Wireless Sensor and Actuator Networks Technologies

364 Pages·2008·4.85 MB·English

by  Roberto Verdone

Preface

Aim of this book
Wireless sensor and actuator networks (WSANs) are among the most addressed research
fi elds in the area of information and communication technologies (ICT) these days, in the
US, Europe and Asia. WSANs are composed of possibly a large number of tiny, autono-
mous sensor devices and actuators equipped with wireless communication capabilities. One
of the most relevant aspects of this research fi eld stands in its multidisciplinarity and the
broad range of skills that are needed to approach their design. Theory of control systems
is involved, networking, middleware, application layer issues are relevant, joint considera-
tion of hardware and software aspects is needed, and their use can range from biomedical
to industrial or automotive applications, from military to civil environments, etc.
This book mainly covers wireless networking and design issues of WSANs with
applications.
This research fi eld attracted enormous and ever increasing attention in the past years.
However, by looking, for example, at the IEEE literature, the fi rst paper having ‘ wireless
sensor network ’ in the title in the online IEEE database of scientifi c papers Xplore, dates
back to the year 2000. A query on the ACM database brings us to the same outcome. So,
this is a new research fi eld that only very recently attracted the interest of many scien-
tists worldwide. On the other hand, the number of papers in the open literature increased
exponentially after the year 2000 (e.g. with a similar query IEEE Xplore shows 3 papers
for 2000, 20 in 2001, 34 in 2002, 98 in 2003, 289 in 2004, 622 in 2005, 952 in 2006):
this clearly testifi es to the relevance of the research fi eld on the one hand; on the other,
owing to the chaotic distribution of effort provided by thousands of separate research
groups worldwide, a consensus on major design rules of WSANs is still lacking, and it
is not unusual to fi nd recent papers using model assumptions which have been proven to
be not realistic by others. This book also aims at defi ning some general design rules for
WSANs and a common set of model assumptions that are real-world-proof. Some myths
will be destroyed.
Why a new book on WSANs
As anticipated by the title, this book covers aspects of WSANs, ranging from channel
modelling, transmission techniques, communication protocols, localization and sig-
nal processing issues. Some of these aspects have already been covered by previously
published books, by this and other publishers. The rationale for providing a new book on
WSANs is the following. The majority of available books provide extensive descriptions
xix
PPrreeffaaccee--PP337722553399..iinndddd xxiixx 1122//3311//22000077 1122::1199::4400 PPMM
xx Preface
of algorithms and protocols learnt from the literature, while minor relevance is given to
their performance evaluation, and to the description of tools, techniques and methodolo-
gies needed to set the most important parameters of such algorithms and protocols. For
instance, how the node density is related to the estimation error of a given spatial ran-
dom process, or how network reachability is linked to the lifetime of nodes. Moreover,
such books do not emphasize the practical issues related to the development of WSANs,
being mainly based on theoretical results published over the scientifi c literature.
This book intends to provide methodologies and tools to design WSANs for real appli-
cations, in real environments. It will describe algorithms and protocols, with the aim of
assessing their performance against the most relevant parameters. The design paradigm
used comes out from the traditional communication society approach: fi rst, derive charts
describing the link between performance metrics and system parameters, then, by fi xing
requirements on the former, the latter are set as design constraints. In other words, contrary
to best effort approaches, we design WSAN-related techniques starting from fi xed require-
ments (on network connectivity, or lifetime, on the desired precision of position estimates,
etc.). The tools that can be used to derive the performance metrics are basically of two types:
either mathematical or simulation approaches can be applied. In both cases proper model-
ling of basic aspects like, for example, wireless channel characterization and physical layer
techniques, is needed. This can be achieved from the literature, or by conducting on-the-fi eld
experiments based on available hardware/software platforms. In this book all these approaches
are used: many experimental results are reported, which are useful to the comprehension of
basic aspects related to the communications between wireless sensor nodes; also, a careful
selection of results taken from the available literature is presented. Many simulation outcomes
are discussed based on various types of simulation frameworks. Finally, the largest possi-
ble emphasis is given to the use of mathematical and formal descriptions of algorithms and
their performance. In fact, we believe that mathematical models allow best comprehension
of the relations between system parameters, provided that the models are suitably tuned to
the real world. As a result of this approach, the book contains many performance charts.
The general approach used in writing this book is thus oriented to algorithm and proto-
col design and performance assessment. This is testifi ed by the subtitle: ‘ Technologies,
analysis and design. ’ Another option was: ‘ From theory to practice ’ .
This book is for PhD students and researchers, who aim at creating a solid scientifi c
background about WSANs. It is also intended for engineers who need to design WSANs
and want to understand the basic rules underlying their performance. Even if less impor-
tance is given to an exhaustive description of the available literature, the table of con-
tents is also designed in order to provide a book useful for beginners.
About the contents of this book
Apart from the fi rst chapter, introducing the main defi nitions related to and features of
WSANs, the book is composed of three parts.
PPrreeffaaccee--PP337722553399..iinndddd xxxx 1122//3311//22000077 1122::1199::4400 PPMM
Preface xxi
The fi rst part covers fundamental issues. A description of some the most interesting appli-
cations for WSANs is given in Chapter 2. This part is taken from the outcomes of the
most relevant European projects on WSANs; moreover, a discussion on how these appli-
cations fall into few categories useful for partitioning the set of design guidelines is given.
Then Chapter 3 deals with an analysis of the characteristics of the wireless channel which
have an impact over link and network performance. The reason for starting with these
two chapters is in the usual approach that engineers apply when designing a telecom-
munication system, that is, building a protocol stack: applications, on top of the stack,
set requirements that drive the selection of protocols and transmission techniques; at the
other end, the wireless channel poses constraints to the communication capabilities and
performance. As a result of the encounter between the requirements set by applications
and the constraints posed by the wireless channel, the communication protocols and tech-
niques are selected. Therefore, after these two chapters we can start analyzing the link and
network performance of WSANs under several viewpoints. Chapters 4 and 5 deal with
network performance in terms of connectivity and lifetime. The impact of node density,
transmission ranges and transmit power levels on such aspects are discussed. This part is
not technology related. Then Chapter 6 introduces the transmission techniques (at physi-
cal and data link layer) that allow achievement of such transmission ranges and network
performance. Emphasis is given to IEEE802.15.4 standards, as they represent an almost
de facto standard for many applications of WSANs. This fi rst part reports many results
of experimental activities performed in order to measure wireless channel characteristics
at 2.4 GHz, the band used by IEEE802.15.4 and other transmission techniques like, for
example, Bluetooth, which are also suitable for some applications of WSANs. Then both
mathematical and simulation models are used to study network performance.
The second part of the book covers issues related to access control and routing (Chapter 7)
localisation and time synchronisation (Chapter 8) and signal processing (Chapter 9) for
WSANs. Many algorithms and protocols presented in the literature are reported and
described and some of them are evaluated through the use of suitable frameworks. This
part of the book provides more detailed insight on some of the key functionalities of
WSANs. The scientifi c literature is reviewed and some protocols and algorithms are
numerically evaluated. This part is based on theoretical analyses.
The fi nal part of this book (Chapter 10) is dedicated to some case studies: real-world
applications of WSANs which are the result of experimental activities performed by the
authors in the context of industrial contracts or large cooperative projects. The aim of
this part is to clarify how the design guidelines provided in previous chapters can be use-
ful when designing real-world networks. As usual when dealing with fi eld trials and in
general with experimental activities, the results are sometimes diffi cult to interpret and
discuss. However, this also provides suggestions on how to build a fi eld trial and how to
anticipate the behaviour of a network prototype. Mention of the industrial contexts of
the experimental activities is given as a necessary tribute to the availability of results.
In summary, the book tries to follow a path. First, the fundamentals are given in order to
establish a suffi cient technical background useful to the comprehension of basic aspects.
PPrreeffaaccee--PP337722553399..iinndddd xxxxii 1122//3311//22000077 1122::1199::4400 PPMM
xxii Preface
Since in this phase the pillars of the network are created, validation of assumptions is a
key issue, and experiments are conducted under such a view. The design of algorithms and
protocols can then be safely based on such pillars, and their performance evaluation can be
realized through the use of theoretical approaches. Finally, the techniques designed must
be proven on the real world, and the case studies discussed here report some examples.
Is it possible to skip some chapters?
Of course, it is. However, we dare to suggest you read the fi rst chapter, because it pro-
vides some clear introductory statements about WSANs, and the whole fi rst part, in
sequential order. Then the chapters of the second part might be read or not depending on
specifi c interests. Concerning the third part, we assume that if you bought the book then
you will read chapter 10.
In any case, all chapters are introduced by a short paragraph describing its scope and
content. Take a minute to read it before deciding whether to skip the chapter.
Acknowledgements
Some young scientists (PhD and Master students, or post-docs) directly contributed to this
book under the supervision of the authors; they merit explicit acknowledgement as the
book would have not been produced without their effort. Chiara Buratti and Flavio Fabbri
were very active in the preparation of Chapter 4; Chiara also took care of the fi rst draft of
Chapter 2. Moreever, both Chiara and Flavio assisted the authors with corrections, inclu-
sion of references, etc in the whole book. Chiara Taddia was involved in writing Chapter 7.
Virginia Corvino was in charge of part of Chapter 6. Enrica Salbaroli and Raffaele Rugin
were involved in writing Chapter 10. Thank you very much for your precious help.
Gratitude has also to be expressed to Alberto Zanella, senior researchers at WiLAB, who
provided useful comments for the realisation of the second part of the book.
The authors are also very thankful to John Orriss from the University of Manchester, UK,
who collaborated for years on connectivity issues with the group of authors, and gener-
ated some of the models used in this book. Davide Dardari and Andrea Conti would like
to thank Professor Marco Chiani as well as Professor Moe Z. Win and his staff from the
Massachusetts Institute of Technology (MIT), Cambridge, USA, for giving them a unique
opportunity to improve their expertise in advanced wireless systems and for the strong col-
laboration which is still fertile both from the professional and personal point of view.
Mention should also go to all those at WiLAB who indirectly provided inputs to this book
simply because they were involved in the past years in common research projects with
the authors, and as such contributed to the creation of a common scientifi c background
on WSANs at WiLAB: Velio Tralli, Gianni Pasolini, Andrea Giorgetti, Alessandro Bazzi,
Barbara Masini, and the younger ones.
PPrreeffaaccee--PP337722553399..iinndddd xxxxiiii 1122//3311//22000077 1122::1199::4411 PPMM
Preface xxiii
Finally, the authors wish to express their immense gratitude to Professor Oreste
Andrisano, head of WiLAB: since the beginning of the 1990s he started believing in the
potential of a small group of young scientists showing an enthusiastic approach to the
fi eld of wireless networks, including the authors of this book.
Authors ’ personal notes
When I started thinking about a book on wireless sensor and actuator networks and I had
the fi rst talks with Tim Pitts, commissioning editor of Elsevier, it was a long time ago.
Both my parents were alive and healthy. Now both are gone. This led to two consequences.
First, despite the continuous push from my colleagues and friends, Gianluca, Davide and
Andrea, whom I had invited since the beginning to collaborate in the preparation of the
book, it resulted in a signifi cant delay. However, it also had positive effects: many interest-
ing scientifi c results have been published in the last year and it was possible to include the
latest achievements; also, some relevant European projects dealing with wireless sensor
networks reached their maturity during this period, which means that part of their results
are reported in the book. Second, I want to thank who shined upon my life during the past
two years, my family, and I can’t refrain from dedicating this book to my Mother and my
Father who largely contributed to it, through me.
Roberto Verdone
I dedicate this book to my wife Paola, son Damiano and daughter Alessia for the support
and understanding they provided me with throughout the preparation of this manuscript.
A special thanks to my father from whom I inherited the enthusiasm for the wireless
world and my mother for her devotion and dedication.
Davide Dardari
Writing a book is a nice adventure. It is a trip that you know when it starts but you have no
idea when it will be fi nished. Writing a scientifi c book is more complex, simply because
researchers are working while you are writing, and it could become a race without a fi nish.
Nothing is possible without students who every day ask me new questions that force me
to study, friends that stimulate me with continuous discussion and family who support me
in any new trip. I dedicate this book to my students, to my friends and to my family.
Gianluca Mazzini
I dedicate this book to my family and to persons who have allowed me to grow and culti-
vate the passion for doing research.
Andrea Conti
PPrreeffaaccee--PP337722553399..iinndddd xxxxiiiiii 1122//3311//22000077 1122::1199::4411 PPMM
About the authors
This book is the outcome of the effort provided by several young and senior scientists
working in the area of WSANs for many years in Italy. Roberto Verdone and Davide
Dardari are with the University of Bologna, while Gianluca Mazzini and Andrea Conti
are with the University of Ferrara; the four authors cooperate in large projects within the
framework of WiLAB, the wireless communication laboratory located in Bologna and in
putting together scientists working in the fi eld of wireless systems. WiLAB is an organi-
zation born under the auspices of the University of Bologna, the National Research
Council and CNIT (the National Inter-University Consortium for Telecommunications).
The research experience accumulated by the authors in the fi eld of WSANs is the
result of several projects carried out both at national and European level at WiLAB.
Participation to either long-term research frameworks like the European networks of
excellence, NEWCOM and CRUISE, or more industry-oriented contexts, has provided
a range of skills that deal with both theoretical aspects and practical implementations.
Such diversifi ed experiences appear in the following chapters.
xxiv
AAbboouutt tthhee aauutthhoorrss--PP337722553399..iinndddd xxxxiivv 1122//3311//22000077 1122::3311::1122 PPMM
1
Introduction
1.1 Introduction
This chapter is the only one in this book intended for beginners who need to start from
scratch and learn what a wireless sensor and actuator network (WSAN) is and what are
its main characteristics. However, even among the category of scientists working in the
area of WSANs, sometimes statements like ‘ WSAN are a specifi c type of ad hoc net-
work ’ or ‘ Power control is an effi cient way to increase network lifetime ’ are found. They
are not true. If you want to learn why, read this short chapter. It will provide a defi ni-
tion of WSANs fi rst, and then the main features of WSANs will be highlighted. Some
practical aspects concerning energy management, a relevant matter for most WSAN
applications, will be discussed. Technical issues that research is still addressing are also
described. Finally, the research agendas of some fora that recently defi ned the goals of
research on embedded systems and WSANs for the next years are summarized, to ensure
better understanding of what are future expectations.
1.2 What is a WSAN?
A wireless sensor network (WSN) in its simplest form can be defi ned as (Chong & Kumar,
2003; Akyildiz, Su, Sankarasubramaniam & Cayirci, 2002; Culler, Estrin & Srivastava, 2004)
a network of (possibly low-size and low-complex) devices denoted as n odes that can sense
the envir onment and communicate the information gathered from the monitored fi eld (e.g.,
an area or volume) through wireless links; the data is forwarded, possibly via multiple hops
relaying, to a s ink (sometimes denoted as controller or monitor ) that can use it locally, or is
connected to other networks (e.g., the Internet) through a gateway. The nodes can be station-
ary or moving. They can be aware of their location or not. They can be homoge neous or not.
This is a traditional single-sink WSN (see Figure 1.1 ). Almost all scientifi c papers in
the literature deal with such a defi nition. This single-sink scenario suffers from the lack
of scalability: by increasing the number of nodes the amount of data gathered by the
sink increases and once its capacity is reached the network size can not be augmented.
1
CChh0011--PP337722553399..iinndddd 11 1122//3311//22000077 1122::3344::3388 PPMM
2 Wireless sensor and actuator networks
Other nets
Gateway
(e.g., Internet)
Sink/controller
Node
Figure 1.1 T raditional single-sink WSN
Moreover, for reasons related to medium access control (MAC) and routing aspects, net-
work performance cannot be considered independent from the network size.
Let us give a simple and approximate evaluation of the capacity of a single-sink WSN,
defi ned in terms of maximum number of nodes that can be attached to the sink. We con-
sider a WSN where nodes are requested to send their samples (composed of D bytes
each) taken from the monitored space every T seconds. Let us start by assuming that all
R
nodes can directly send their data to the sink (single-hop network with star topology).
1.2.1 Single-sink single-hop WSN
Denote as N the number of nodes, R the channel bit rate. Then we defi ne a factor, (cid:2) (cid:2) 1,
b A
taking account of the overhead introduced by all protocol stack layers: if S is the maxi-
A
mum data throughput measured at the application layer, it is given by S (cid:3) R (cid:4) (cid:2) . The
A b A
smaller (cid:2) , the lower is the throughput even if the channel bit rate is unchanged. All
A
protocol layers contribute to lower (cid:2) ; the MAC sub-layer is often the main contributor
A
when random channel access schemes are used. In modern communication systems (cid:2)
A
typically takes values between 0.5 and 0.1.
Under such assumptions, the application throughput will be approximately equal to N D
8/ T . Then, we reach the following inequality: N D 8/ T (cid:5) R (cid:2) ; therefore,
R R b A
N (cid:2) R (cid:2) T /(8D). (1.1)
b A R
This equation provides an approximate estimation of the number of nodes that can be part
of a single-sink single-hop WSN. To give a numerical example, assume R = 250 Kbit/s,
b
T (cid:3) 1 s, (cid:2) (cid:3) 0.1, D (cid:3) 3; then the maximum number of nodes is approximately 1000. On
R A
the other hand, if T (cid:3) 10 ms, then N can not exceed 10. It is clear that the requirements
R
CChh0011--PP337722553399..iinndddd 22 1122//3311//22000077 1122::3344::3399 PPMM
Introduction 3
Other nets
Gateway
(e.g., Internet)
Sink/controller
Node
Figure 1.2 Multi-sink WSN
set by the application scenario play a very relevant role when defi ning the capacity of
a single-sink WSN.
Note also that the protocol overhead can play a signifi cant role, through (cid:2) .
A
In the case discussed above, the N nodes are all within range of the sink. If the transmis-
sion range of links between sink and nodes is R , then the density of nodes is (no smaller
than) N / (cid:3) R2 .
1.2.2 Single-sink multi-hop WSN
If we now assume that the N nodes are distributed according to a smaller density, then some
of them must reach the sink through multiple hops. If a node can send its sample to the sink
through h hops, then the delivery of the data sample requires h transmissions. Let us denote
by h the average number of hops per data sample taken from the fi eld; if no smart reuse of
m
radio resources is introduced, then we have for a single-sink multi-hop WSN:
N(cid:5)R (cid:2) T /(8Dh ). (1.2)
b A R m
Therefore, the capacity of the network is reduced by a factor of h .
m
1.2.3 Multi-sink multi-hop WSN
A more general scenario includes multiple sinks in the network (see Figure 1.2 ). Given
a level of node density, a larger number of sinks will decrease the probability of isolated
clusters of nodes that cannot deliver their data owing to unfortunate signal propagation
conditions. In principle, a multiple-sink WSN can be scalable (i.e., the same perform-
ance can be achieved even by increasing the number of nodes), while this is clearly not
CChh0011--PP337722553399..iinndddd 33 1122//3311//22000077 1122::3344::3399 PPMM