} ∈ R

(3)

where R are the Real numbers. 4.2

Cooperative Sensor Coalitions

The possible coalitions that the sensors will form, will be limited by their position, that is, the coalitions can only be formed by neighbouring sensors. Let’s consider the matrix of the sensors and a pair of sensors si,j and sk,m will be in the same neighbourhood if and only if:  (i − k)2 − (j − m)2 ≤ 1

(4)

that is, if each sensor to which the game is applied, is the center of a Von Neumann neighbourhood, its neighbours are those lying within a Manhattan distance (in the matrix) equal to one. In addition, the following conditions have to be fulfilled by the allowed coalitions: 1. Coalition sensors have to be in the same neighborhood as defined in Eq. 4. 2. Coalitions cannot be formed by a single sensor.

Cooperative Algorithm Healthcare

4.3

55

A Characteristic Function to Find cooperative temperatures.

In the proposed game, we want the neighbourhood coalitions to democratically decide the temperature of the main sensor. To do this, they will form coalitions that will decide on the final temperature of the sensor, which will be determined by whether they can vote or not in the process. From the characteristic function defined in Eq. (2), if the value is 1(0), the coalition can vote (not vote) respectively. si is the main sensor with its associated temperature tsi , the characteristic function is built in the following way: 1. First, the average temperature of all the sensors is calculated: Tski =

V 1  ts V i i

(5)

here Ts1i represents the average temperature of the sensors’ neighbourhood si (including it) in the first iteration of the game and V is the number of neighbours in the coalition. 2. The next step is to compute an absolute value for the temperature difference between the temperatures of each sensor and the average temperature: k T si

=

V 1  | tsi − Tski |2 V i

12 (6)

3. Using the differences in temperature values with regards to the average k temperature T si (see Eq. (6)) a confidence interval is created and defined as follows: k

T Iski = Tski ± t(V −1, α2 )√ si (7) V in Eq. (7) we use the Student’s-t distribution with an error of 1%. 4. In this step we use a hypothesis test. If the temperature of the sensor lies in the interval Iski , it belongs to the voting coalition, otherwise, it is not in the voting coalition:  1 if tsi ∈ Iski k (8) u (s1 , . . . , sn ) = 0 if tsi ∈ Iski 5. The characteristic function will repeat this process iteratively (k is the number of the iteration) until all the sensors in that iteration belong to the voting coalition. In each iteration k, the following payoff vector of the coalition Sj (with 1 ≤ j ≤ n where n is the number of sensors in the coalition) in the step k (P V (Sjk )) is available: PV

(Sjk )

k

k

= (u (s1 ), . . . , u (sn )) where

n  i

uk (si ) ≤ n

(9)

56

R. Casado-Vara et al.

The stop condition of the game iterations is P V (Sjk ) = P V (Sjk+1 ) the process end. That is, let P V (Sjk ) = (uk (s1 ), . . . , uk (sn )) and let P V (Sjk+1 ) = (uk+1 (s1 ), . . . , uk+1 (sn )). The iteration process ends when both payoff vectors contain the same elements. This process is shown in the following equation: ⎧ k k+1 (s1 ) ⎪ ⎨ u (s1 ) = u .. (10) . ⎪ ⎩ k u (sn ) = uk+1 (sn ) Solution Concept of the Cooperative Game. Once the characteristic function has been applied to all sensors involved in this step of the game a payoff vector in the step k is available (see Eq. 9). Since the proposed game is a cooperative game, the solution concept is a coalition of players that we have called game equilibrium (GE). The GE of the proposed game is defined as the minimal coalition with more than half of the votes cast. Let n be the number of players involved in this step of the game. Winning coalition must satisfy the following conditions: 1. Sum of the elements of the coalition PV must be higher than half plus 1 of the votes cast: n  n (11) uk (si ) ≥ + 1 2 i 2. The coalition is maximal (i.e., coalition with the greatest number of elements, different from 0, in its payoff vector P V (Sjk )). Therefore, the solution to the proposed game is the coalition that verifies both conditions from among all possible coalitions that are formed at each step k of the game. 4.4

Temperatures of the Winning Coalition

Once the characteristic function decides which is the winning coalition, it is possible to calculate the temperature of the main sensor. Let {s1 , . . . , sj } be the winning coalition’s sensors and {ts1 , . . . , tsj } be their associated temperature. The temperature that the game has voted to be the main sensor’s temperature (MST) is calculated as follows: M ST =

max

j∈|Swinner |

{j ∗ tsi }si∈Swinner

(12)

where |S| is the number of elements in the winning coalition. Therefore, the MST will be the maximum temperature that has the highest relative frequency. In case of a draw, it is resolved by the Lagrange criterion.

Cooperative Algorithm Healthcare

57

Diffuse Convergence. In each game iteration, there is a matrix with temperature (see Eq. 1). Hence we define a sequence of arrays {Mn }n∈N where the Mi element corresponds to the temperature matrix in step i of the game. Therefore, it can be said that the sequence of matrices is convergent if: ∀ > 0, there is N ∈ N such that |Mi−1 − Mi | ≤  ∀i ∈ N.

(13)

i That is, if the element mi−1 n,m ∈ Mi−1 and the element mn,m ∈ Mi are set and the convergence criterion is applied, we have: i ∀n,m > 0 there is N ∈ N such that |mi−1 n,m − mn,m | ≤ n,m i ∀i ∈ N , ∀i ≥ N and mi−1 n,m ∈ Mi−1 , mn,m ∈ Mi

(14)

Therefore, by applying the criterion of convergence in Eq. (14) to all the elements, a new matrix is obtained with the temperature differences between the temperatures obtained in previous and in the next step of the game. ⎛ i−1 ⎞ i |m1,1 − mi1,1 | . . . |mi−1 1,m − m1,m | ⎜ ⎟ .. .. .. (15) ⎝ ⎠ . . . i−1 i i−1 i |mn,1 − mn,1 | . . . |mn,m − mn,m | For the succession of matrices to be convergent, each of the sequences of elements i that are formed with the |mi−1 n,m − mn,m | must be less than the fixed  > 0. In this work, it is established that  = 0.01. The game reach the equilibrium if at least 80% of the elements of the matrix are convergent.

5

Empirical Results of the Case Study

In this case study the temperature of a recovery unit of a hospital in Salamanca (Spain) was collected. The IoT nodes collect data from the rooms and corridors. Then, the cooperative algorithm is used to increase the quality of the collected data and false data detection. In this way the hospital can monitor the temperature of the environment more effectively. Finally, data that has been transformed is stored in the blockchain. Once the mesh is made with the position of the IoT temperature nodes, the temperature is collected and stored in a matrix. In this case study, the cooperative algorithm is executed in the edge computing layer, transforms data before it reaches blockchain. The type of sensor used (in the IoT temperature nodes) is a combination of the ESP8266 microcontroller in its commercial version “ESP-01” and a DHT11 temperature and humidity sensor (Fig. 1). Their combination allows for greater flexibility when collecting data and adaptability to the case study, since the DHT11 sensor is designed for indoor environments (has an operating range of between 0 ◦ C and 50 ◦ C) according to its datasheet. The microcontroller collects data from this sensor through a onewire protocol and communicates it to the environment via WiFi using HTTP standards and GET/POST requests. The device is programmed using the ESP-IDF programming environment provided by the manufacturer of the microcontroller.

58

R. Casado-Vara et al.

Fig. 2. Evolution of surface temperatures over the different steps in the game until reaching game equilibrium.

In Fig. 2, the first image shows the initial temperature and in the rest of the images the iterations of the game until the GE is reached. In the successive images, the temperature clusters are being formed, this can be observed by the changes in the colour gradient. It can also be seen that some areas with inaccurate temperatures self-correct smoothly on the basis of the temperatures in their environment. This is the intended process, since the game is executing its iterations depending on the environment surrounding the sensor. This makes sense because the temperature of the sensor will be similar to the average temperature of the environment in which it is located. Also, we show the evolution of the temperature on the surface. To this end, we have represented the temperature on the z axis to facilitate visualization in the form of a surface. Experimental results about convergence and sensors that are providing inaccurate measurements are shown in Fig. 3. In the graph on the left it can be found that the algorithm reaches the diffuse convergence defined in this paper (i.e., 80% of the elements of the matrix reach convergence). The algorithm takes less than 10 steps to produce reliable temperatures relative to their neighbors. On the other hand, on the right side of the figure, the number of sensors that are performing inaccurate measurements (assuming a 0.01 ◦ C error) is 15%. While in 10 stages of the algorithm this number decreases below 5%. The algorithm increases the number of sensors that are providing precise temperatures according to their neighbors in less than 10 steps.

Cooperative Algorithm Healthcare

59

Fig. 3. Left: The algorithm achieves convergence in less than 10 steps. While from the 50 stages onwards the temperatures reach the optimal temperature regarding to their neighborhood. Right: Assuming that we have allowed an error of 0.01 ◦ C, at the first step of the algorithm one notices that 15% of the sensors are providing inaccurate temperatures. Once 10 stages of the algorithm have occurred, sensors that are sensing imprecise temperatures are under 5%.

6

Conclusions

This paper presents a distributed and self-organized cooperative algorithm using game theory. The algorithm has been applied to data collected by IoT e-health devices. In addition, a blockchain-based architecture is suggested to improve data security. The novelty of this architecture lies in the fact that it provides an edge computing layer in which the cooperative algorithm is executed to improve data quality and the detection of false data. On the other hand, this new algorithm improves the energy efficiency of healthcare facilities by applying algorithms in the edge computing layer. Acknowledgment. This work was developed as part of “Virtual-Ledgers-Tecnolog´ıas DLT/Blockchain y Cripto-IOT sobre organizaciones virtuales de agentes ligeros y su aplicaci´ on en la eficiencia en el transporte de u ´ ltima milla”, ID SA267P18, project cofinanced by Junta Castilla y Le´ on, Consejer´ıa de Educaci´ on, and FEDER funds.

References 1. Agrawal, R., Gehrke, J., Gunopulos, D., Raghavan, P.: Automatic subspace clustering of high dimensional data for data mining applications. In: Proceedings of ACM SIGMOD International Conference Management of Data, pp. 94–105 (1998) 2. Alfian, G., Syafrudin, M., Ijaz, M.F., Syaekhoni, M.A., Fitriyani, N.L., Rhee, J.: A personalized healthcare monitoring system for diabetic patients by utilizing BLE-based sensors and real-time data processing. Sensors 18, 2183 (2018) 3. Biswas, S., Das, R., Chatterjee, P.: Energy-efficient connected target coverage in multi-hop wireless sensor networks. In: Industry Interactive Innovations in Science, Engineering and Technology, pp. 411–421. Springer, Singapore (2018)

60

R. Casado-Vara et al.

4. Casado-Vara, R., Prieto-Castrillo, F., Corchado, J.M.: A game theory approach for cooperative control to improve data quality and false data detection in WSN. Int. J. Robust Nonlinear Control 28, 5087–5102 (2018) 5. Casado-Vara, R., Gonz´ alez-Briones, A., Prieto, J., Corchado, J.M.: Smart contract for monitoring and control of logistics activities: pharmaceutical utilities case study. In: The 13th International Conference on Soft Computing Models in Industrial and Environmental Applications, pp. 509–517. Springer, Cham (2018) 6. Morente-Molinera, J.A., Kou, G., Gonz´ alez-Crespo, R., Corchado, J.M., Herrera-Viedma, E.: Solving multi-criteria group decision making problems under environments with a high number of alternatives using fuzzy ontologies and multigranular linguistic modelling methods. Knowl. Based Syst. 137, 54–64 (2017) 7. Distefano, S., Bruneo, D., Longo, F., Merlino, G., Puliafito, A.: Hospitalized patient monitoring and early treatment using IoT and cloud. BioNanoScience 7(2), 382–385 (2017) 8. Casado-Vara, R., Prieto, J., De la Prieta, F., Corchado, J.M.: How blockchain improves the supply chain: case study alimentary supply chain. Procedia Comput. Sci. 134, 393–398 (2018) 9. Firouzi, F., Rahmani, A.M., Mankodiya, K., Badaroglu, M., Merrett, G.V., Wong, P., Farahani, B.: Internet-of-Things and big data for smarter healthcare: from device to architecture, applications and analytics. Future Gener. Comput. Syst. 78, 583–586 (2018) 10. Han, Z., Niyato, D., Saad, W., Ba¸sar, T., Hjørungnes, A.: Game theory in Wireless and Communication Networks: Theory, Models, and Applications. Cambridge University Press, Cambridge (2012) 11. Casado-Vara, R., de la Prieta, F., Prieto, J., Corchado, J.M.: Blockchain framework for IoT data quality via edge computing. In: Proceedings of the 1st Workshop on Blockchain-Enabled Networked Sensor Systems, pp. 19–24. ACM, November 2018 12. Karthik, B.N., Parameswari, L.D., Harshini, R., Akshaya, A.: Survey on IOT & Arduino Based Patient Health Monitoring System (2018) 13. Kupriyanovsky, Y., et al.: Smart container, smart port, BIM, Internet things and blockchain in the digital system of world trade. Int. J. Open Inf. Technol. 6(3), 49–94 (2018) 14. Chamoso, P., Gonz´ alez-Briones, A., Rodr´ıguez, S., Corchado, J.M.: Tendencies of technologies and platforms in smart cities: a state-of-the-art review. Wireless Commun. Mobile Comput. 2018, 17 (2018) 15. Kumar, P.M., Gandhi, U.D.: A novel three-tier Internet of Things architecture with machine learning algorithm for early detection of heart diseases. Comput. Electr. Eng. 65, 222–235 (2018) 16. Casado-Vara, R., Novais, P., Gil, A.B., Prieto, J., Corchado, J.M.: Distributed continuous-time fault estimation control for multiple devices in IoT networks. IEEE Access 7, 11972–11984 (2019) 17. Minoli, D., Sohraby, K., Occhiogrosso, B.: IoT security (IoTsec) mechanisms for e-health and ambient assisted living applications. In: Proceedings of the Second IEEE/ACM International Conference on Connected Health: Applications, Systems and Engineering Technologies, pp. 13–18. IEEE Press, July 2017 18. Moinet, A., Darties, B., Baril, J.-L.: Blockchain based trust & authentication for decentralized sensor networks. arXiv preprint arXiv:1706.01730 (2017) 19. Gonzalez-Briones, A., Prieto, J., De La Prieta, F., Herrera-Viedma, E., Corchado, J.M.: Energy optimization using a case-based reasoning strategy. Sensors (Basel) 18(3), 865–865 (2018). https://doi.org/10.3390/s18030865

Cooperative Algorithm Healthcare

61

20. Parthasarathy, P., Vivekanandan, S.: A typical IoT architecture-based regular monitoring of arthritis disease using time wrapping algorithm. Int. J. Comput. Appl. 1–11 (2018) 21. Prieto Tejedor, J., Chamoso Santos, P., de la Prieta Pintado, F., Corchado Rodr´ıguez, J.M.: A generalized framework for wireless localization in gerontechnology. 17th IEEE International Conference on Ubiquitous Wireless Broadband ICUWB 2017. IEEE, September 2017 22. Singh, D., Tripathi, G., Alberti, A.M., Jara, A.: Semantic edge computing and IoT architecture for military health services in battlefield. In: Consumer Communications & Networking Conference (CCNC), 2017 14th IEEE Annual, pp. 185–190. IEEE, January 2017 23. Rodr´ıguez, S., Zato, C., Corchado, J.M., Li, T.: Fusion system based on multiagent systems to merge data from WSN. In: 2014 17th International Conference on Information Fusion (FUSION), pp. 1–8 (2014) 24. Shae, Z., Tsai, J.: On the design of a blockchain platform for clinical trial and precision medicine. In: International Conference on Distributed Computing Systems (ICDCS 2017), Atlanta (2017) 25. Casado-Vara, R., Vale, Z., Prieto, J., Corchado, J.: Fault-tolerant temperature control algorithm for IoT networks in smart buildings. Energies 11(12), 3430 (2018) 26. Casado-Vara, R., Chamoso, P., De la Prieta, F., Prieto, J., Corchado, J.M.: Nonlinear adaptive closed-loop control system for improved efficiency in IoT-blockchain management. Inf. Fusion 49, 227–239 (2019) 27. Varatharajan, R., Manogaran, G., Priyan, M.K., Sundarasekar, R.: Wearable sensor devices for early detection of Alzheimer disease using dynamic time warping algorithm. Cluster Comput. 21, 1–10 (2017) 28. Gazafroudi, A.S., Corchado, J.M., Kean, A., Soroudi, A.: Decentralized flexibility management for electric vehicles. IET Renew. Power Gener. 13, 952–960 (2019) 29. Yue, X., Wang, H., Jin, D., Li, M., Jiang, W.: Healthcare data gateways: found healthcare intelligence on blockchain with novel privacy risk control. J. Med. Syst. 40(10), 218 (2016) 30. Gazafroudi, A.S., Soares, J., Ghazvini, M.A.F., Pinto, T., Vale, Z., Corchado, J.M.: Stochastic interval-based optimal offering model for residential energy management systems by household owners. Int. J. Electr. Power Energy Syst. 105, 201–219 (2019) 31. Di Mascio, T., Vittorini, P., Gennari, R., Melonio, A., De La Prieta, F., Alrifai, M.: The learners’ user classes in the TERENCE adaptive learning system. In: 2012 IEEE 12th International Conference on Advanced Learning Technologies, pp. 572– 576. IEEE, July 2012 32. Chamoso, P., Rivas, A., Mart´ın-Limorti, J.J., Rodr´ıguez, S.: A hash based image matching algorithm for social networks. In: Advances in Intelligent Systems and Computing, vol. 619, pp. 183–190 (2018) 33. Garc´ıa, O., Chamoso, P., Prieto, J., Rodr´ıguez, S., De La Prieta, F.: A serious game to reduce consumption in smart buildings. In: Communications in Computer and Information Science, vol. 722, pp. 481–493 (2017) 34. Chamoso, P., de La Prieta, F., Eibenstein, A., Santos-Santos, D., Tizio, A., Vittorini, P.: A device supporting the self management of tinnitus. In: Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), LNCS, vol. 10209, pp. 399–410 (2017) 35. Rom´ an, J.A., Rodr´ıguez, S., de da Prieta, F.: Improving the distribution of services in MAS. In: Communications in Computer and Information Science, vol. 616 (2016)

62

R. Casado-Vara et al.

36. Buciarelli, E., Silvestri, M., Gonz´ alez, S.R.: Decision economics, in commemoration of the birth centennial of Herbert A. Simon 1916-2016 (Nobel Prize in Economics 1978). In: Distributed Computing and Artificial Intelligence, 13th International Conference. Advances in Intelligent Systems and Computing, vol. 475. Springer (2016) 37. Prieto, J., Mazuelas, S., Bahillo, A., Fernandez, P., Lorenzo, R.M., Abril, E.J.: Adaptive data fusion for wireless localization in harsh environments. IEEE Trans. Signal Process. 60(4), 1585–1596 (2012) 38. Mu˜ noz, M., Rodr´ıguez, M., Rodr´ıguez, M.E., Rodr´ıguez, S.: Genetic evaluation of the class III dentofacial in rural and urban Spanish population by AI techniques. In: Advances in Intelligent and Soft Computing, AISC, vol. 151 (2012) 39. Rodr´ıguez, S., Tapia, D.I., Sanz, E., Zato, C., De La Prieta, F., Gil, O.: Cloud computing integrated into service-oriented multi-agent architecture. In: IFIP Advances in Information and Communication Technology, AICT, vol. 322 (2010) 40. Chamoso, P., De La Prieta, F.: Simulation environment for algorithms and agents evaluation. In: ADCAIJ Adv. Distrib. Comput. Artif. Intell. J. 4(3), 87–96 (2015). (ISSN: 2255-2863)

Author Index

A Alemany, Jose, 1 Andrade, Jose Eduardo Reinoso, 15

J Jordán, Jaume, 5

B Baggetro, Alejandro Fuster, 15

M Márquez, Sergio, 27, 38 Muñoz, Marcel Vicente, 10

C Casado-Vara, Roberto, 27, 38, 49 Chimeno, Samuel Gallego, 10 Corchado, Juan M., 27, 38, 49

O Ospina, Óscar Mauricio Salazar, 10

D De la Prieta, Fernando, 49 del Val, Elena, 1 Diez, Carlos, 1 Domenech, Araceli Teruel, 15

P Palanca, Javier, 5 Prieto, Javier, 27, 38, 49

F Fernández, Joaquín Delgado, 10

R Ramón, Pablo Pueyo, 10 Rincon, Jaime, 5 Rodriguez, Sara, 49 Ruiz, Ramon, 1

G Giménez, Maite, 5 González, Carlos Peiró, 15 González-Briones, Alfonso, 27, 38

S Sánchez, Sergio Márquez, 10 Sanchis, Ana Gutiérrez, 20

H Hernández, Aarón González, 10

T Taverner, Joaquin, 1

© Springer Nature Switzerland AG 2020 S. Omatu et al. (Eds.): DCAI 2018, AISC 802, p. 63, 2020. https://doi.org/10.1007/978-3-030-00524-5