83,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: Wiley - IEEE
- Sprache: Englisch
This book provides the most up-to-date research and development on wearable computing, wireless body sensor networks, wearable systems integrated with mobile computing, wireless networking and cloud computing
This book has a specific focus on advanced methods for programming Body Sensor Networks (BSNs) based on the reference SPINE project. It features an on-line website (http://spine.deis.unical.it) to support readers in developing their own BSN application/systems and covers new emerging topics on BSNs such as collaborative BSNs, BSN design methods, autonomic BSNs, integration of BSNs and pervasive environments, and integration of BSNs with cloud computing. The book provides a description of real BSN prototypes with the possibility to see on-line demos and download the software to test them on specific sensor platforms and includes case studies for more practical applications.
• Provides a future roadmap by learning advanced technology and open research issues
• Gathers the background knowledge to tackle key problems, for which solutions will enhance the evolution of next-generation wearable systems
• References the SPINE web site (http://spine.deis.unical.it) that accompanies the text
• Includes SPINE case studies and span topics like human activity recognition, rehabilitation of elbow/knee, handshake detection, emotion recognition systems
Wearable Systems and Body Sensor Networks: from modeling to implementation is a great reference for systems architects, practitioners, and product developers.
Giancarlo Fortino is currently an Associate Professor of Computer Engineering (since 2006) at the Department of Electronics, Informatics and Systems (DEIS) of the University of Calabria (Unical), Rende (CS), Italy. He was recently nominated Guest Professor in Computer Engineering of Wuhan University of Technology on April, 18 2012 (the term of appointment is three years). His research interests include distributed computing and networks, wireless sensor networks, wireless body sensor networks, agent systems, agent oriented software engineering, streaming content distribution networks, distributed multimedia systems, GRID computing.
Raffaele Gravina received the B.Sc. and M.S. degrees both in computer engineering from the University of Calabria, Rende, Italy, in 2004 and 2007, respectively. Here he also received the Ph.D. degree in computer engineering. He's now a Postdoctoral research fellow at University of Calabria. His research interests are focused on high-level programming methods for WSNs, specifically Wireless Body Sensor Networks. He wrote almost 30 scientific/technical articles in the area of the proposed Book. He is co-founder of SenSysCal S.r.l., a spin-off company of the University of Calabria, and CTO of the wearable computing area of the company.
Stefano Galzarano received the B.S. and M.S. degrees both in computer engineering from the University of Calabria, Rende, Italy, in 2006 and 2009, respectively. He is currently pursuing a joint Ph.D. degree in computer engineering with University of Calabria and Technical University of Eindhoven (The Netherlands). His research interests are focused on high-level programming methods for wireless sensor networks and, specifically, novel methods and frameworks for autonomic wireless body sensor networks.
Das E-Book können Sie in Legimi-Apps oder einer beliebigen App lesen, die das folgende Format unterstützen:
Seitenzahl: 358
Veröffentlichungsjahr: 2018
Ähnliche
Table of Contents
Cover
Title Page
Preface
Acknowledgments
1 Body Sensor Networks
1.1 Introduction
1.2 Background
1.3 Typical m‐Health System Architecture
1.4 Hardware Architecture of a Sensor Node
1.5 Communication Medium
1.6 Power Consumption Considerations
1.7 Communication Standards
1.8 Network Topologies
1.9 Commercial Sensor Node Platforms
1.10 Biophysiological Signals and Sensors
1.11 BSN Application Domains
1.12 Summary
References
2 BSN Programming Frameworks
2.1 Introduction
2.2 Developing BSN Applications
2.3 Programming Abstractions
2.4 Requirements for BSN Frameworks
2.5 BSN Programming Frameworks
2.6 Summary
References
3 Signal Processing In‐Node Environment
3.1 Introduction
3.2 Background
3.3 Motivations and Challenges
3.4 The SPINE Framework
3.5 Summary
References
4 Task‐Oriented Programming in BSNs
4.1 Introduction
4.2 Background
4.3 Motivations and Challenges
4.4 SPINE2 Overview
4.5 Task‐Oriented Programming in SPINE2
4.6 SPINE2 Node‐Side Middleware
4.7 SPINE2 Coordinator
4.8 SPINE2 Communication Protocol
4.9 Developing Application in SPINE2
4.10 Summary
References
5 Autonomic Body Sensor Networks
5.1 Introduction
5.2 Background
5.3 Motivations and Challenges
5.4 State‐of‐the‐Art
5.5 SPINE‐*: Task‐Based Autonomic Architecture
5.6 Autonomic Physical Activity Recognition
5.7 Summary
References
6 Agent‐Oriented Body Sensor Networks
6.1 Introduction
6.2 Background
6.3 Motivations and Challenges
6.4 State‐of‐the‐Art: Description and Comparison
6.5 Agent‐Based Modeling and Implementation of BSNs
6.6 Engineering Agent‐Based BSN Applications: A Case Study
6.7 Summary
References
7 Collaborative Body Sensor Networks
7.1 Introduction
7.2 Background
7.3 Motivations and Challenges
7.4 State‐of‐the‐Art
7.5 A Reference Architecture for Collaborative BSNs
7.6 C‐SPINE: A CBSN Architecture
7.7 Summary
References
8 Integration of Body Sensor Networks and Building Networks
8.1 Introduction
8.2 Background
8.3 Motivations and Challenges
8.4 Integration Layers
8.5 State‐of‐the‐Art: Description and Comparison
8.6 An Agent‐Oriented Integration Gateway
8.7 Application Scenarios
8.8 Summary
References
9 Integration of Wearable and Cloud Computing
9.1 Introduction
9.2 Background
9.3 Motivations and Challenges
9.4 Reference Architecture for Cloud‐Assisted BSNs
9.5 State‐of‐the‐Art: Description and Comparison
9.6 BodyCloud: A Cloud‐based Platform for Community BSN Applications
9.7 Engineering BodyCloud Applications
9.8 Summary
References
10 Development Methodology for BSN Systems
10.1 Introduction
10.2 Background
10.3 Motivations and Challenges
10.4 SPINE‐Based Design Methodology
10.5 Summary
References
11 SPINE‐Based Body Sensor Network Applications
11.1 Introduction
11.2 Background
11.3 Physical Activity Recognition
11.4 Step Counter
11.5 Emotion Recognition
11.6 Handshake Detection
11.7 Physical Rehabilitation
11.8 Summary
References
12 SPINE at Work
12.1 Introduction
12.2 SPINE 1.x
12.3 SPINE2
Index
End User License Agreement
List of Tables
Chapter 01
Table 1.1 List of commercial sensor node platforms.
Table 1.2 Summary of representative BSN systems.
Chapter 02
Table 2.1 BSN‐application development approaches comparison.
Table 2.2 Common tasks of
BSN
applications.
Table 2.3 Requirements for BSN frameworks.
Chapter 03
Table 3.1 API exposed by SPINE at the coordinator station.
Table 3.2 SPINE‐tested mobile personal devices.
Chapter 04
Table 4.1 SPINE2 application‐level messages.
Chapter 05
Table 5.1 Activity recognition accuracy affected by short faults over all channels and
C
= 3.
Table 5.2 Activity recognition accuracy affected by short faults over a specific channel and
C
= 3.
Table 5.3 Accuracy improvements over all channels and
C
= 3.
Chapter 06
Table 6.1 Comparison among agent‐oriented platforms (Agilla, ActorNet, AFME, and MAPS) for WSNs.
Chapter 08
Table 8.1 Sensing services of the BSN system for human activity recognition
Chapter 09
Table 9.1 Architectures for the integration of wireless sensor networks with cloud computing: a comparison.
Table 9.2 Architectures for the integration of body area networks with cloud computing: a comparison.
Chapter 11
Table 11.1 Posture/movement recognition accuracy.
Table 11.2 Stress threshold for HRV parameters.
List of Illustrations
Chapter 01
Figure 1.1 Common wearable sensors and their location on the human body.
Figure 1.2 A three‐tier hierarchical BSN architecture: (1) body sensor tier, (2) personal area network tier, and (3) global network tier.
Figure 1.3 Typical hardware architecture of a sensor node.
Figure 1.4 Peer‐to‐peer topology.
Figure 1.5 Star topology.
Figure 1.6 Mesh topology.
Figure 1.7 Clustered topology.
Figure 1.8 Mica Mote.
Figure 1.9 TelosB Tmote Sky.
Figure 1.10 Different revisions of the Shimmer platform.
Chapter 02
Figure 2.1 Reference model of a middleware‐based programming framework.
Chapter 03
Figure 3.1 The SPINE middleware architecture.
Figure 3.2 The SPINE
Node
software architecture.
Figure 3.3 The SPINE
Coordinator
software architecture.
Figure 3.4
Java desktop
implementation of the SPINE Management GUI (sensor‐node configuration dialog window).
Figure 3.5
Android
implementation of the SPINE Management GUI (sensor and function configuration dialog windows).
Figure 3.6 Data processing chain supported by the SPINE High‐level Data Processing plug‐in.
Figure 3.7 High‐Level Data Processing layered software architecture.
Chapter 04
Figure 4.1 The software layering approach in the SPINE2 middleware.
Figure 4.2 A task‐oriented application with tasks instantiated on different nodes.
Figure 4.3 Software architecture of the node‐side part of the framework.
Figure 4.4 The two‐layer protocol stack (a) and the packet fields (b).
Figure 4.5 The SPINE2 components interacting with the user applications.
Chapter 05
Figure 5.1 The multiplane autonomic architecture of a SPINE‐* application.
Figure 5.2 Examples of application with self‐configuring property; (a) the reconfiguration task is driven by the output results of the Processing task; (b) the reconfiguration task is driven by the desktop application running on the BSN coordinator.
Figure 5.3 Example of application with self‐healing property.
Figure 5.4 Example of application with self‐optimization property.
Figure 5.5 Example of application with self‐protection property.
Figure 5.6 The task‐based applications on the waist node (a) and on the thigh node (b).
Figure 5.7 The tested activities’ sequence.
Figure 5.8 The autonomic application running on the waist node.
Chapter 06
Figure 6.1 Architecture of MAPS.
Figure 6.2 Agent behavior model of MAPS.
Figure 6.3 Agent modeling of BSNs. (a) Master/slave model, (b) Master/slave + peer‐to‐peer model, and (c) Super peer model.
Figure 6.4 Architecture of the agent‐based activity monitoring system.
Figure 6.5 Finite state machine of the sensor agents: WaistSensorAgent and ThighSensorAgent.
Chapter 07
Figure 7.1 BSN infrastructures based on the logical communications among individuals and base stations.
Figure 7.2 The reference CBSN Network Architecture.
Figure 7.3 Activity diagram of basic CBSN operations.
Figure 7.4 CBSN Functional Architecture components.
Figure 7.5 C‐SPINE Functional Architecture components.
Figure 7.6 Inter‐BSN component interaction.
Chapter 08
Figure 8.1 An example of the Building Network environment.
Figure 8.2 The overall BMF framework architecture.
Figure 8.3 BN/BSN integration: a scenario.
Figure 8.4 BN/BSN integration layers.
Figure 8.5 BMF‐BN/SPINE‐BSN integration based on the gateway approach.
Figure 8.6 Class diagram of the agent‐based gateway.
Figure 8.7 Interaction between the agent‐based gateway (pair <BMFAgent, SPINEAgent>) and the BMF Coordinator.
Figure 8.8 Architecture of the in‐building physical activity recognition system.
Chapter 09
Figure 9.1 The cloud computing ecosystem.
Figure 9.2 Reference architecture for the integration of BSN and cloud computing.
Figure 9.3 The BodyCloud architecture.
Figure 9.4 Workflow schema of the BodyCloud approach for developing community BSN applications.
Figure 9.5 ECGMonitoring DataFeed modality.
Figure 9.6 ECGMonitoring GroupAnalysis modality.
Figure 9.7 EcgToRR workflow.
Figure 9.8 GUI view. (a) ECG wave plotting and (b) beat per minute instantaneous value.
Figure 9.9 CDRDetection DataFeed modality.
Figure 9.10 SingleCDRAnalysis modality.
Figure 9.11 SingleCDR workflow.
Figure 9.12 GUI view: detection of a CDR.
Figure 9.13 RehabMonitoring DataFeed modality.
Figure 9.14 Single RehabMonitoringAnalysis modality.
Figure 9.15 RehabMonitoring workflow.
Figure 9.16 GUI view: knee rehabilitation.
Figure 9.17 RawAccelerationDataFeed modality.
Figure 9.18 Single ActivityMonitoring Analysis modality.
Figure 9.19 ActivityMonitoring workflow.
Figure 9.20 GUI view: activity statistics.
Chapter 10
Figure 10.1 Architecture and function platforms.
Figure 10.2 Mapping of function and architecture.
Figure 10.3 Pattern architectural schemas: (a) Sensor Data Collection for Monitoring; (b) Multisensor Data Fusion for Detection/Classification of Events.
Figure 10.4 SPINE‐based Platform Design process schema.
Chapter 11
Figure 11.1 Two screenshots of the developed activity recognition Android app.
Figure 11.2 The main monitoring window of the stress detection system.
Figure 11.3 Block diagram of the proposed adaptive QRS detection algorithm.
Figure 11.4 The proposed CDR detection algorithm applied to RRi series.
Figure 11.5 A screenshot of the developed CDR detection mobile application.
Figure 11.6 The E‐Shake application.
Figure 11.7 The E‐Shake system architecture.
Figure 11.8 Two screenshots of the rehabilitation digital assistant.
Chapter 12
Figure 12.1 The task‐oriented application defined and deployed in “SPINE2SimpleTest.java” application.
Guide
Cover
Table of Contents
Begin Reading
Pages
iii
xi
xii
xiii
xiv
xv
xvi
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
20
21
22
23
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
239
240
Wearable Computing
From Modeling to Implementation of Wearable Systems Based on Body Sensor Networks
Giancarlo Fortino, Raffaele Gravina, and Stefano Galzarano
University of CalabriaRende, Italy
This edition first published 2018© 2018 John Wiley & Sons, Inc.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Giancarlo Fortino, Raffaele Gravina, and Stefano Galzarano to be identified as the authors of this work has been asserted in accordance with law.
Registered OfficeJohn Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
Editorial Office111 River Street, Hoboken, NJ 07030, USA
For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of WarrantyIn view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging‐in‐Publication Data
Names: Fortino, Giancarlo, 1971– author. | Gravina, Raffaele, 1982– author. | Galzarano, Stefano, 1984– author.Title: Wearable computing : from modeling to implementation of wearable systems based on body sensor networks / Giancarlo Fortino, Raffaele Gravina, Stefano Galzarano.Description: 1st edition. | Hoboken, NJ : John Wiley & Sons, 2018. | Includes bibliographical references and index. |Identifiers: LCCN 2017053912 (print) | LCCN 2017059016 (ebook) | ISBN 9781119078821 (pdf) | ISBN 9781119078838 (epub) | ISBN 9781118864579 (cloth)Subjects: LCSH: Wearable computers. | Sensor networks.Classification: LCC QA76.592 (ebook) | LCC QA76.592 .F67 2018 (print) | DDC 004.167–dc23LC record available at https://lccn.loc.gov/2017053912
Cover design by WileyCover images: © nopporn/Shutterstock; © Sergey Nivens/Shutterstock
Preface
Wearable computing is a relatively new area of research and development that aims at supporting people in different application domains: health care, fitness, social interactions, video games, and smart factory. Wearable computing is based on wearable sensor devices (e.g. to measure heart rate, temperature, or blood oxygen), common life objects (e.g. watch, belt, or shoes), and personal handheld devices (e.g. smartphones or tablets). Wearable computing has been recently boosted by the introduction of body sensor networks (BSNs), i.e. networks of wireless wearable sensor nodes coordinated by more capable coordinators (smartphones, tablets, and PCs).
In particular, BSNs enable a very wide range of application scenarios in different industry sectors. We can categorize them into different domains: e‐Health, e‐Emergency, e‐Entertainment, e‐Sport, e‐Factory, and e‐Social.
e‐Health applications span from early detection or prevention of diseases, elderly assistance at home, to post‐trauma rehabilitation after surgeries. e‐Emergency applications include BSN systems to support fire fighters, response teams in large‐scale disasters due to earthquakes, landslides, terrorist attacks, etc. e‐Entertainment domain refers to human–computer interaction systems typically based on BSNs for real‐time motion and gesture recognition. e‐Sport applications are related to the e‐Health domain, although they have a nonmedical focus. Specifically, this domain includes personal e‐fitness applications for amateur and professional athletes, as well as enterprise systems for fitness clubs and sport teams offering advanced performance monitoring services for their athletes. e‐Factory is an emerging and very promising domain involving industrial process management and monitoring, and workers’ safety and collaboration support. Finally, e‐Social applications may use BSN technologies to recognize user emotions and cognitive states to enable new forms of social interactions with friends and colleagues. An interesting example is given by a system that involves the interaction between two people’s BSNs to detect handshakes and, subsequently monitor their social and emotional interactions.
Although the basic elements (sensors, protocols, and coordinators) of a BSN are available (already from a commercial point of view), developing BSN systems/applications is a complex task that requires design methods based on effective and efficient programming frameworks. In this book, we will provide programming approaches and methods to effectively develop efficient BSN systems/applications. Moreover, we also provide new techniques to integrate BSN‐based wearable systems with more general Wireless Sensor Network systems and with Cloud computing.
This book, entitled Wearable Computing: From Modeling to Implementation of Wearable Systems Based on Body Sensor Networks, is based on an intense and extensive basic and applied research activity driven by the SPINE project (http://spine.deis.unical.it), whose authors are cofounders, responsible, and main developers. Thus, the book is connected to the SPINE website to provide readers with software and tools for the development of their wearable computing systems.
This book is aimed at a large audience in the Wearable Computing domain, that is gaining considerable research interest and momentum, and is expected to be of increasing interest to academic researchers and particularly to commercial developers. Upon reading this book the audiences will perceive the following benefits:
Learn the state‐of‐the‐art in research and development on wearable computing, wireless BSNs, wearable systems integrated with mobile computing, wireless networking, and cloud computing.
Obtain a future roadmap by learning advanced technology and open research issues.
Gather the background knowledge to tackle key problems, whose solutions will enhance the evolution of next‐generation wearable systems.
Use the book as a valuable reference for a technical professional in a related industry.
Use the book as a text book in the late undergraduate or the graduate level to prepare students who intend to perform research in the field of the book or intend to be employed in a related industry.
The main topics of the book are the following:
Wearable Computing
, the study or practice of inventing, designing, building, or using miniature body‐borne computational and sensory devices. Wearable computers may be worn under, over, or in clothing, or may also be themselves clothes.
Wireless Sensor Networks (WSNs)
, collections of tiny devices capable of sensing, computation, and wireless communication operating in a certain environment to monitor and control events of interest in a distributed manner and collectively react to critical situations. WSN applications span various domains such as environmental and building monitoring and surveillance, pollution monitoring, agriculture, health care, home‐automation, energy management, earthquake, and eruption monitoring.
Body Sensor Networks (BSNs)
, involving wireless wearable physiological sensors applied to the human body for medical and nonmedical purposes. In particular, they allow for the continuous measurement of body movements and physiological parameters, such as heart rate, muscular tension, skin conductivity, and breathing rate and volume, during the daily life of a user.
In‐node Signal Processing
, a central computing method in advanced wireless sensor platforms through which data processing is carried out directly on the sensor node to preprocess data acquired from sensors, to fuse data coming from other sensor nodes, and, notably, to perform higher level computation such as classification and decision making.
Mobile Computing
, human–computer interaction by which a computer is expected to be transported during normal usage. Mobile computing involves mobile communication, mobile hardware, and mobile software. Communication issues include ad‐hoc and infrastructure networks as well as communication properties, protocols, data formats, and concrete technologies. Hardware includes mobile devices or device components. Mobile software deals with the characteristics and requirements of mobile applications.
Cloud Computing
, the use of computing resources (hardware and software) that are delivered as a service over a network (typically the Internet). The name comes from the use of a cloud‐shaped symbol as an abstraction for the complex infrastructure it contains in system diagrams. Cloud computing entrusts remote services with a user’s data, software, and computation.
Platform‐Based Design (PBD)
, an embedded computing design methodology that consists of a sequence of design/development steps that leads the initial high‐level description of a digital system to its final implementation. Each step is a refinement process that transforms the design from a higher level description to a lower level description that is progressively closer to the final implementation.
Software Framework
, an abstraction in which software providing generic functionality can be selectively changed by user code, thus providing application‐specific software. A software framework is a universal, reusable software platform used to develop applications, products, and solutions. Software Frameworks include support programs, compilers, code libraries, an application programming interface (API), and tool sets that bring together all the different components to enable development of a project or solution.
Autonomic Computing
is a paradigm born as a response to the increasing complexity of managing computing systems. It faces the problem by introducing a series of self‐* properties (self‐configuration, self‐healing, self‐optimization, and self‐protection) into complex systems, through which such systems can be capable of performing several self‐management actions without any human intervention.
Activity Recognition
aims to recognize the actions and goals of one or more agents from a series of observations on the agents’ actions and the environmental conditions. Since the 1980s, this research field has captured the attention of several computer science communities due to its strength in providing personalized support for many different applications and its connection to many different fields of study such as medicine, human–computer interaction, or sociology. Specifically, we are mainly interested in sensor‐based single‐user and multiuser activity recognition that integrates the emerging area of sensor networks with novel data mining and machine learning techniques to model a wide range of human activities.
Specifically, this book is organized into 12 chapters:
Chapter 1
, Body Sensor Networks (BSNs), covers the state‐of‐the‐art about wearable sensor nodes, network architecture/protocols/standards, and applications/systems.
Chapter 2
, BSN Programming Frameworks, analyzes the state‐of‐the‐art about the most known software frameworks (CodeBlue, Titan, RehabSPOT, and others) for programming BSN applications/systems.
Chapter 3
, Signal Processing In‐Node Environment, describes in detail the SPINE framework (
http://spine.deis.unical.it
) from architectural and programming perspectives.
Chapter 4
, Task‐Oriented Programming, discusses task‐oriented programming of BSN applications through SPINE2.
Chapter 5
, Autonomic BSNs, illustrates how to make BSNs autonomic, by using SPINE*, an extension of SPINE2.
Chapter 6
, Agent‐oriented BSNs, presents the use of the Agent paradigm for programming BSN systems. Specifically, the MAPS (Mobile Agent Platform for SunSPOT) framework is used to design and implement agent‐based BSNs.
Chapter 7
, Collaborative BSNs, provides an introduction of methods and architectures to make BSNs interact with each other for supporting multiuser BSN applications.
Chapter 8
, Integration of BSNs and Wireless Sensor Networks, covers gateway‐based solution for interoperability between BSNs and infrastructural WSNs (e.g. building indoor sensor networks). This would enable “invisible” interaction between BSN‐worn people and the surrounding environment.
Chapter 9
, Integration of Wearable and Cloud Computing, presents an architecture for the integration of BSNs and the Cloud, called BodyCloud, based on Google App Engine. It is crucial now to move the data acquired or preprocessed on the human body to the cloud for storing and nonreal‐time analysis purposes.
Chapter 10
, Development Methodology for BSN Systems, describes a SPINE‐based methodology for the development of BSN systems. The methodology guides the BSN system developer from requirement analysis to implementation and deployment.
Chapter 11
, SPINE‐based BSN Applications, presents several applications developed through SPINE in different application domains (Activity Recognition: recognition of human postures and movements, Emotion Recognition: recognition of stress and fear, Handshake Detection: collaborative recognition of two people’s handshake, and Rehabilitation: real‐time computation of extension angles of elbow/knee).
Chapter 12
, SPINE at Work, provides a quick yet effective reference for BSN programmers interested in developing their applications using the SPINE framework. The chapter provides the necessary information for setting up the SPINE environment so as to start programming as well as insights on how the framework itself can be customized and extended.
Acknowledgments
This book is the result of direct and indirect involvement of many researchers, academics, and industry professionals.
We sincerely thank all the other members of the SPINE team: Fabio Bellifemine, Roberta Giannantonio, Antonio Guerrieri, Roozbeh Jafari, and Alessia Salmeri. Our gratitude also goes to all the international researchers and internal alumni that contributed to the SPINE Project with studies, programming efforts, and novel ideas; in particular let us remind Andrea Caligiuri, Giuseppe Cristofaro, Philip Kuryloski, Vitali Loseu, Ville‐Pekka Seppa, Edmund Seto, Marco Sgroi, and Filippo Tempia.
This work has been partially carried out under the framework of INTER‐IoT, Research and Innovation action – Horizon 2020 European Project, Grant Agreement 687283, financed by the European Union.
We thank Wiley’s publication staff for handling the book project and supporting its publication.
We hope that this book will serve as a valuable text for academic researchers and particularly to commercial developers working in the wearable computing domain.
1Body Sensor Networks
1.1 Introduction
This chapter provides an overview of the state‐of‐the‐art and technology in the field of wireless body sensor networks (BSNs). After introducing the motivations and the potential applications of this emerging technology, the chapter focuses on the analysis of the architecture of sensor nodes, communication techniques, and energy issues. We will then present and compare some of the programmable sensing platforms that are most commonly used in the context of wireless sensor networks (WSNs), and in particular those applied to remote monitoring of patients. The chapter also contains an analysis of relevant vital human signals and physical sensors used for their recording. Finally, the chapter presents the hardware/software characteristics that must be taken into consideration during the design stages of a healthcare monitoring system based on BSNs. For instance, important characteristics are sensor wearability, biocompatibility, energy consumption, security, and privacy of the acquired biophysical information.
1.2 Background
The widespread use of mobile applications for patient monitoring over the last few years is radically changing the approach to the health care. In today’s society, this is gaining an increasingly important role in the prevention of diseases; the convenience, for instance in terms of health‐care costs, is significant. The BSN technology makes often use of mobile applications that allow for the transmission to a coordinator node, such as a smartphone or a tablet, information about vital signs and physical activities (movements and gestures) [1, 2]. The miniaturization and the production cost reduction are leading to the realization of extremely small‐sized sensing and computing devices with high processing capacity thus giving a great impulse to the development of WSNs, and, as a direct consequence, of BSNs. Very heterogeneous information and diversified physiological signals can be transmitted, possibly after the application of sensor fusion techniques [3], by the sensor nodes to the coordinator device.
Figure 1.1 shows a number of wearable sensing devices and their typical location on the body:
Electrocardiography
(ECG): the ECG is used to record the electrical activity (including the heart rate) of the heart over a period of time using electrodes placed on the skin.
Blood pressure meter
: also known as sphygmomanometer, it is a device used to measure (typically, both diastolic and systolic) blood pressure.
Pulse oximetry
: the oximeter is a medical device that allows us to measure noninvasively the amount of hemoglobin in the blood. Since hemoglobin binds with oxygen, it is therefore possible to obtain an estimate of the amount of oxygen present in the blood.
Electromyography
(EMG): the EMG sensor is used to monitor muscle activity, using a needle electrode inserted into the muscle for high accuracy, or, more practical and noninvasive, with simple skin electrodes. It records the activity of the muscle fibers under different conditions: at rest, during voluntary contraction up to the maximum effort, and during a sustained average contraction.
Electroencephalography
(EEG): the EEG sensor uses electrodes placed on the scalp to monitor the brain activity and capture different types of brain waves.
Motion
inertial sensors (e.g. accelerometers and gyroscopes) monitor human movements and even gestures.
Figure 1.1 Common wearable sensors and their location on the human body.
BSN systems are commonly characterized by a number of hardware and software requirements:
Interoperability
: it is necessary to ensure the continuous data transfer through different standards (e.g. Bluetooth and ZigBee) to promote the exchange of information and ensure interaction between devices. In addition, it should provide an adequate level of scalability in relation to the number of sensor nodes and the workload of the BSN.
System device
: the sensors must be of low complexity, small size, lightweight, energy efficient, easy to use, and reconfigurable. In addition, patient biosignal storage, retrieval, visualization, and analysis must be facilitated.
Security
at the device and system level: particular attention must be paid to secure transmission and authenticated access to such sensible data.
Privacy
: the BSN could be considered as a “threat” to the freedom of the individual, if the purpose of the applications goes “beyond” the medical purposes. Social acceptance to these systems is the key to their wider dissemination.
Reliability
: the whole system must be reliable at hardware, network, and software levels. Reliability affects directly the quality of monitoring because, in the worst case, the failure to observe and/or successfully notify a “critical risk event” can be lethal for the patient. Because of the limitations and requirements on communication and power consumption, the reliability techniques used in traditional networks are not easily applicable in the BSN domain and, both at the design and implementation phase, this must be taken seriously.
Validation and accuracy
of sensory data: sensing devices are subject to hardware constraints that can affect the quality of the acquired data; both wired and wireless connections are not always reliable; environmental interference and limited energy availability also affect this aspect. This can cause inconsistencies in the transmitted data and might lead to critical errors in their interpretation. It is very important that all data transmitted from the sensor nodes to the coordinator are adequately “validated” either in hardware or software, trying to identify the “critical points” of the system.
Data consistency
: for large‐scale BSNs, with many and heterogeneous sensors, a single biophysical phenomenon may be “fragmented” and only partially detectable into individual signals. This aspect arises problems of information consistency, which must be addressed through appropriate synchronization strategies, data fusion techniques [3], and/or mutual exclusion in the access to data.
Interference
: wireless links used in the BSN should try to minimize the interference issues and favor the coexistence of sensor nodes with other network devices available within the radio range.
Biological compatibility
: the wearable sensors and skin electrodes must be biocompatible and stable, as they might operate on the user for a long period of time without interruptions.
In addition to the hardware and software features, we highlight some aspects that could encourage the wide diffusion and exploitation of BSN systems:
Costs
: users expect low costs for health monitoring, yet preserving high performance of the devices used.
Different levels of monitoring
: users may require different levels of monitoring, for example, to control the risk of ischemic heart disease or of falling during movements. Depending on the operating mode, the energy level required for the power supply of the devices can also vary.
Noninvasive easy‐to‐use devices
: the devices must be wearable, lightweight, and noninvasive. They should not hinder users in their daily activities; their operation must be “transparent” to users who should ignore the details of the monitoring task.
Consistent performance
: sensors must be calibrated and accurate, and they should provide consistent measurements even if the BSN is stopped and restarted several times. Wireless links should be as robust as possible and be able to operate correctly in different (noisy) working environments.
1.3 Typical m‐Health System Architecture
Figure 1.2 shows the typical architecture of an m‐Health system based on BSN technology. It usually consists of three different tiers communicating through wireless (or sometimes wired) channels [4].
Figure 1.2 A three‐tier hierarchical BSN architecture: (1) body sensor tier, (2) personal area network tier, and (3) global network tier.
Tier 1 represents the Body Sensor Tier and includes a set of wireless wearable medical sensor nodes composing the BSN. Each node is able to detect, sample, and process one or more physiological signals. For example, a motion sensor for discriminating postures, gestures, and activities; an electrocardiogram (ECG) sensor can be used for monitoring cardiac activity; and an electroencephalogram (EEG) sensor for monitoring cerebral electrical activity, and so on.
Tier 2 is the Personal Area Network Tier and contains the personal coordinator device (often a smartphone or a tablet, but possibly a PC) running an end‐user application. This tier is responsible for a number of functions providing a transparent interface to the BSN, to the user, and to the upper tier. The interface to the BSN provides functionalities to configure and manage the network, such as sensor discovery and activation, sensory data recording and processing, and establishment of a secure communication with both Tier 1 and Tier 3. When the BSN has been configured, the end‐user monitoring application starts providing feedback through a user‐friendly graphical and/or audio interface. Finally, if there is an active channel of communication with the upper tier, it can report raw and processed data for off‐line analysis and long‐term storage. Conversely, if Internet connectivity is temporary unavailable, the coordinator device should be able to store the data locally and perform the data transfer as soon as the connectivity is restored.
Tier 3 is the Global Network Tier and comprises one or more remote medical servers or a Cloud computing platform. Tier 3 usually provides services to medical personnel for off‐line analysis of a patient’s health status, real‐time notification of life‐critical events and abnormal conditions, and scientific and medical visualization of collected data. In addition, this tier can provide a web interface for the patient itself and/or relatives too.
1.4 Hardware Architecture of a Sensor Node
A typical sensor node architecture is shown in Figure 1.3 and consists of the following main components:
Sensing
unit, each node usually includes one or multiple built‐in sensors and an expansion bus through which it is possible to attach further sensors that might be necessary for specific applications. A sensor is generally composed of a transducer and an analog‐to‐digital converter (see next bullet point). The transducers are realized by exploiting the characteristics of some materials that vary their “electrical properties” to varying environmental conditions. Many transducers used on wireless sensor nodes are based on MEMS (Micro‐ElectroMechanical Systems) technology. MEMS sensors are more efficient and require less power consumption with respect to piezoelectric sensors; furthermore, MEMS sensors are characterized by low production costs, although this could lead to less precision if compared with piezoelectric sensors.
Analog‐to‐Digital Converter
(ADC) converts the voltage value of a transducer into a digital value, which will then be used for post‐processing.
Processing
unit, the Micro‐Controller Unit (MCU) of a sensor node is usually associated with a built‐in limited memory unit to improve the processing speed and enable local online sensory data processing. The sensor node is, therefore, able to perform signal processing such as “background noise” filtering, data fusion and aggregation, and feature extraction (e.g. mean, variance, maximum/minimum value, entropy, and signal amplitude/energy). The MCU is also responsible for the management of the other hardware resources.
Transceiver
unit is the component that connects the node to the network. It can be an optical or a radio frequency (RF) device. It is also possible, and actually very useful, to use the radio with a low duty‐cycle, to help reducing the power consumption.
External memory
is needed to store the binary code of the program running on the sensor node. Some sensor platforms also include a further memory (usually a microSD flash memory) as a mass storage unit for sensory data recording.
Power supply
is the scarcest resource of a sensor node and must be preserved as much as possible to prolong its lifetime; it could be notably supported by a unit for energy harvesting (e.g. from solar light, heat, or vibration).
Figure 1.3 Typical hardware architecture of a sensor node.
1.5 Communication Medium
In a multi‐hop sensor network the nodes can interact with each other via a wireless communication medium. One choice is to use the ISM (industrial, scientific, and medical) radio spectrum [5], i.e. a predefined set of frequency bands that can be used freely in many countries. Most of the sensors currently on the market do in fact make use of a RF circuit. Another option is given by infrared (IR) communication. On the one hand, the IR communication does not require permits or licenses, it is protected from interference, and IR transceivers are very cheap and easy to realize. On the other hand, however, IR requires line‐of‐sight between the transmitter and the receiver, which makes it hardly usable for WSNs and BSNs as nodes very often cannot be deployed in such a way.
1.6 Power Consumption Considerations
A sensor node is normally equipped with a very limited energy source. The lifecycle of a sensor node heavily depends on the battery dimensions and on the processing and communication duty‐cycling. For these reasons, many research efforts are focusing on the design of power‐aware communication protocols and algorithms, with the aim of optimizing energy consumption. While in traditional mobile networks and ad‐hoc networks energy consumption is not the most important constraint, in the WSN domain it is a crucial aspect. This is true even in the specific subdomain of the BSNs. Although it is generally easier to recharge or replace the batteries of the wearable nodes, due to wearability reasons, the battery dimension (and hence its capacity) is generally much smaller than in other WSN scenarios.
In a sensor node, the energy consumption is mainly due to three tasks:
Communication
: it is the most affecting factor. Low‐power radios, strict radio duty‐cycling, power‐aware WSN‐specific communication protocols and standards, and on‐node data fusion and aggregation techniques are critical design choices for reducing the activation of the transceiver module as much as possible. It is worth noting that both transmission and listening/reception time must be optimized.
Sensing
: the power required to carry out the sampling depends on the nature of the application and, as a consequence, on the type of the physical transducers involved.
Data processing
: it must be taken into account, even though the energy consumed for processing a given amount of data is very small compared to the energy requirements for transmitting the same amount of data. Experimental studies showed that the energy cost for transmitting 1 kB of data is about the same that would be obtained by performing 3–100 million instructions on the sensor node microcontroller [6].
1.7 Communication Standards
The aforementioned requirements impose very tight restrictions on the type of network protocols that can be used in WSNs. The short‐range wireless technologies are a prerequisite, given the limited power budget available for each node. The implementation of a wireless network communication protocol that must be robust, fault tolerant, and capable of self‐configuration even in hostile environments represents a considerable technological challenge, which required (and still requires) the efforts of several standardization bodies, such as IEEE and IETF.
The IEEE 802.15.4 [7] is to date the most widely adopted standard in the WSN domain. Indeed, it is intended to offer the fundamental lower network layers (physical and MAC) of Wireless Personal Area Networks (WPANs) focusing on low‐cost, low‐speed ubiquitous communication between devices. The emphasis is on very low‐cost communication of nearby devices with little to no underlying infrastructure. The basic protocol conceives a 10 m communication range with a transfer rate of 250 kbit/s. Tradeoffs are possible to favor more radically embedded devices with even lower power requirements, through the definition of several physical layers. Lower transfer rates of 20 and 40 kbit/s were initially defined, with the 100 kbit/s rate being added later. Even lower rates can be considered with the resulting effect on power consumption. The main identifying feature of 802.15.4 is the importance of achieving extremely low manufacturing and operation costs, and technological simplicity, without sacrificing flexibility or generality. Important features include real‐time suitability by reservation of guaranteed time slots, collision avoidance through CSMA/CA, and integrated support for secure communications. It operates on one of three possible unlicensed frequency bands:
868.0–868.6 MHz: Europe, allows 1 communication channel.
902–928 MHz: North America, up to 30 channels.
2400–2483.5 MHz: Worldwide use, up to 16 channels.
To complete the IEEE 802.15.4 standard, the ZigBee [8] protocol has been realized. ZigBee is a low‐cost, low‐power, wireless mesh network standard built upon the physical layer and medium access control defined in the 802.15.4. It is intended to be simpler and less expensive than, for instance, Bluetooth. ZigBee chip vendors typically sell integrated radios and microcontrollers with 60 to 256 kB flash memory. The ZigBee network layer natively supports both star and tree networks, and generic mesh networks. Every network must have one coordinator device. In particular, within star networks, the coordinator must be the central node. Specifically, the ZigBee specification completes the 802.15.4 standard by adding four main components:
Network layer
, which enables the correct use of the MAC sublayer and provides a suitable interface for the application layer.
Application layer
is the highest‐level layer defined by ZigBee and represents the interface to the end users.
ZigBee device object
(ZDO) is the protocol responsible for overall device management, security keys, and policies. It is responsible for defining the role of a device (i.e. coordinator or end device).
Manufacturer‐defined application objects
, which allow for customization and favor total integration.
Bluetooth
