Download the full version of the ebook at ebookfinal.

com

Wireless sensor networks deployment strategies for

outdoor monitoring First Edition Fadi Al-Turjman

https://ebookfinal.com/download/wireless-sensor-networks-
deployment-strategies-for-outdoor-monitoring-first-edition-
fadi-al-turjman/

OR CLICK BUTTON

DOWNLOAD EBOOK

Download more ebook instantly today at https://ebookfinal.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Protocols and Architectures for Wireless Sensor Networks

1st Edition Holger Karl

https://ebookfinal.com/download/protocols-and-architectures-for-
wireless-sensor-networks-1st-edition-holger-karl/

ebookfinal.com

Dynamic Wireless Sensor Networks 1st Edition Sharief M.A.

Oteafy

https://ebookfinal.com/download/dynamic-wireless-sensor-networks-1st-
edition-sharief-m-a-oteafy/

ebookfinal.com

Topology Control in Wireless Ad Hoc and Sensor Networks

1st Edition Paolo Santi

https://ebookfinal.com/download/topology-control-in-wireless-ad-hoc-
and-sensor-networks-1st-edition-paolo-santi/

ebookfinal.com

Wireless Ad Hoc and Sensor Networks Management Performance

and Applications 1st Edition He

https://ebookfinal.com/download/wireless-ad-hoc-and-sensor-networks-
management-performance-and-applications-1st-edition-he/

ebookfinal.com
Wireless Sensor Networks From Theory to Applications 1st
Edition Ibrahiem M. M. El Emary (Editor)

https://ebookfinal.com/download/wireless-sensor-networks-from-theory-
to-applications-1st-edition-ibrahiem-m-m-el-emary-editor/

ebookfinal.com

Security in Sensor Networks 1st Edition Yang Xiao

https://ebookfinal.com/download/security-in-sensor-networks-1st-
edition-yang-xiao/

ebookfinal.com

Sensor Technologies for Civil Infrastructures Applications

in Structural Health Monitoring 1st Edition Ming L. Wang

https://ebookfinal.com/download/sensor-technologies-for-civil-
infrastructures-applications-in-structural-health-monitoring-1st-
edition-ming-l-wang/
ebookfinal.com

Hacking Wireless Networks For Dummies 1st ed Edition Kevin

Beaver

https://ebookfinal.com/download/hacking-wireless-networks-for-
dummies-1st-ed-edition-kevin-beaver/

ebookfinal.com

Distributed Sensor Networks Image and Sensor Signal

Processing Volume One 2nd Edition S. Sitharama Iyengar
(Editor)
https://ebookfinal.com/download/distributed-sensor-networks-image-and-
sensor-signal-processing-volume-one-2nd-edition-s-sitharama-iyengar-
editor/
ebookfinal.com
Wireless Sensor
Networks
 Wireless Sensor
Networks
Deployment Strategies for
Outdoor Monitoring

Fadi Al-Turjman
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks
does not warrant the accuracy of the text or exercises in this book. This book’s use or discussion of
MATLAB® software or related products does not constitute endorsement or sponsorship by The
MathWorks of a particular pedagogical approach or particular use of the MATLAB® software.

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2018 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed on acid-free paper

International Standard Book Number-13: 978-0-8153-7581-4 (Hardback)

This book contains information obtained from authentic and highly regarded sources.
Reasonable efforts have been made to publish reliable data and information, but the
author and publisher cannot assume responsibility for the validity of all materials or the
consequences of their use. The author and publisher have attempted to trace the copyright
holders of all material reproduced in this publication; we apologize to copyright holders if
permission to publish in this form has not been obtained. If any copyright material has not
been acknowledged, please write and let us know so we may rectify this in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted,
reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means
now known or hereafter invented, including photocopying, microfilming, and recording, or in
any information storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.
copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc.
(CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a nonprofit organization
that provides licenses and registration for a variety of users. For organizations that have been
granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks

and are used only for identification and explanation without intent to infringe.

Library of Congress Cataloging‑in‑Publication Data

Names: Al-Turjman, Fadi, author.

Title: Wireless sensor networks: deployment strategies for outdoor
monitoring / Fadi Al-Turjman.
Description: Boca Raton: Taylor & Francis, a CRC title, part of the Taylor &
Francis imprint, a member of the Taylor & Francis Group, the academic
division of T&F Informa, plc, [2017] | Includes bibliographical references.
Identifiers: LCCN 2017040978 | ISBN 9780815375814 (hb : acid-free paper) |
ISBN 9781351207034 (e)
Subjects: LCSH: Wireless sensor networks. | Environmental monitoring.
Classification: LCC TK7872.D48 A3953 2017 | DDC 006.2/5--dc23
LC record available at https://lccn.loc.gov/2017040978

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com
 This book is dedicated to my beloved parents, my mother and
my father, to honor their major support and encouragement
over the course of my entire life. In this way, I wish to
express maybe a little part of my great gratitude to them.
My mother, Lama, thank you very much for the tender care
you have been always providing. Your delicious morning
sandwiches’ taste in the cold winter is still in my mouth.
Great mother, words can never express the deepest gratitude I have
for you. You have been there for me my whole life and I love you so
much for it. I mostly admire and love you for the person YOU are.
My father, Ferzat, completed his BSc and MSc degrees in
mathematics at Damascus University, Damascus, Syria, in 1974.
He spent his entire life teaching not only math but also ethics.
Throughout his career, Mr. Al-Turjman has made significant
contributions in multiple areas of the education system. He
received several academic honors for his outstanding teaching,
including the Prince of Kuwait Prize for Excellence in Teaching
twice, in 2000 and 2010. Indeed, he is a brilliant teacher and
scientific communicator. I owe him a lot, and I believe his
support was a key factor in any achievement I have ever made.
Thank you very much great father for your sincere love.
I love you always,
Your son,
Fadi Al-Turjman
Contents

P R E FA C E xi
AUTHOR xiii

CHApTER 1 INTRODUCTiON 1
1.1 Contributions 2
1.2 Book Outline 4
References 6

CHApTER 2 D E p L OY M E N T O F W i R E L E S S S E N S O R
N E T wO RKS i N O U T D O O R E N Vi RO N M E N T
MONiTORiNG: A N OVERViE w 7
2.1  esired Network Properties in OEM
D 8
2.1.1 Connectivity 8
2.1.2 Fault-Tolerance 9
2.1.3 Lifetime 10
2.2 R andom vs. Deterministic WSNs Deployment 11
2.2.1 R andom Deployment 12
2.2.2 Deterministic (Grid-Based) Deployment 12
2.3 Summary 16
References 18

CHApTER 3 E F F i C i E N T D E p L OY M E N T O F W i R E L E S S
S E N S O R N E T w O R K S TA R G E T i N G E N V i R O N M E N T
M O N i T O R i N G A pp L i C AT i O N S 21
3.1 R
 elated Work 25
3.2 S ystem Models and Problem Definition 27
3.2.1 Network Model and Placement Problem 28

VII
VIII C O N T EN T S

3.2.2 Cost and Communication Models 30

3.2.3 Lifetime Model 31
3.3 Deployment Strategy 38
3.3.1 First Phase of the O3DwLC Strategy 39
3.3.2 Second Phase of the O3DwLC Strategy 42
3.4 Performance Evaluation 52
3.4.1 Simulation Model 53
3.4.2 Simulation Results 54
3.5 Conclusion 59
References 60

CHApTER 4 O p T i M i Z E D R E L AY P L A C E M E N T F O R W i R E L E S S
S E N S O R N E T w O R K S F E D E R AT i O N i N
E N V i R O N M E N TA L A pp L i C AT i O N S 63
4.1 R
 elated Work 66
4.2 O ptimized WSN Federation 67
4.2.1 Definitions and Assumptions 67
4.2.2 Deployment Strategy 68
4.2.2.1 G
 rid-Based ORP (GORP) 68
4.2.2.2 Th
 e General Non-Grid (ORP) 74
4.3 Performance Evaluation 75
4.3.1 Simulation Environment 75
4.3.2 Performance Metrics and Parameters 75
4.3.3 Baseline Approaches 77
4.3.4 Simulation Model 78
4.3.5 Simulation Results 78
4.4 Conclusion 82
References 82

CHApTER 5 TO wA R D S A U G M E N T i N G F E D E R AT E D W i R E L E S S
85
S E N S O R N E T w O R KS i N F O R E S T RY A pp L i CAT i O N S
5.1  ackground and Related Work
B 87
5.2 System Model 88
5.2.1 Problem Definition 88
5.2.2 Communication Model 89
5.2.3 Network Model 90
5.2.4 Grid Model 91
5.3 Fixing Augmented Network Damage Intelligently
(FADI): The Approach92
5.4 Performance Evaluation 99
5.4.1 Performance Metrics and Parameters 100
5.4.2 Baseline Approaches 100
5.4.3 Simulation Setup and Results 101
5.5 Conclusion 104
References 105
 C O N T EN T S IX

CHApTER 6 O p T i M i Z E D H E X A G O N - B A S E D D E p L OY M E N T F O R
L A R G E -S C A L E U bi Q U i T O U S S E N S O R N E T w O R K S 107
6.1 Related Work 109
6.2 System Models 111
6.2.1 N  etwork Model 111
6.2.2 E  nergy Consumption Model 112
6.2.3 C  ommunication Model 112
6.2.4 Cost Model 113
6.2.4.1 Relay Node 113
6.2.4.2 C  ognitive Node 114
6.2.5 Problem Definition 116
6.3 The O2D Deployment Strategy 117
6.4 Case Study 125
6.5 Simulation Results and Discussion 128
6.5.1 Simulation Setup 131
6.5.2 Evaluation of the Square-Based Grid 131
6.5.2.1 N  ode Reliability (NR) 132
6.5.2.2 I nstantaneous Throughput (IT) 132
6.5.2.3 Delay 132
6.5.3 Evaluation of the Hexagon-Based Grid 134
6.5.3.1 I nstantaneous Throughput (IT) 135
6.5.3.2 N  ode Reliability (NR) 136
6.6 Conclusions 138
References 139

CHApTER 7 TO wA R D S P R O L O N G E D L i F E T i M E F O R
D E p L OY E D W i R E L E S S S E N S O R N E T w O R K S iN

OUTDOOR ENViRONMENT MONiTORiNG 143

7.1 R
 elated Work 145
7.1.1 Contributions 149
7.2 S ystem Models 150
7.2.1 Communication Model 151
7.2.2 Network Model 151
7.2.3 Lifetime and Energy Models 153
7.3 Deployment Strategy 154
7.3.1 First Phase of the O3D Strategy 155
7.3.2 Second Phase of the O3D Strategy 158
7.4 Lifetime Theoretical Analysis 166
7.5 P erformance Evaluation and Discussion 168
7.5.1 Simulation Model 169
7.5.2 Simulation Results 169
7.6 Conclusions 175
References 176
X C O N T EN T S

CHApTER 8 PAT H P L A N N i N G F O R M O bi L E D ATA

C O L L E C T O R S i N FU T U R E C i T i E S 181
8.1 Related work 184
8.2 System Models 185
8.2.1 N  etwork Model 186
8.2.2 Energy Model 187
8.2.3 Communication Model 187
8.3 Hybrid Genetic-based Path Planning (HGPP)
Approach187
8.3.1 Chromosome Representation 188
8.3.2 Initial Population Creation 188
8.3.3 Path Planning 188
8.4 Performance Evaluation 190
8.4.1 Experimental Setup 190
8.4.2 Performance Metrics and Parameters 190
8.4.3 Simulation Results 191
8.5 Conclusions 197
References 199

CHApTER 9 C ON CLUSiONS AND FU T U R E D i R E C T i O N S 201

9.1 Summary 202
9.2 Future Work 203
INDEX 205
Preface

Wireless Sensor Networks (WSNs) overcome the difficulties of other

monitoring systems because they require no human attendance on
site, provide real time interaction with events, and maintain efficient
cost and power operations. However, further efficiencies are required
especially in the case of Outdoor Environment Monitoring (OEM)
applications due to their harsh operational conditions, huge targeted
areas, limited energy budgets, and required three-dimensional (3-D)
setups. A fundamental issue in defeating these practical challenges is
the deployment planning of the WSNs. The deployment plan is a key
factor of many intrinsic properties of OEM networks, summarized
in connectivity, lifetime, fault-tolerance, and cost-effectiveness. In
this book, we investigate the problem of WSNs’ deployments which
address these properties in order to overcome the unique challenges
and circumstances in OEM applications.

XI
X II P REFAc E

MATLAB® is a registered trademark of The MathWorks, Inc. For

product information, please contact:

The MathWorks, Inc.

3 Apple Hill Drive
Natick, MA, 01760-2098 USA
Tel: 508-647-7000
Fax: 508-647-7001
E-mail: info@mathworks.com
Web: www.mathworks.com
Author

Fadi Al-Turjman  is an associate professor at

Middle East Technical University Northern
Cyprus Campus, Turkey. He received his PhD
degree in computing science from Queen’s
University, Kingston, Ontario, Canada, in
2011. He is a leading authority in the areas of
­smart/­cognitive, wireless, and mobile networks’
architectures, protocols, deployments, and per-
formance evaluation. His record spans more
than 150 publications in journals, conferences, patents, books, and
book chapters, in addition to numerous keynotes and plenary talks
at flagship venues. He has received several recognitions and best
papers’ awards at top international conferences and led a number
of international symposia and workshops in flagship ComSoc con-
ferences. He is serving as the lead guest editor in several journals,
including the IET Wireless Sensor Systems, Sensors (Multidisciplinary
Digital Publishing Institute), IEEE Access, and Wireless Communication
and Mobile Computing (Wiley). He is also the general workshops
chair for the IEEE International Conference on Local Computer
Networks (LCN ’17). Recently, he has published his book entitled
Cognitive Sensors & IoT: Architecture, Deployment, and Data Delivery

X III
XIV AU T H O R

with Taylor & Francis Group, CRC Press, New York. Since 2007, he
has been working on international wireless sensor networks projects
related to remote monitoring, as well as Smart Cities-related deploy-
ments and data delivery protocols using integrated RFID–sensor
networks.
 1
I NTRODUCTi ON

Wireless networking and advanced sensing technology have enabled

the development of low-cost and power-efficient Wireless Sensor
Networks (WSNs) which can be used in various domains such as
health care, home intelligence, and Outdoor Environment Monitoring
(OEM). Devices in WSNs, called Sensor Nodes (SNs), are used to
sense certain properties of the surrounding environment, including
physical/chemical properties, and transmit the sensed data to a central
unit, called a Base Station (BS), either periodically or on demand.
According to different application requirements, a WSN may consist
of just a few or as many as thousands of wireless nodes, operating in
a collaborative and coherent manner for a few days or several years to
fulfill a specific task [1].
Currently, WSNs are proving to be a promising wireless monitor-
ing technology in various OEM applications. In OEM applications,
sensor nodes are assumed to monitor and report the status of the sur-
rounding outdoor environments, as in air pollution, forest fire, and
flood detection applications [2,3]. The wide use of WSNs in OEM is
due to their enormous potential benefits and advantages. The direct
and continuous existence of WSNs surrounding the monitored phe-
nomena allows them to provide localized measurements and detailed
information that is hard to obtain through traditional instrumenta-
tion (i.e., traditional sensors). For example, the Moderate Resolution
Imaging Spectroradiometer (MODIS) [4,5], a traditional instrument,
was launched to provide forest fire surveillance by capturing satel-
lite images and processing them every 1–2 days, which is too late for
the forest to survive. In contrast, in Reference [2], a WSN is used to
provide real time interaction with forest fires due to the cooperative
efforts of sensor nodes. Another example is the redwood tree appli-
cation in Reference [6]. The entire surrounding of the redwood tree
is known to have substantial variations in temperature and moisture

1
2 WIREL E S S SENS O R NE T W O RKS

along its length (~70 m). In order to study these variations, a winch
near the top is needed to carry a set of weather monitoring instru-
ments with a very long serial cable connected to a battery powered
data logger at the tree base. However, a few small nodes constructing
the WSN were able to collect the targeted data wirelessly at different
scales and resolutions.
Measuring natural species with a WSN can also enable long term
data collection at scales and resolutions that are difficult, if not impos-
sible, to obtain otherwise. Moreover, due to the local computing and
networking capabilities of WSNs, sensor nodes can be reprogrammed
to do different tasks after deployment on site, based on changes in
the conditions of the WSN itself and the targeted phenomena [1].
Therefore, laboratories can be extended to the monitored site where
no human intervention is required, and real time interaction can be
provided remotely.
Nonetheless, deploying* WSNs in OEM is challenging and needs
further investigation due to harsh operational conditions, vast mon-
itored areas, limited energy budgets, and required 3-D setups. For
instance, most of the proposed OEM applications [2,3,6–9] are applied
in harsh and large-scale areas, such as forests and river surfaces, and
experience bad channel conditions as well as great chances of node
failures and damages. Furthermore, some applications required more
complicated deployment planning when 3-D setups [3,6,8,10], and
long lifetime periods [2,6,9] are required. Therefore, we investigate
efficient deployment plans to address these challenges and fill the lit-
erature gap in this critical research direction. In the following sub-
sections, OEM deployment challenges and motivations behind this
work are detailed. And major research contributions are reviewed in
addition to outlining the remaining contents of this book.

1.1 Contributions

This work aims primarily at optimizing the WSN deployment in

OEM applications by exploiting static/mobile RNs in grid-based
deployments. We study the optimization problem when SNs’ posi-
tions are known in advance and static RNs are utilized. Afterward, we
* Deployment is the process of identifying numbers and positions of the utilized
nodes (devices) in the WSN.
 IN T R O D U c Ti O N 3

examine a more complicated scenario where SNs’ positions are anony-

mous and a hybrid of static and mobile RNs are in use. Accordingly,
we present novel deployment schemes (strategies) to place the set of
RNs/SNs on any grid model in these scenarios. The main contribu-
tions of this book are summarized as follows:

1. Generally, SNs/RNs can be placed on any vertex of the 3-D

grid which results in a huge search space in large-scale appli-
cations, such as an OEM. Thus, this leads to extreme com-
plexity in repositioning extra static or mobile RNs during the
run time. We overcome this challenge by finding a subset of a
relatively small number of grid vertices for these RNs without
affecting the solution optimality.
2. Assuming prepositioned SNs and static RNs only, we con-
sider them along with their communication links as a graph
to mathematically represent their connectivity property. And
thus, we formulate a solid mathematical optimization problem
that has the aforementioned limited search space and aims at
maximizing the network connectivity while maintaining cost
and lifetime constraints.
3. Key to our contribution is also the use of proper cost and
communication models in addition to a revised definition
for the network lifetime, which is more appropriate in OEM
applications.
4. Assuming the mobility feature in a subset of the available
RNs, we also prolong the network lifetime. We mathemati-
cally formulate this problem as an Integer Linear Program
(ILP) and provide a two-phase solution for it. We consider
an effective power level metric at the objective function of
the ILP, which is the minimum node residual energy along
with the total consumed energy. This consideration results
in more influential movement for the mobile RNs, and thus
more equilibrium in the traffic load is achieved.
5. We derive an upper bound for the maximum network life-
time under idealistic operational conditions in OEM. The
proposed two-phase solution shows significant improvements
in terms of the network lifetime in comparison to this upper
bound.
4 WIREL E S S SENS O R NE T W O RKS

6. We investigate mobile vs. static grid-based deployments in

order to overcome harsh operational conditions in OEM
applications while considering hexagon and square grid
shapes for these deployment solutions.
7. Genetic-based solutions, as well as solid mathematical opti-
mization, have been exploited in significant OEM applica-
tions such as the emerging Smart City project.
In summary, this book provides guidelines for device deployment of
typical heterogeneous WSNs in OEM applications.

1.2 Book Outline

The rest of this book is organized as follows. Chapter 2 highlights

relevant background and related work in the literature. In Chapter 3,
we introduce a novel 3-D deployment strategy, called Optimized 3-D
Grid deployment with Lifetime Constraint (O3DwLC), for relay
nodes in environmental applications. The strategy optimizes network
connectivity while guaranteeing specific network lifetime and limited
cost. Key to our contribution is a very limited search space for the
optimization problem in addition to a revised definition for network
lifetime, which is more appropriate in environment monitoring. The
effectiveness of our strategy is validated through extensive simulations
and comparisons, assuming practical considerations of signal propa-
gation and connectivity. In Chapter 4, we propose two Optimized
Relay Placement (ORP) strategies with the objective of federating
disjoint WSN sectors with the maximum connectivity under a cost
constraint on the total number of RNs to be deployed. The perfor-
mance of the proposed approach is validated and assessed through
extensive simulations and comparisons assuming practical consid-
erations in outdoor environments. In Chapter 5, we propose a 3-D
grid-based deployment for relay nodes in which the relays are effi-
ciently placed on grid vertices. We present a novel approach, named
Fixing Augmented network Damage Intelligently (FADI), based on
a minimum spanning tree construction to reconnect the disjointed
WSN sectors. The performance of the proposed approach is validated
and assessed through extensive simulations; comparisons with two
mainstream approaches are presented. Our protocol outperforms the
 IN T R O D U c Ti O N 5

related work in terms of the average relay node count and distribution,
the scalability of the federated WSNs in large scale applications, and
the robustness of the topologies formed. We elaborate further on this
strategy in Chapter 6 where the hexagonal virtual grid is assumed
instead of the square one. We propose the use of cognitive nodes
(CNs) in the underlying sensor network to provide intelligent infor-
mation processing and knowledge-based services to the end users. We
identify tools and techniques to implement the cognitive functionality
and formulate a strategy for the deployment of CNs in the underlying
sensor network to ensure a high probability of successful data recep-
tion among communicating nodes. From MATLAB® simulations, we
were able to verify that in a network with randomly deployed sensor
nodes, CNs can be strategically deployed at predetermined positions,
to deliver application-aware data that satisfies the end user’s qual-
ity of information requirements, even at high application payloads.
Chapter 7 proposes a 3-D grid-based deployment for heterogeneous
WSNs (consisting of SNs, RNs, and MRNs). The problem is cast
as a Mixed Integer Linear Program (MILP) optimization problem
with the objective of maximizing the network lifetime while main-
taining certain levels of fault-tolerance and cost efficiency. Moreover,
an Upper Bound (UB) on the deployed WSN lifetime, given that
there are no unexpected node/link failures, has been driven. Based on
practical/harsh experimental settings in OEM, intensive simulations
show that the proposed grid-based deployment scheme can achieve an
average of 97% of the expected UB.
Additionally, a typical scenario has been discussed and analyzed in
smart cities while considering genetic based approaches in Chapter 8.
In this chapter, we study the path planning problem for these DCs
while optimizing their counts and their total traveled distances. As
the total collected load on a given DC route cannot exceed its stor-
age capacity, it is important to decide on the size of the exchanged
data packets (images, videos, etc.) and the sequence of the targeted
data sources to be visited. We propose a hybrid heuristic approach for
public data delivery in smart city settings. In this approach, public
vehicles are utilized as DCs that read/collect data from numerously
distributed Access Points (APs) and relay it back to a central process-
ing base station in the city. We also introduce a cost-based fitness
function for DCs election in the smart city paradigm. Our cost-based
6 WIREL E S S SENS O R NE T W O RKS

function considers resource limitations in terms of DCs’ count, stor-

age capacity and energy consumption. Extensive simulations are per-
formed; results confirm the effectiveness of the proposed approach in
comparison to other heuristic approaches with respect to total trav-
eled distances and overall time complexity. Finally, Chapter 9 con-
cludes the book and provides directions for future work.

References
1. I. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci, A survey
on sensor networks, IEEE Communications Magazine, vol. 40, no. 8,
pp. 102–114, 2002.
2. M. Hefeeda, Forest fire modeling and early detection uses wireless
sensor networks, School of Computing Science, Simon Fraser University,
Vancouver, BC, Technical Report TR-2007-08, August 2007.
3. F. Al-Turjman, Information-centric sensor networks for cognitive IoT:
An overview, Annals of Telecommunications, vol. 72, no. 1, pp. 3–18,
2017.
4. G. Guo and M. Zhou, Using MODIS land surface temperature to eval-
uate forest fire risk of Northeast China, IEEE Geoscience and Remote
Sensing Letters, vol. 1, no. 2, pp. 98–100, 2004.
5. NASA, MODIS—Moderate Resolution Imaging Spectroradiometer.
http://modis.gsfc.nasa.gov/, 2009.
6. G. Tolle, J. Polastre, R. Szewczyk, and D. Culler, A macroscope in the
redwoods, In Proceedings of the ACM Conference on Embedded Networked
Sensor Systems (SenSys), San Diego, CA, 2005, pp. 51–63.
7. J. Cox, MIT’s tree-powered wireless network, Network World. www.
networkworld.com/community/node/33278, 2009.
8. A. Singh, M. Batalin, V. Chen, M. Stealey, B. Jordan, J. Fisher,
T. Harmon, M. Hansen, and W. Kaiser, Autonomous robotic sensing
experiments at San Joaquin river. In Proceedings of the IEEE International
Conference on Robotics and Automation (ICRA), Rome, Italy, 2007,
pp. 4987–4993.
9. B. Son, Y. Her, and J. Kim, A design and implementation of forest-
fires surveillance system based on wireless sensor networks for South
Korea mountains, International Journal of Computer Science and Network
Security (IJCSNS), vol. 6, no. 9, pp. 124–130, 2006.
10. J. Thelen, D. Goense, and K. Langendoen, Radio wave propagation
in potato fields, In Proceedings of the Workshop on Wireless Network
Measurements, Riva del Garda, Italy, April 2005.
 2
D EPLOYMENT O F W iRELES S
S ENSO R N E T wORKS iN
O UTD O O R E N V iRONMENT
M ONiTORiN G
An Overview

Deployment planning is of the utmost importance in the context of

Wireless Sensor Networks (WSNs) as it decides the available resources
and their configuration for system setup. This in turn plays a key role
in the network performance. A significant amount of research has
been made to enhance the network performance during its opera-
tional time by optimizing data routing and Medium Access Control
(MAC) protocols [1]. Nevertheless, even with the best suit of routing
and MAC protocols, a designed sensor network cannot achieve the
targeted performance level unless it has been appropriately configured
in advance. For example, if the installed devices are insufficient or if
there are architecture deficiencies due to an ineffective deployment
plan, the connectivity between the deployed nodes and the system
lifetime will be degraded, in addition to causing nonoperational net-
works in some critical situations like flood and fire detection [2,3].
Accordingly, recent efforts are moving toward enhancing the net-
work performance by optimizing the network deployment planning.
Several network properties, considered as deployment objectives and
constraints, have led to a rich research field. Using the categorization
criteria adopted in Reference [4], we classify the technical approaches
of WSN deployment into two groups, random deployment and deter-
ministic (grid-based) deployment. In random deployment, nodes are
arbitrarily scattered and are managed in an ad hoc manner. In con-
trast, in the grid-based deployment, nodes are placed according to
virtual grid vertices which leads to more efficiency in terms of the

7
8 WIREL E S S SENS O R NE T W O RKS

targeted network properties. The shortcomings of these approaches

will be pointed out where appropriate in the following sections.

2.1 Desired Network Properties in OEM

Due to aforementioned challenges and characteristics, we will dis-

cuss the most important properties in the Outdoor Environment
Monitoring (OEM) deployment planning, namely connectivity, fault-
tolerance, and network (system) lifetime.

2.1.1 Connectivity

Connectivity is defined as the ability of the deployed nodes to stay

connected with the Base Station (BS) to satisfy the targeted applica-
tion objectives [5]. As previously mentioned, nodes and communica-
tion links in OEM are subject to several risks. Therefore, the deployed
network connectivity is essential to ensure that the sensed data is deliv-
ered correctly to the BS for processing. In the context of WSNs, con-
nectivity could be presented as k-connectivity. k-connectivity has two
different meanings, namely, k-path connectivity and k-link connec-
tivity. k-path connectivity means that there are k independent paths
between the deployed nodes and the BS, while the k-link connectivity
means that each node is directly connected to k neighboring nodes.
In fact, a WSN could be disconnected even if k-link connectivity is
satisfied. However, when k > 1, the network can tolerate some node
and link failures. At the same time, the higher degree of connectivity
improves communication availability among the deployed nodes. In
the literature, some OEM deployments, such as the one in Reference
[6], support k-path connectivity while others support k-link connec-
tivity for more reliable and longer lasting networks, as in References
[7–9]. In fact, it may not be necessary to maintain k-connectivity
between all nodes, but only among the nodes which form the network
backbone, which we call irredundant (critical) nodes.
In general, connectivity problems can be repaired after their
occurrence either by using extra redundant nodes or by utilizing the
node mobility feature. Node redundancy [10] is used to overcome
disconnected networks. Redundant nodes, which are not being used
for communication or sensing, are turned off. When the network
 D E P L OY M EN T O F WIREL E S S SENS O R NE T W O RKS 9

becomes disconnected, one or more of the redundant nodes is turned

on to resume the network connectivity. In Reference [11], the low-
est number of redundant nodes are added to a disconnected static
network, so that the network remains connected. In addition, over-
lapping clusters of sensor nodes are used to enhance the network con-
nectivity in Reference [12]. In Reference [13], a distributed recovery
algorithm is developed to address k-link connectivity requirements,
where k is equal to 1 and 2. The idea is to identify the least set of nodes
that should be repositioned in order to reestablish a particular level of
connectivity. A shortcoming of such techniques is the requirement
for extra nodes regardless of their roles, types, and physical positions,
which may not be cost-effective in environment monitoring applica-
tions. Also, when some redundant nodes fail, it may no longer be
possible to repair the network connectivity.
Meanwhile, node mobility can also be used to maintain network
connectivity. Typically, mobile nodes are relocated after deploy-
ment to carry data between disconnected partitions of the network
[14]. Providing radio connectivity using mobile nodes while consid-
ering ongoing missions, traveling distance, and message exchange
complexity has also been considered recently in Reference [13].
However, relocating nodes without considering grid connectivity
properties and characteristics can have severe effects on the direc-
tion of movement and the choice of the most appropriate node to
be moved.
Although the above techniques can aid with repairing connectiv-
ity problems, they do not address the sources of these problems. In
this book, we present a radically different approach toward addressing
connectivity problems and complementing the work of the aforemen-
tioned techniques. We address the properties of grid connectivity in
practice and under realistic scenarios, such as inaccurate positioning
and communications irregularity. Thus, more efficient connectivity
planning and maintenance can be achieved.

2.1.2 Fault-Tolerance

Due to the harsh operational conditions where OEM sensor networks

are deployed, WSNs are susceptible to unpredictable events such as
hardware failures, power leakage, physical damage, environmental
10 WIREL E S S SENS O R NE T W O RKS

interference, etc. Thus, it is crucial to ensure that WSNs are also fault-
tolerant. As mentioned before, the sources of failure in OEM can
occur because of either node or link failures. These failures are char-
acterized by the Probability of Node Failure (PNF) and Probability
of Disconnected Nodes (PDN) to indicate the harshness level of the
monitored environment.
Three scenarios are considered in the deployed networks to be fault-
tolerant: fault detection, fault diagnosis, and fault recovery [15]. There
are many techniques for fault detection and diagnosis in the literature;
they are classified into self-diagnosis, group detection, and hierar-
chal detection [16]. On the other hand, few fault recovery techniques
are addressed in the literature. They can be classified into node and
link recovery techniques. These recovery techniques assure the ability
of the deployed network to stay operational and fulfill the assigned
task in the presence of a set of nonfunctional nodes/links, so that
sensed and carried data can be recovered. Although both node and
link failures are important to be tolerated together, some deployments
have considered only node failure recovery such as [9], while others
have considered only the link failure recovery as in Reference [17],
which weakens the deployed network performance. Compared to the
complexity of considering a single type of these recovery techniques,
it is trivial to consider both of them in one deployment technique
by employing redundant nodes and maximizing the communication
connectivity.

2.1.3 Lifetime

As for the system lifetime, it is the most challenging problem in any

OEM deployment as the utilized nodes are energy constrained, and
deployed networks are sometimes required to work for longer inter-
vals measured in years [18,19]. In addition, it is undesirable to revisit
harsh environments, such as those targeted by the OEM applications,
simply for node replacement or recharging. Accordingly, accurate life-
time prediction in the early deployment planning stages is required.
There are two types of lifetime predictions in the literature: node and
network lifetime predictions. Node lifetime can be measured in sev-
eral ways. For example, it can be measured based on the number of
cycles over which the data is collected, on the cumulative active time
 D E P L OY M EN T O F WIREL E S S SENS O R NE T W O RKS 11

of the node before it is depleted, or on the cumulative traffic volume

of the node before its energy is depleted. However, all of the afore-
mentioned measuring methods require accurate energy consumption
models which consider the effects of surrounding environments for
more realistic and practical predictions. This issue is addressed in
our research. In fact, nodes in WSNs consume power in three main
domains: sensing, communication, and data processing. Because
the majority of power consumed per node is used in communica-
tion, related work in literature has concentrated on proposing energy
consumption models for receiving and transmitting wireless signals.
Network lifetime has several definitions in the literature. One of the
most common states that, “it is the time until the first node death
occurs” [20]. Such a definition may not be suitable if we are monitor-
ing, for example, the forest temperature or humidity as long as we
are still receiving the same information from other nodes in the same
area. Similarly, the definition, which says that “it is the time until
the last node death occurs”, can serve as an upper bound of other
definitions. Thus, both of these definitions are unrealistic for OEM
applications. Another way to define lifetime relies on the percent-
age of nodes alive [20]. But choosing such a percentage threshold is
usually arbitrary in the literature and does not reflect the application
requirements. In addition, being alive does not mean that the node
is still connected to the sink node (or BS). Yet another approach to
evaluate network lifetime is in terms of connectivity and network par-
titions [21]. These definitions still do not satisfy OEM applications.
For instance, if a set of nodes is destroyed because of the movement of
an animal or a falling tree in the monitored site, the network can still
be considered functional as long as some other nodes are still alive and
connected to the BS. Accordingly, redundant nodes and links must be
used in such applications. In order to take into consideration node/
link redundancy, an OEM specific lifetime definition is proposed in
this book.

2.2 Random vs. Deterministic WSNs Deployment

In the literature, deployments targeting network properties men-

tioned above are mainly classified into two categories: random vs.
deterministic (grid-based) deployments.
12 WIREL E S S SENS O R NE T W O RKS

2.2.1 Random Deployment

In random deployments, nodes can be placed randomly to reduce the

deployment cost, or based on a weighted random deployment plan-
ning, they can be placed where the distributed nodes’ density is not
uniform in the monitored areas. For instance, K. Xu et al. [22] studied
the random RN deployment in a 2-D plane. The authors proposed an
efficient network lifetime maximization deployment when the RNs
are directly communicating with the BS. In this study, it was estab-
lished that different energy consumption rates at different distances
from the BS render uniform RN deployment, thus being a poor can-
didate for network lifetime extension. Alternatively, a weighted ran-
dom deployment is proposed. In this random deployment, the density
of RN deployment increases as the distance to the BS increases, and
thus distant RNs can split the traffic among themselves. This in turn
extends the average RN lifetime. We note that some attempts have
been made toward the 3-D deployment as well. For example, authors
in Reference [23] consider the implications of sensing and communica-
tion ranges on the network connectivity in random 3-D deployments.
In Reference [24], the authors proposed a distributed algorithm that
achieves k-connectivity in homogenous WSNs randomly deployed in
the 3-D space. Simply, the idea is to adapt the nodes’ transmission
power until either the distance separating two consecutive neighbors
is greater than a specific threshold or the maximal power is reached.
However, this method is not cost-effective due to hardware complex-
ity required to adapt the transmission ranges, which is not worthy in
OEM applications. In addition, adapting the transmission range to
reach farther distances would increase the energy consumption, and
hence degrade the network lifetime.

2.2.2 Deterministic (Grid-Based) Deployment

Proposals described above are suitable for applications which are not
interested in the exact node positioning. In contrast, some propos-
als have advocated deploying nodes exactly on specific predefined
locations, called grid vertices. These locations are optimized in
terms of the aforementioned network properties and the feasibil-
ity of the location itself in reality (e.g., non-reachable locations are
 D E P L OY M EN T O F WIREL E S S SENS O R NE T W O RKS 13

not feasible for node deployment). Due to the interest posed by

OEM applications in the exact physical positioning of sensor and
relay nodes, this type of deployment (i.e., grid-based deployment)
serves this purpose more appropriately, and hence is adopted in our
work. Moreover, coupling this type of deployment with guaranteed
multipath routing can significantly enhance the network lifetime
and fault-tolerance chances, in addition to repairing network con-
nectivity problems.
In the literature, two categories of approaches have been pursued
to enhance these three main properties (i.e., network lifetime, con-
nectivity, and fault-tolerance): (1) populating additional (redundant)
nodes, and (2) employing mobile nodes. The approach presented in
Reference [13] counters faulty sensor nodes by repositioning mobile
pre-identified spare sensors from different parts of the network. The
most appropriate candidate spare node is the closest one. In order to
detect the closest spare to the faulty sensor, a grid-based approach
is proposed. The targeted region is divided into cells. Each cell has
a head advertising available spare nodes in its cell or requesting the
spare ones for it. Once the spares are located, they are moved to the
cell of failure without affecting the data traffic and the network topol-
ogy. However, assuming that the closest node will always be moved
to recover the faulty ones will rapidly deplete their batteries unless a
load balance constraint is added.
W. Alsalih et al. [33] proposed a deterministic deployment for
data gathering using a mobile node, called a data collector (DC). In
this deployment, a mobile data collector moves along a predefined
track through the sensing field. SNs whose transmission ranges
overlap with this track are called relay nodes (RNs). In addition to
forwarding their own measurements to the data collector, the RNs
collect the measurements of neighboring SNs whose transmission
ranges do not overlap with the track. It was shown that employ-
ing mobile data collectors can extend the network lifetime in com-
parison to conventional WSNs employing static nodes only. In fact,
the concept of mobile nodes, or data collectors, was used earlier
in References [25,26]. These references assumed the existence of a
number of predefined locations where data collectors can be placed.
The network lifetime was divided into equal length time periods,
called rounds (a more rigorous definition of a round will be given
14 WIREL E S S SENS O R NE T W O RKS

subsequently in this book), and data collectors are moved to new

locations at the beginning of each round. In Reference [25], the
problem of finding the optimal locations for the mobile nodes was
formulated as a Mixed Integer Linear Program (MILP) problem
whose objective is minimizing the total consumed energy during a
round. It was also shown that the identified optimal locations were
optimal when the objective was to minimize the maximum energy
spent by a single SN. However, these two objective functions are
not suitable for the placement of mobile nodes since the optimal
solutions will not change over the time, i.e., the maximum or total
energy spent per round might not change; hence, locations of data
collectors (mobile nodes) will not be changed. In contrast, a heu-
ristic mobile node placement scheme that considers nodes’ residual
energy was proposed in Reference [26]. The locations of data col-
lectors were chosen according to local information only. In other
words, the decision of whether or not a mobile node is placed at a
given location is made based on the residual energy of SNs that are
one hop away from that location. Consequently, the locations calcu-
lated may be far from optimal.
Even so, assuming all nodes are able to move may not be valid
in a number of OEM applications. It is more appropriate if a sub-
set of the deployed nodes is assumed to have the mobility feature, or
more static nodes are populated to repair connectivity (by recovering
faulty nodes) and prolong the network lifetime. In an earlier work
[27], we proposed an Integer Linear Program (ILP) that assumes a
number of the deployed nodes have the ability to change their loca-
tions during the network operational time. This deployment strategy
enforces the minimal energy consumption while maintaining connec-
tivity requirements, fault-tolerant constraints, and cost-effectiveness.
A key parameter in formulating the deployment problem is limiting
the 3-D search space to a finite set of points by using grid models. The
performance of the proposed strategy is validated through extensive
simulations. Network lifetime enhancements of up to 50% have been
achieved as compared to other deployment schemes under practical
settings in OEM applications.
Unlike the previous references, the authors in Reference [28]
proposed an algorithm to achieve fault-tolerant and long lasting
WSNs by populating static nodes which have at least two disjointed
 D E P L OY M EN T O F WIREL E S S SENS O R NE T W O RKS 15

paths between every pair of sensor nodes. This algorithm simply

identifies candidate positions for relay nodes that cover the maxi-
mum number of sensors. Such candidate positions are found at the
intersections of the communication ranges of neighboring sensor
nodes. Relay nodes with the widest coverage span are then placed at
these candidate positions. The algorithm checks whether the relays
form a 2-connected graph and every sensor can reach at least two
relays. If not, more relays are added and the connectivity and cov-
erage are rechecked. This algorithm is repeated until the objective
is achieved. However, relay positions may not be accurate in real-
ity due to communication range irregularity. This work becomes
difficult when 3-D space is considered. Furthermore, lifetime and
energy constrains are not addressed in this algorithm, in addition
to ignoring node recovery (i.e., ignoring the lost data by the faulty
node) leading to link fault-tolerance deployment rather than link
and node fault-tolerance. This makes the proposed deployment
unsuitable for OEM applications. Similarly, static nodes are used
in Reference [29] to repair connectivity in grid-based deployments.
In this work, the authors aim at populating additional minimum
numbers of relay nodes to establish connections with partitioned
segments in the network using a heuristic approach. It shows ben-
eficial aspects of the resulting topology with respect to connectivity
and traffic balance. Nevertheless, node mobility, which may have a
dramatic impact on both connectivity and lifetime of the network,
hasn’t been addressed in this work. Moreover, adding additional
relay nodes during the monitoring process is not always applicable
under the harsh operational conditions experienced by the OEM
applications. Finally, this work is adopted in a 2-D plane only.
Contrarily, in Reference [30], maximal 3-D connectivity using the
least number of nodes has been discussed, and thus it provides a
cost-effective deployment plan. However, this work assumes regu-
lar communication range represented by a binary sphere around
each grid vertex, which is not the case in practice. In addition, there
is no fault-tolerance and lifetime constraints in this work, mak-
ing it unsuitable for OEM applications. We note that this work is
restricted to a specific type of grid model, which is the octahedron
3-D grid, and thus it is not applicable on other 3-D grid shapes
(e.g., cubes, pyramids, etc.).
16 WIREL E S S SENS O R NE T W O RKS

2.3 Summary

In this section, we conclude our literature overview. Reviewed deploy-

ment proposals are classified and summarized in Table 2.1. In addi-
tion, corresponding proposals to reviewed deployments are compared
based on the desired network properties in OEM applications. The
notation (1) in Table 2.1 indicates that the technique of additional
(redundant) nodes is used to achieve the desired network properties,
while the notation (2) indicates that node mobility is used for that pur-
pose. As for 2-D and/or 3-D, it means that the deployment is designed
for two-dimensional and/or three-dimensional space, respectively. The
Network-based and/or Node-based under the lifetime property indicate
that the system lifetime is measured based on the overall network sta-
tus (e.g., network partition) or on a single node status (e.g., first node
death). Under fault-tolerance, link fault-tolerance means that the toler-
ated source of failure is the link, while node fault-tolerance means that
the tolerated source of failure is the node. The 1-path under connectiv-
ity refers to providing at least one path from each node to the BS dur-
ing the monitoring process. Finally, the “—” means that the network
property is not addressed in the corresponding proposal.
Unsurprisingly, most of the proposed deployments, which focus
on the network connectivity and lifetime and/or fault-tolerance, have
addressed the cost factor of the deployed network in terms of total
number of nodes used. This is due to high probabilities of node dam-
age and loss in applications requiring these network properties. In
contrast, none of the proposed deployments have addressed all of the
required properties, except Reference [33]. Even in Reference [33], a
1-path connectivity property is addressed rather than k-path ­and/­or
k-link connectivity, where k ≥ 1. Also, it is a random deployment
technique. Moreover, lifetime and fault-tolerance properties in this
reference are restricted to node-based and node fault-tolerance only,
while it is necessary in OEM to consider the network-based and
link fault-tolerant properties as well. Hence, there is no deployment
proposal which considers all of the required network properties in
OEM applications. Also, there is no proposal considering both link
and node fault-tolerance simultaneously. Similarly, there is no pro-
posal considering both node-based and network-based system life-
time simultaneously. Furthermore, it is obvious from Table 2.1 that
 D E P L OY M EN T O F WIREL E S S SENS O R NE T W O RKS
Table 2.1 A Comparison between Various Deployment Proposals in the Literature Considering the Desired Network Properties in OEM

NETWORK PROPERTIES DEPLOYMENT

TARGETED
REFERENCE COSTa CONNECTIVITY FAULT-TOLERANCE LIFETIME TYPE TECHNIQUE SPACE
[13] √ 1-path — Network-based Grid-based (2) 2-D
[31,30] √ k-path Link fault-tolerant — Grid-based (1) 3-D
[32] √ k-path Link fault-tolerant — Grid-based (1) 2-D
[24] — k-link Link fault-tolerant — Random (1) 3-D
[29] √ 1-path — Network-based Grid-based (1) 2-D
[23] — k-link Link fault-tolerant — Random (1) 3-D
[22] √ 1-path — — Random (1) 2-D
[33] √ 1-path Node fault-tolerant Node-based Random (1) 2-D
[27] √ 1-path — Node-based Grid-based (2) 2-D
[27] √ k-path Node-link fault-tolerant Network-based Grid-based (2) 3-D
a The √ in this column indicates that the cost factor is considered in the corresponding reference.

17
18 WIREL E S S SENS O R NE T W O RKS

there is no proposal considering the deployment technique of type (2)

in the 3-D space. Although there are some ­proposals considering the
deployment technique (2) in 2-D plane, they are not addressing the
case where heterogeneous hybrid nodes are used (i.e., when a subset
of the RNs is mobile). Consequently, we are in a dire need for generic
3-D grid-based deployments considering both techniques, (1) and (2),
to achieve the desired OEM network properties. We need a deploy-
ment that addresses technique (2) where hybrid nodes are utilized. In
addition, we need a deployment that addresses technique (1) such that
the additional nodes are considered and deployed at the beginning
of the monitoring process. Moreover, a deployment plan consider-
ing both node- and network-based lifetime definitions, in addition to
communication links irregularity in 3-D space, is required.

References
1. I. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci, A survey
on sensor networks, IEEE Communications Magazine, vol. 40, no. 8,
pp. 102–114, 2002.
2. M. Hefeeda, Forest fire modeling and early detection uses wireless
sensor networks, School of Computing Science, Simon Fraser University,
Vancouver, BC, Technical Report TR-2007-08, August 2007.
3. F. Al-Turjman, Information-centric sensor networks for cognitive IoT:
An overview, Annals of Telecommunications, vol. 72, no. 1, pp. 3–18,
2017.
4. M. Younis and K. Akkaya, Strategies and techniques for node place-
ment in wireless sensor networks: A survey, Elsevier Ad Hoc Network
Journal, vol. 6, no. 4, pp. 621–655, 2008.
5. M. Ibnkahla, Ed., Adaptation and Cross Layer Design in Wireless Networks,
CRC Press, Boca Raton, 2008.
6. G. Tolle, J. Polastre, R. Szewczyk, and D. Culler, A macroscope in the
redwoods, In Proceedings of the ACM Conference on Embedded Networked
Sensor Systems (SenSys), San Diego, CA, 2005, pp. 51–63.
7. M. Hamilton, E. Graham, P. Rundel, M. Allen, W. Kaiser, M. Hansen,
and D. Estrin, New approaches in embedded networked sensing for
terrestrial ecological observatories, Environmental Engineering Science,
vol. 24, no. 2, pp. 192–204, 2007.
8. R. Pon, A. Kansal, D. Liu, M. Rahimi, L. Shirachi, W. Kaiser, G. Pottie,
M. Srivastava, G. Sukhatme, and D. Estrin, Networked infomechanical
systems (NIMS): Next generation sensor networks for environmental
monitoring, IEEE MTT-S International Microwave Symposium Digest,
vol. 1, pp. 373–376, 2005.
 D E P L OY M EN T O F WIREL E S S SENS O R NE T W O RKS 19

9. N. Ramanathan, L. Balzano, M. Burt, D. Estrin, E. Kohler, T. Harmon,

C. Harvey, J. Jay, S. Rothenberg, and M. Srivastava, Rapid deployment
with confidence: Calibration and fault detection in environmental sen-
sor networks, Center of Embedded Network Systems, Los Angeles, CA.
www. scholarship.org/uc/item/8v26b5qh, 2006.
10. A. Cerpa and D. Estrin, Ascent: Adaptive self-configuring sensor net-
works topologies, IEEE Transactions on Mobile Computing, vol. 3, no. 3,
pp. 272–285, 2004.
11. N. Li and J. C. Hou, Improving connectivity of wireless ad hoc net-
works. In Proceedings of the IEEE International Conference on Mobile and
Ubiquitous Systems: Networking and Services (MobiQuitous), San Diego,
CA, 2005, pp. 314–324.
12. M. Youssef, A. Youssef, and M. Younis, Overlapping multihop clus-
tering for wireless sensor networks, IEEE Transactions on Parallel
Distributed Systems, vol. 20, no. 12, pp. 1844–1856, 2009.
13. A. Abbasi, U. Baroudi, M. Younis, and K. Akkaya, C2AM: An algo-
rithm for application-aware movement-assisted recovery in wireless
sensor and actor networks, In Proceedings of the ACM International
Wireless Communications and Mobile Computing Conference (IWCMC),
Leipzig, Germany, 2009, pp. 655–659.
14. M. Dunbabin, P. Corke, I. Vasilescu, and D. Rus, Data muling over
underwater wireless sensor networks using an autonomous underwater
vehicle, In Proceedings of the IEEE International Conference on Robotics
and Automation (ICRA), Orlando, FL, 2006, pp. 2091–2098.
15. M. Yu, H. Mokhtar, and M. Merabti, A survey on fault manage-
ment in wireless sensor networks, ACM Journal of Network and Systems
Management, vol. 15, no. 2, pp. 171–190, 2007.
16. L. Moreira, H. Vogt, and M. Beigl, A survey on fault tolerance in wire-
less sensor networks, Scientific Commons. http://en.scientificcommons.
org/23851530, 2009.
17. J. Thelen, D. Goense, and K. Langendoen, Radio wave propaga-
tion in potato fields, In Proceedings of the Workshop on Wireless Network
Measurements, Riva del Garda, Italy, April 2005.
18. A. Mainwaring, J. Polastre, R. Szewczyk, D. Culler, and J. Anderson,
Wireless sensor networks for habitat monitoring, In Proceedings of the
ACM International Workshop on Wireless Sensor Networks and Applications
(WSNA), Atlanta, GA, 2002, pp. 399–423.
19. P. Juang, H. Oki, Y. Wang, M. Martonosi, L. Peh, and D. Rubenstein,
Energy-efficient computing for wildlife tracking: Design trade-offs
and early experiences with ZebraNet, ACM SIGPLAN Notices, vol. 37,
no. 10, pp. 96–107, 2002.
20. L. van Hoesel, T. Nieberg, J. Wu, and P. Havinga, Prolonging the life-
time of wireless sensor networks by cross layer interaction, IEEE Wireless
Communications, vol. 11, no. 6, pp. 78–86, 2004.
21. Y. Chen and Q. Zhao, On the lifetime of wireless sensor networks,
IEEE Communications Letters, vol. 9, no. 11, pp. 976–978, 2005.
20 WIREL E S S SENS O R NE T W O RKS

22. K. Xu, H. Hassanein, G. Takahara, and Q. Wang, Relay node deploy-

ment strategies in heterogeneous wireless sensor networks, vol. 9, no. 2,
pp. 145–159, 2010.
23. V. Ravelomanana, Extremal properties of three-dimensional sensor net-
works with applications, IEEE Transactions on Mobile Computing, vol. 3,
no. 3, pp. 246–257, 2004.
24. M. Ishizuka and M. Aida, Performance study of node placement
in sensor networks, In Proceedings of the International Conference on
Distributed Computing Systems Workshops (ICDCSW), Tokyo, Japan,
2004, pp. 598–603.
25. S. Gandham, M. Dawande, R. Prakash, and S. Venkatesan, Energy
efficient schemes for wireless sensor networks with multiple mobile base
stations, In Proceedings of the IEEE Global Telecommunications Conference
(GLOBECOM), San Francisco, CA, 2003, pp. 377–381.
26. A. Azad and A. Chockalingam, Mobile base stations placement and
energy aware routing in wireless sensor networks, In IEEE Wireless
Communications and Networking Conference (WCNC), Las Vegas, NV,
2006, pp. 264–269.
27. F. Al-Turjman, H. Hassanein, and M. Ibnkahla, Towards prolonged
lifetime for deployed WSNs in outdoor environment monitoring,
Elsevier Ad Hoc Networks Journal, vol. 24, no. A, pp. 172–185, Jan 2015.
28. B. Hao, H. Tang, and G. Xue, Fault-tolerant relay node placement in
wireless sensor networks: Formulation and approximation, In Proceedings
of the Workshop on High Performance Switching and Routing (HPSR),
Phoenix, AZ, April 2004.
29. S. Lee and M. Younis, Optimized relay placement to federate seg-
ments in wireless sensor networks, IEEE Journal on Selected Areas in
Communications, vol. 28, no. 5, pp. 742–752, 2010.
30. S. N. Alam and Z. Haas, Coverage and connectivity in three-dimensional
networks, In Proceedings of the ACM International Conference on Mobile
Computing and Networking (MobiCom), Los Angeles, CA, 2006,
pp. 346–357.
31. K. Xu, H. Hassanein, G. Takahara, and Q. Wang, Relay node deploy-
ment strategies in heterogeneous wireless sensor networks: Multiple-
hop communication case, In Proceedings of the IEEE Conference on Sensor
and Ad Hoc Communications and Networks (SECON), Santa Clara, CA,
2005, pp. 575–585.
32. F. Al-Turjman, M. Ibnkahla, and H. Hassanein, An overview of wire-
less sensor networks for ecology and forest monitoring, In Proceedings of
the IEEE International Workshop on Signal Processing and its Applications
(WoSPA), Sharjah, UAE, 2008.
33. W. Alsalih, H. Hassanien, and S. Akl, Placement of multiple mobile
data collectors in underwater acoustic sensor networks, ACM Journal
of Wireless Communications and Mobile Computing, vol. 8, no. 8,
pp. 1011–1022, 2008.
 3
E FFi CiENT D EPLOYMENT
O F W iRELES S S ENSOR
N E T wO RKS TARG E TiN G
E N V iRO NMENT M ONiTORiN G
A PPLi CATiONS*

Wireless Sensor Networks (WSNs) enable long term environment

monitoring at scales and resolutions that are difficult, if not impossi-
ble, to obtain using conventional techniques. WSNs can be re-tasked
after deployment in the field based on changes in the environment,
conditions of the sensor network themselves, or scientific endeavor
requirements [3]. However, most environmental applications require
all seasons data-gathering which can include extreme conditions.
Thus, long-lasting, tightly connected networks in the monitored field
are needed in order to have better understanding of the monitored
phenomena such as the life cycles of huge redwood trees [34] or to
satisfy specific application objectives for several years as in forestry
fire detection [31].
One of the most efficient techniques used for maintaining long-
lasting connected networks is achieved by deploying Relay Nodes
(RNs) [19]. Relay nodes can have extra storage space and much more
powerful transceivers to forward sensed data for long distances in huge
monitored sites; thus, the energy of Sensor Nodes (SNs) is saved for
further data sensing and gathering. Nevertheless, deployment of relay
nodes [11,24] in environmental applications is a challenging problem
due to harsh environments and required 3-D setups.
Harshness of the environment arises because of the nature of
outdoor monitoring applications where sensor networks may work

* This chapter has been coauthored with Hossam S. Hassanein and Mohamed A.
Ibnkahla.

21
22 WIREL E S S SENS O R NE T W O RKS

under heavy rain, extreme temperature variations, and sometimes

stormy days destroying the deployed nodes and/or their communica-
tion links (edges). Many nodes and communication links may also
be destroyed by unexpected visitors such as birds and wild fauna. In
addition, dense trees and growing foliage can weaken communication
links and affect connectivity. Nodes and links are prone to risks which
lead to high probabilities of failures. Many nodes may become discon-
nected, which can degrade the overall network lifetime. As a result,
we are characterizing harshness of the monitored environment by the
Probability of Node Failure (PNF) and Probability of Disconnected
Nodes (PDN). Due to high PNF and PDN in environmental applica-
tions, it is reasonable to have redundant nodes in the deployed sensor
networks [23]. In the scope of this chapter, a redundant node is the
node which can be removed from the network without affecting the
targeted data. Contrarily, an irredundant node is defined as a unique
source of information in the monitored site that cannot be recovered
by other nodes in the network.
Meanwhile, deployment in environmental applications becomes
more challenging when 3-D setups are required. In environmental
applications, relay nodes are not only forwarding data from different
variations in the horizontal plane but also from different vertical levels
(e.g., on trees, at soil surface, and even underground). For instance,
in monitoring the massive redwood trees in California, some experi-
ments required sensor placements at different heights on the trees
spanning a range of tens of meters [34]. There is also increased inter-
est in 3-D environmental applications such as CO2 flux monitoring
and imagery [31], where sensors are placed at different vertical lev-
els to fulfill coverage and data accuracy requirements. Accordingly,
communication links between deployed sensor nodes have to be con-
sidered in a 3-D space rather than a 2-D plane only, raising the com-
plexity of the deployed network connectivity [12,26].
Nonetheless, existing deployment schemes for such applica-
tions [28,29,31] are not always based on sound connectivity models
but rather on simplified lifetime models. They have not efficiently
addressed the problem of connectivity in a 3-D space, which is a nat-
ural model in environmental applications. For example, deployment
in Reference [18] focused on connectivity in 2-D outdoor applica-
tions. Deployment algorithms are proposed in References [25,33] to
 EF FI CIEN T D E P L OY M EN T O F WIREL E S S 23

guarantee the 2-D connectivity and/or ensure survivability in case of

node failure without lifetime considerations. In contrast, the deploy-
ment in Reference [15] is aimed at maximizing the network lifetime
under specific energy budget. The energy provisioning and relay node
placement are formulated as a mixed-integer linear programming
problem in a 2-D plane. Heuristic algorithms are then introduced to
overcome the computational complexity. Nevertheless, relay nodes in
Reference [15] are assumed to adapt their transmission range to reach
any other node in the network, and connectivity is not considered
as an issue (which is not practical in environmental applications). A
hybrid approach has been proposed in Reference [35] to balance con-
nectivity and lifetime in 2-D outdoor deployments. Even in the most
recent papers in environmental WSN applications, including volcano
[30] and harsh industrial [21] environment monitoring, network con-
nectivity is considered in a 2-D plane with the assumption of a very
basic binary communication disc model which is not realistic. Thus,
these approaches are prone to failure in practical large-scale environ-
mental applications.
It is worth mentioning that there are some attempts toward the
3-D deployment. As an example, authors in Reference [18] studied
the effects of sensing and communication ranges on connectivity in
a 3-D space. However, connectivity optimization with lifetime con-
straints has not been investigated so far in 3-D environmental appli-
cations. As noted in Reference [37], many of the popular deployment
strategies, which are optimally solved in polynomial time in a 2-D
plane, become NP-Hard in 3-D settings.
We also note that relay nodes in environment monitoring are gen-
erally more expensive than sensing nodes due to the cost of the wide
range transceivers used to cover large-scale areas, such as forests and
cities. Therefore, efficient 3-D deployment must maintain connectiv-
ity and lifetime that limits the number of these expensive nodes. In
efficient environmental deployments, it is also undesirable to revisit
the monitored sites (e.g., for node replacement or battery recharging).
Therefore, the deployed wireless sensor network must be guaranteed
to function for a pre-specified lifetime period. For more accurate and
practical lifetime guarantees, an environment-specific lifetime defi-
nition should be considered. Even though there are several defini-
tions for WSN lifetime in the literature, there is no agreement on a
24 WIREL E S S SENS O R NE T W O RKS

definition for lifetime in environmental applications. An inappropri-

ate definition might lead to incorrect lifetime estimation, and hence
cause a waste of resources.
In this chapter, we investigate an efficient way for the relays’ place-
ment that address aforementioned challenges and desired WSNs’
features in environmental applications. Such node placement prob-
lem has been shown in Reference [5] to be NP-hard. Finding non-
optimal approximate solutions is also NP-hard in some cases [13]. To
address this complexity, we propose an efficient two-phase relay node
deployment in a 3-D space, called O3DwLC strategy. The first phase
of O3DwLC is used to set up a connected network backbone using
minimum number of relay nodes for cost efficiency. In the second
phase, we aim at finding a set of a relatively small number of candi-
date positions. We do this by optimizing the relay node placement
on these positions to achieve the maximum backbone connectivity
for guaranteed lifetime period and within a limited cost budget. This
two-phase deployment scheme will provide a reliable interaction with
the network end users monitoring outdoor environments. It will opti-
mize the relay node deployment in forests to detect fires and report
wildlife activities in water bodies to record events concerning floods,
water pollution, coral reef conditions, and oil spills. In addition, it
will target other rural and hazardous areas such as deserts, polar and
volcanic terrains, and battlefields.
Major contributions of this chapter are listed as follows. We explore
the most suitable lifetime definition in environment monitoring. The
appropriateness of the proposed definition is evaluated and compared
to other lifetime definitions in the literature based on harsh environ-
mental characteristics. We introduce a generic 3-D relay node place-
ment problem which aims at maximizing connectivity with constraints
on wireless sensor network cost and lifetime. We propose an efficient
two-phase solution for the 3-D deployment problem, which considers
a limited search space, generic communication model, most appropri-
ate lifetime definition, and harsh operational conditions. Performance
of the proposed two-phase solution is evaluated and compared to cur-
rently used strategies in environmental applications in the presence of
varying probabilities of node failure and disconnection.
The remainder of this chapter is organized as follows. In Section
3.1, related work is outlined. Practical system models and placement
 EF FI CIEN T D E P L OY M EN T O F WIREL E S S 25

problems are presented in Section 3.2. In Section 3.3, our two-phase

deployment strategy is described. The performance of the proposed
strategy is evaluated and compared to other deployment strategies
in Section 3.5. Finally, conclusions and future work are given in
Section 3.5.

3.1 Related Work

Extensive work has been reported in the literature relating to relay

node deployment strategies which are classified into random vs. grid-
based deployments [37]. In random deployment, nodes are randomly
scattered and are organized in an ad hoc manner. In grid-based
deployment, nodes are placed on grid vertices leading to more accu-
rate positioning and data measurements. In addition, applying grid
deployments in a 3-D space has several other benefits. Such deploy-
ments precisely limit the search space of relay node positions and pos-
sible paths between them. Thus, the formulation of the placement
optimization problem is simplified. Moreover, grid models can reflect
channel conditions using specific signal propagation models that con-
sider harsh environment characteristics; hence, better connectivity
can be maintained. Due to the interest of environmental applications
in the exact physical positioning of sensor and relay nodes, grid-based
deployment is the most appropriate option and is adopted for this
work. However, more efficiency is required in grid-based deployments
to enhance deployed node connectivity and limit the huge number
of candidate positions (or grid vertices) in large-scale environmental
applications.
Connectivity of the deployed grid-based network could be pre-
sented as k-connectivity. k-connectivity has two different meanings,
namely, k-path connectivity and k-link connectivity [17]. The k-path
connectivity means that there are k independent paths between
every pair of nodes, while k-link connectivity means that each node
is directly connected to k neighboring nodes. Nevertheless, a wire-
less sensor network could be disconnected even if k-link connectivity
is satisfied. With k-path connectivity where k ≥ 1, the network can
tolerate some node and link failures. At the same time, the higher
degree connectivity improves communication capacity among nodes.
In some cases, it may not be necessary to maintain k-connectivity
26 WIREL E S S SENS O R NE T W O RKS

among all of the network nodes, but only among nodes which form
the communication backbone of the network, and thus is adopted
in this chapter. Nodes constructing the network backbone are called
irredundant (critical) nodes in this research. Failure of connectivity
between these nodes may lead to severe effects on the WSNs’ perfor-
mance such as network partitioning and data loss. Accordingly, node
redundancy in Reference [8] is used to overcome such connectivity
problems. Redundant nodes are deployed; the ones not being used for
communication or sensing are turned off. When the network becomes
disconnected, one or more of the redundant nodes is turned on to
repair connectivity. In Reference [22], the lowest number of redun-
dant nodes is added to a disconnected static network, so the network
remains connected. Similarly, authors in Reference [20] focus on
designing an optimized approach for connecting disjointed WSNs
segments by populating the least number of relays. The deployment
area is modeled as a grid with equal-sized cells. The optimization
problem is then mapped by selecting the fewest count of cells to pop-
ulate relay nodes such that all segments are connected. Overlapping
clusters of sensor nodes, which rely on the concept of redundancy as
well, are then used to enhance the network connectivity in Reference
[37]. In Reference [1], a distributed recovery algorithm is developed to
address 1- and 2-connectivity requirements. The idea is to identify the
least set of nodes that should be repositioned in order to reestablish a
particular level of connectivity. Nonetheless, redundant node deploy-
ment becomes an intricate task in huge 3-D spaces, where numer-
ous positioning options are possible with different connectivity levels
(degrees). Therefore, more efforts are required to optimize the deploy-
ment process under such circumstances.
Meanwhile, grid-based deployment should not only guarantee
connectivity in harsh environment monitoring but also should guar-
antee specific network lifetime. Lifetime has several definitions in
the literature. One of the most common lifetime definitions states
“it is the time till the first node death occurs” [15]. Such a defini-
tion may not be appropriate if we are monitoring forest temperature
or humidity because if a node dies, we can still receive similar (or
redundant) information from other nodes in the same area. Therefore,
this definition may only be considered as a lower boundary of other
lifetime definitions and should not reflect the actual network lifetime.
 EF FI CIEN T D E P L OY M EN T O F WIREL E S S 27

Similarly, the definition, “it is the time till the last node death occurs,”
can serve as an upper bound of other definitions. Both of these defini-
tions are unrealistic for environmental applications. Another way to
define lifetime could rely on the percentage of alive nodes (which have
enough energy to accomplish their assigned tasks) [15]. But choosing
a percentage threshold is usually arbitrary and does not reflect the
application requirements. In addition, being alive does not mean that
the node is still connected. Therefore, other approaches to evaluate
network lifetime are relying on connectivity and network partitions
[9]. Still these definitions do not satisfy the environment monitoring
applications. For instance, if a set of nodes is destroyed because of the
movement of wild animals or falling trees in the monitored site, the
network can still be considered functional as long as other nodes are
still alive and connected. In References [6,32], the lifetime definitions
rely on the percentage of covered area in the monitored site. But they
are not suitable for environmental applications which are data-driven.
In data-driven applications, we are more interested in sampling data
from the monitored site rather than providing full coverage. Tightly
connecting these data gathering nodes can reduce redundant informa-
tion transmissions, and, hence prolong the overall network lifetime.
Motivated by the benefits of device heterogeneity, as well as the
3-D grid model, our research provides an efficient grid-based deploy-
ment for provisioning WSNs of maximum backbone (critical nodes)
connectivity degree under lifetime constraints and limited cost budget
in environmental applications. For more efficiency and unlike other
grid-based deployments, we find the most feasible grid vertices to be
searched for are those with optimal deployment rather than search-
ing a massive number of grid vertices in large-scale environmental
applications.

3.2 System Models and Problem Definition

In this section, we outline our assumed wireless sensor network models.

Additionally, we introduce a general definition for the targeted relay
node deployment problem. We assume hierarchical network architec-
ture to address the node heterogeneity problem. A graph topology is
considered for easy network extension and accurate (mathematical)
connectivity computation. Furthermore, a detailed discussion of the
28 WIREL E S S SENS O R NE T W O RKS

utilized cost and communication models is proposed, and the appro-

priateness of the considered lifetime definition is examined.

3.2.1 Network Model and Placement Problem

In this chapter, a two-layer hierarchical architecture is assumed to

be a natural choice in large-scale environmental applications as well
as providing a more energy-efficient deployment plan. The lower
layer consists of sensor nodes that sense the targeted phenomena and
send measured Rdata to cluster heads (CHs) in the upper layer, as
shown in Figure 3.1a. These sensor nodes usually have fixed and lim-
ited transmission ranges and do not relay traffic in order to conserve
more energy. The upper layer consists of cluster heads and relay nodes
which have better transmission range (=r) and communicate periodi-
cally with the base station to deliver the measured data in the lower
layer. Cluster heads aggregate the sensed data and coordinate the
medium access; they also support relay nodes in relaying data from
other CHs to the BS in the upper layer. Assuming sensor nodes have
enough energy to perform their effortless tasks, we focus this work on
the upper layer devices which are relay nodes and cluster heads. The
topology of the upper layer is modeled as a graph G = (V, E), where
V = {n0, n1, …, nnc} is the set of nc candidate grid vertices, E is the set
of edges in graph G, and (i, j) ∊ E, if nodes at ni and nj have enough
probabilistic connectivity percentage to establish a communication

(a) (b)

Upper
layer

Lower
layer

SN CH RN BS

Figure 3.1 (a) Two-layer hierarchical architecture and (b) Cubic 3-D grid model for the targeted
wireless sensor network deployment where dashed lines and empty circles represent grid edges and
vertices, respectively.
 EF FI CIEN T D E P L OY M EN T O F WIREL E S S 29

link (edge). We assert that deployment of relay nodes in this research

is independent of the underlying Medium Access Control (MAC)
protocol, where we assume a transmission rate limit T for each node
during a one-time unit (measured in days). This limit can be adjusted
to comply with any MAC protocol. For simplicity and without losing
generality, we assume S-MAC protocol [35] is handling the medium
access in this research with a 50% duty cycle and with only 6 bytes
controlling fields in the exchanged packets for more energy savings.
Moreover, the assumed traffic generation model is flexible enough to
support different requirements of numerous WSN applications as we
have specific parameters, T and Y, controlling the assumed packet
transmission rate and arrival rate respectively. For example, if we
increase T, packets will be generated more frequently. If we decrease
Y, packets will take more time to reach the destination; thus, a more
congested network can be experienced. In this research, we assume
that the arrival rate Y is following a passion process which is very
natural in WSN simulations.
Figure 3.1b depicts the 3-D grid model assumed in this chapter;
the grid edge length is supposed to be equal to a relay node trans-
mission range r. It is assumed that all relay nodes have a common
transmission range r. We think our deployment planning is applicable
for other types of grid models, not just the cubic one. In this cubic
grid model, each sensor node (SN) is placed near to phenomena of
interest for more accurate estimates in terms of the spatial proper-
ties of the collected data. Cluster heads (CHs) are then placed on the
most appropriate grid vertices, which can serve the largest number of
sensor nodes distributed around each cluster head. The base station is
placed based on the application requirements in a fixed position and
is the data sink for the system. Then, we seek to optimize the relay
node positions on the 3-D grid to get cluster heads connected to the
base station efficiently, that is, in terms of cost and network lifetime.
Hence, we define a general relay node placement problem in environ-
mental wireless sensor networks as follows:
Problem Statement Definition: Given a specific sensing task with
pre-specified SNs, CH and BS locations, determine the positions of RNs so
that connectivity between CHs and BS is maximized while lifetime and
cost constraints are satisfied.
30 WIREL E S S SENS O R NE T W O RKS

3.2.2 Cost and Communication Models

Device cost in environmental applications depends on its function-

alities and hardware components. The more functionality the device
has, the more complex and expensive it is. As relay nodes are assumed
to have more functionality and dominate other devices in terms of
transmission range, the cost is modeled in this chapter by the number
of relay nodes placed in the monitored site. We assume identical cost
for these RNs.
For the communication model, we consider a probabilistic con-
nectivity between the deployed devices, in which wireless signals not
only decay with distance, but also are attenuated and reflected by sur-
rounding obstacles such as trees, animals, hills, etc. Accordingly, the
communication range of each device must be represented by an arbi-
trary shape as depicted in Figure 3.2. For realistic estimation of the
ability to communicate within this arbitrary shape, we need a signal
propagation model that reflects the effects of the surrounding obsta-
cles and the environmental characteristics on the propagated signals.
This model can describe the path loss* in the monitored environment
as follows [28].
Pr = K 0 − 10γ log ( d ) − µd (3.1)
where Pr is the received signal power, d is the Euclidian distance
between the transmitter and receiver, γ is the path loss exponent
­calculated based on experimental data, µ is a random variable that

y
z

Figure 3.2 An arbitrary shape of the communication range in 3-D space, due to attenuation and
shadowing affecting outdoor wireless signals.

* Path loss is the difference between the transmitted and received power of the signal.
 EF FI CIEN T D E P L OY M EN T O F WIREL E S S 31

follows a log-normal distribution function with zero mean and vari-

ance δ 2 to describe signal attenuation effects* in the monitored site,
and K0 is a constant calculation based on the transmitter, receiver, and
monitored site mean heights.
Let Pr equal the minimal acceptable signal level to maintain con-
nectivity. Assume γ and K0 in Equation 3.1 are also known for the
specific site to be monitored. Thus, a probabilistic communication
model which gives the probability that two devices separated by dis-
tance d can communicate with each other is given by
γ
Pc = Ke − µd (3.2)
where K 0 = 10log ( K )
Thus, the probabilistic connectivity Pc is not only a function of the
distance separating the wireless nodes but also a function of the sur-
rounding obstacles and terrain, which can cause shadowing and mul-
tipath effects (represented by µ ). Thus, the ability to communicate
between two nodes is defined as follows:
Probabilistic Connectivity Definition: Two nodes (devices) i and
j, separated by distance d, are connected with a threshold parameter
τ (0 ≤ τ ≤ 1), if Pc (i , j ) ≥ τ.
Note that this communication model is generic in terms of the
parameters (K, µ , τ, and γ ), which specify the surrounding environ-
ment characteristics. Setting these parameters to values obtained from
experimental data would provide more practical connectivity estima-
tion, and thus more efficient deployment planning.

3.2.3 Lifetime Model

Models in the literature differ in the way they consider a wireless

sensor network to be still operational. These models can rely on
­connectivity of the deployed nodes or on percentage of alive nodes
(which have enough energy to accomplish their assigned tasks) in the
network. Connectivity-based (CB) and Percentage of Alive Nodes (PAN)
models are defined as follows:

* Wireless signals are weakened because of shadowing and multipath effects. This
refers to the fluctuation of the average received power.
32 WIREL E S S SENS O R NE T W O RKS

Connectivity-Based Definition: Lifetime of a WSN is the time span

from deployment to the instant when a network partition occurs *.
Percentage of Alive Nodes Definition: Lifetime of a WSN is the
time span from deployment to the instant when the percentage of alive nodes
falls below a specific threshold.
However, losing a few nodes may not significantly affect the over-
all wireless sensor network performance especially when redundant
nodes and communication links (edges) are used in tolerating high
probabilities of disconnected nodes as in environmental applications.
In environment monitoring, several nodes are generally assigned to
measure single specific criteria of the monitored space, such as tem-
perature in forestry fire detection [31]. Consequently, the concept of
node redundancy should be addressed. Furthermore, lifetime models
relying on the above definitions do not take into consideration the
node type which could be cluster head, relay node, or sensor node.
Therefore, we propose the following Environment-specific (Env.) life-
time definition.
Environment-Specific Lifetime Definition: Lifetime of a WSN is
the time span from deployment to the instant when the percentage of alive
and connected irredundant nodes is below a pre-defined specific threshold.
Using this definition, we benefit from device redundancy by con-
sidering the network to be operational as long as a specific percentage
of CHs providing the targeted data is still alive. These CHs need
not only be alive but also must be connected to the base station via
single or multi-hop path(s). Note that a cluster head i is connected to
another node j, if Pc (i , j ) ≥ τ, according to the probabilistic connectivity
definition.
In order to mathematically translate the aforementioned lifetime
definitions, we assume number of rounds for which a wireless sensor
network can stay operational as the unit measure of the network life-
time. A complete round is defined in this chapter as the time span tround
in which each irredundant cluster head (i.e., responsible for different
sensor nodes) transmits at least once to the base station without violat-
ing cutoff criteria of the lifetime definitions. tround is identical for all
rounds due to a constant data delivery assumed per round. In addition,

* Network partition occurs when one or more nodes are not able to communicate with
the base station.
 EF FI CIEN T D E P L OY M EN T O F WIREL E S S 33

we adopt the general energy consumption model proposed in Reference

[35] in which energy consumed for receiving a packet of length L is

J rx = Lβ (3.3)

and the energy consumed for transmitting a packet of length L for

distance d is:

(
J tx = L ε 1 + ε 2 d γ ) (3.4)
where ε 1, ε 2 and β are hardware specific parameters of the utilized
transceivers, and γ is the path loss exponent.
Based on Equations 3.3 and 3.4, in addition to knowing the initial
energy Ei of each node with its relative position to other nodes, we
can calculate the remaining energy Er per node after the completion
of each round by

Er = Ei − TJ tx − RJ rx − AJ a (3.5)

where T, R, and A are the arrival rates of transmitted, received and

aggregated packets per round respectively. This follows a Poisson dis-
tribution, and J a is the energy consumed for a single packet aggrega-
tion. Considering Er calculated in Equation 3.5 and assuming the
cutoff criterion associated with each lifetime definition is represented
by a binary variable* C, we can calculate the total number of rounds
for which a wireless sensor network can stay operational.
To assess the environment-specific definition, we use simulation to
compare it to connectivity-based and percentage of alive nodes defini-
tions using four main performance metrics: (1) percentage of alive CHs,
(2) percentage of disconnected CHs/RNs, (3) ratio of remaining energy
(RRE), and (4) total rounds, Percentage of alive CHs is the percentage
of cluster heads which have enough energy to aggregate and forward
data to the base station at least once. Percentage of disconnected CHs/
RNs is the percentage of cluster heads and relay nodes which have
enough energy to aggregate and forward data at least once but are not
able to communicate with the base station. Ratio of remaining energy is
the ratio of total energy amount still available at all nodes (CHs/RNs)

* The cutoff criterion is not satisfied and the network is still considered operational if
C = 0 and vice versa if C = 1.
34 WIREL E S S SENS O R NE T W O RKS

to the total energy at deployment when the network is not opera-

tional. The network is not operational when the cutoff criterion of the
lifetime definition is satisfied. Finally, total rounds is the total num-
ber of rounds in which a wireless sensor network can be considered
operational. These four performance metrics are chosen to reflect the
ability of environment-specific lifetime definition to (1) accurately esti-
mate the network lifetime and (2) effectively utilize the network (i.e.,
maximize network operational time by delaying the assumption of the
network death.) Thus, better energy and resource utilization can be
achieved. Using MATLAB®, we simulate randomly generated wire-
less sensor networks which have the hierarchical architecture and the
graph topology proposed in Section 3.3.1. Each generated network
consists of 12 CHs and a total of 50–80 RNs which are randomly
deployed on grid vertices in 700 × 700 × 200 (m3) 3-D space using a
Linear Congruential random number generator. The parameters used
in the simulations are listed in Table 3.1.
Therein, τ is set to high value for practicality in simulating fluctu-
ated and attenuated signals in environment monitoring applications
[28]. For simplicity, we apply cubic 3-D grid model with identical
grid edges of the length equal to 100 (m). We assume a predefined
fixed time schedule for traffic generation (=100 packets per round
from each CH) and a (PNF) varying from 10% to 60%. We define
the PNF as the probability of physical damage for each node in the
network which is very common in outdoor environment monitoring.
Thus, a higher PNF indicates a higher possibility for the node to be
damaged while still having enough energy to sense and communi-
cate. We intend to assume a very high PNF (up to 60%) to reflect

Table 3.1 Parameters of the Simulated WSNs

PARAMETER VALUE PARAMETER VALUE
T 70% L 512 (bits)
nc 110 (vertex) Ei 15.4 (J)
ε1 50e−9 (J/bit) T 100 (packet/round)
ε2 10e−12 (J/bit/m2) Pr −104 (dB)
β 50e−9 (J/bit) tround 24 (h)
γ 4.8 K0 42.152
δ2 10 r 100 (m)
Ja 50e−7 (J) PNF 10%–60%
 EF FI CIEN T D E P L OY M EN T O F WIREL E S S 35

some actual situations in outdoor environmental applications. Three

different cutoff criteria are used for the simulated network to be con-
sidered operational. According to environmental-specific (Env.) life-
time definition and the proposed network model, the network is still
operational as long as the percentage of connected irredundant cluster
heads, which have enough energy to communicate with the base sta-
tion, is greater than or equal to 50%. The CB definition considers the
network nonoperational when one or more irredundant cluster heads
are unable to reach the base station. Finally, the network is not opera-
tional, based on the PAN definition, when 50% or more of the nodes
run out of energy. After the network is considered not operational, we
measure these performance metrics. This experiment is repeated 500
times. The average results are reported in Figures 3.3 through 3.6. We
observe that the average results hold a confidence interval no more
than 5% of the average (over 500 runs) at a 95% confidence level.
Figure 3.3 shows the percentage of disconnected nodes obtained
when the network becomes nonoperational. Obviously, CB definition
underestimates the network lifetime by considering it nonoperational

40

35
Percentage of disconnected CHs/RNs (%)

30

25

20

15

10 Env.
CB

5
–10 0 10 20 30 40 50 60 70 80
Probability of node failure (%)

Figure 3.3 Percentage of disconnected CHs/RNs vs. different probabilities of CH/RN failure.
Exploring the Variety of Random
Documents with Different Content
Fig. 238.—History of St. Benedict.—On the left are the monks of a
neighbouring monastery, who have come to seduce him from his
hermitage in order to place him at their head; but the austerity of his rule
soon dissatisfies them, and they resolve to rid themselves of him. On the
right the monks are offering him a cup of poison, but, on his making the
sign of the cross on the vase, it is shattered to pieces.—From a Fresco by
Spinello d’Arezzo (1390) in the Church of San Miniato, near Florence.
 Before the monkish order had been in existence a century, in the
East as in the West, the monastic regulations underwent
considerable relaxation. The monks having become part of the
clerical hierarchy (by force of circumstances, for there was often a
deficiency of clergy), took precedence of the latter; their abbots,
called archimandrites in the Eastern Church, were raised to the
priesthood and to the episcopate; they even took part in councils,
though these functions and duties interfered with their cenobitic life.
This manifest infraction of the primitive discipline, while it lowered
the moral position of the monks, rather increased than otherwise
their social influence, and gave them greater weight in the world.
Their piety, too, was only temporarily distracted from its original
purpose, for men of note like St. Honoratus, St. Maximus, St. Hilary,
St. Dalmatius, the two brothers Romanus and Lupicius, maintained
the true tradition of monkish life; and the famous abbeys of Lérins
and Mount Jura were built. The ascetics of Constantinople, too, were
spoken of in high terms, as keeping up a perpetual psalmody (401–
405). In Palestine, not far from Jerusalem, a multitude of hermits,
under the guidance of St. Euthymius, practised the most rigorous
abstinence.
In Africa, St. Fulgentius, exiled by the Arians, was the promoter of
regular observances—that is, he preached strict obedience to
monastic rule (501–523); while in the West there were founded, in
the midst of the Romagnol Alps, in the towns of Arles and St.
Maurice d’Agaune, three model monasteries, the first
superintendents of which were St. Hilary, St. Cæsarius, and St.
Severinus; and its principal benefactors, Theodoric, King of the
Goths, Theodoric the Great, and Sigismund, King of Burgundy (504–
522). In the monastery of Kildare, governed by St. Bridget, and in
the monastery founded by St. Colomba, in Ireland, which was
afterwards so justly called the Isle of Saints, the teaching of
Christian art, of the liturgy, of ecclesiastical lore and profane
literature, was unequalled in its perfection, and the fame of it
reached even to Gaul.
 Fig. 239.—History of St. Benedict.—As his disciples were attempting to put
a stone in place for the construction of their chapel, the devil placed
himself upon it, and the united efforts of several persons failed to
dislodge him; but St. Benedict having blessed the stone, the devil
took flight.—From a Fresco by Spinello d’Arezzo (1390), in the Church
of San Miniato, near Florence.
Such was the general position of cenobitism when St. Benedict
(Fig. 238), the future patriarch of the monks of the West, the
supreme legislator of the monastic order, abandoned his humble cell
at Subiaco (528) to found the immense abbey of Monte-Cassino
(Fig. 239), which was the glory of the age. The Benedictine rule, the
result of profound physiological and philosophical studies, a work of
moral science, of wisdom, and of piety, divided the monks’ time
between prayer and manual labour, to be succeeded by the
cultivation or exercise of the intellect whenever the glory of God, the
interests of the monastery, and the education of the people might
require it. St. Benedict soon had under his control an army of monks
who spread throughout the whole Christian world the rules of their
illustrious chief. Amongst them were St. Maurus and Cassiodorus,
the former minister of Theodoric the Great: one of them founded the
monastery of St. Maur-sur-Loire, in France; the other, that of Vivieri,
in Calabria. Cassiodorus took great pains to collect books of the Old
and the New Testament, with their commentaries. He went to great
expense in collecting all the writings of the Greek and Latin Fathers,
of the Jewish historians as well as of those of the Church, and the
principal works on geography, grammar, and rhetoric, and even the
best treatises on medicine, so that the monks attached to the
infirmary might be fully capable of tending the sick. The monastery
of Vivieri contained one of the richest libraries of the period. We find
in the collection of the Institutions of Cassiodorus the following
remarkable homage paid to the calligraphist monks, who were the
greatest men of letters in that day: “I confess, my brethren, that of
all your physical labours, that of copying books has always been the
avocation most to my taste; the more so, as by this exercise of the
mind upon the Holy Scriptures, you convey to those who will read
what you have written a kind of oral instruction. You preach with the
hand, converting the fingers into organs of speech, announcing
silently to men a theme of salvation; it is as it were fighting the evil
one with pen and ink. For every word written by the antiquary, the
demon receives a severe wound. At rest in his seat, as he copies his
books, the recluse travels through many lands without quitting his
room, and the work of his hands has its influence in places where he
has never been.”
Those whom Cassiodorus calls antiquaries were simply scribes—
that is to say, clerks or monks who deciphered the old manuscripts
and transcribed the books. In the monastery of St. Martin, at Tours,
calligraphy was the sole art practised. St. Fulgentius, a prelate
distinguished both for his learning and eloquence; St. Gregory,
Bishop of Agrigentum, no less celebrated; Mamertus Claudius, who
was a regular walking library, themselves copied manuscripts which
they gave to the Church. Calligraphy and illuminating were also
favourite occupations with many nuns, amongst whom may be cited
St. Melanie the younger, St. Cesarie, St. Harnilde, and St. Renilde, all
Frenchwomen (Fig. 240), who, to use the language of the Christian
annalist, wrote with elegance, rapidity, and correctness.
From the time that the monks were raised to the clerical rank,
after first undergoing an examination, the clerks and monks studied
together just as they had before prayed and lived in common, a
monastery being a complete school of ecclesiastical research and
administration. At Monte-Cassino, at St. Ferréol, at St. Calais, at
Tours, and in many other flourishing abbeys in the sixth century, the
monks, and especially the novices, were instructed in religious and
secular subjects as well as in the duties of the priesthood.
The monastic dress was not in every case the same, for, though
always simple and coarse, it varied in shape and appearance with
the statutes of each order, and according to the necessities of
climate. The cenobites in Egypt wore the lebitus or the colobium, the
pera or melote, and the cuculla. The lebitus was a linen garment
with long sleeves open at the hands, and sometimes up to the wrist.
The pera, a jacket of goatskin, is spoken of in one of the epistles of
St. Paul, who alludes to it as especially worn by holy men and
prophets, when they were driven by threats of persecution into the
desert. The cuculla covered the head, and came half-way over the
shoulders. St. Benedict, who borrowed it from the early monks, had
it so much lengthened as to envelop the whole body; but as in this
shape it would have embarrassed the monks in their manual labour,
he made it a garment only to be worn at ceremonials, and replaced
it for ordinary wear by the scapulary (scapulum), which covered the
head and the back. The Western monks also wore a short mantle—a
sort of cape, called a maforte, according to Sulpicius Severus. The
Greeks and Orientals adopted the pallium, which led to their being
designated agmina palliata (an army in robes), when they assembled
in large numbers. Every Greek who devoted himself to the cenobitic
life was compelled to wear a black pallium.

Fig. 240.—St. Radegonde, Wife of King Clotaire (Sixth Century),

receiving the religious garb from the hands of St. Médard,
Bishop of Noyon.—“Histoire et Cronicque de Clotaire” (16mo,
Paris, Jean Mesnage, 1513).
 Pope Gregory the Great, who had been a Benedictine, was most
ardent in the establishment of monasteries, of which he himself
founded a large number. He was the chief promoter of two
important missions which took place in 585 and 596; the first in
Gaul, consisting of missionaries from Ireland, headed by St. Columba
and St. Gall; the second in Great Britain, with monks from the Abbey
of St. Andrew, headed by another monk, St. Augustine. This latter,
who converted the Anglians and their king, Ethelbert, was the first
Archbishop of Canterbury. Colomba founded the Abbey of Luxeuil,
upon the southern side of the Vosgian forests; while Gall, his
disciple, much younger than himself, penetrated into the country of
the Helvetians, who were as deeply sunk in barbarism as the
Anglians, where he founded a monastery which afterwards became
famous under the name of its founder, and which owed its celebrity
to the variety of subjects which were taught there.
St. Colomba was the first to draw up a complete set of monastic
rules, which were generally adopted in France, just as the rules of
St. Isidore, Bishop of Seville, and those of St. Augustine of Ireland,
were followed in the British Isles. These three codes, very similar in
their general principles, varied from each other in many particulars,
for they were applicable to monks living in different countries. In the
communities which adhered to the rules of St. Colomba, as in all the
great Benedictine monasteries, prayer, mental culture, and manual
labour were the invariable occupations of cloister life (Fig. 241). The
rules drawn up by St. Colomba and his imitators, St. Isidore and St.
Augustine, thus remained in force down to the eighth century, in
spite of the new system of education and religious teaching
inaugurated so zealously throughout Gaul by the Anglo-Saxon monk,
St. Boniface; in spite, too, of the industrial and artistic, rather than
scientific and contemplative, turn which St. Eloi gave to the studies
of the monks in his abbey of St. Martin at Limoges, and in the other
monasteries founded or reorganized by him. That illustrious Bishop
of Noyon, Master of the Mint to King Clotaire II., afterwards
treasurer, goldsmith, and minister to Dagobert I. (568–659), was at
great pains to make the cultivation of art an important feature in
monastic life.
It would be incorrect to suppose that the interior of a monastery
in the seventh century presented the same appearance of asceticism
and penance that were afterwards characteristic of certain
communities subject to the most austere regulations. In the country
districts the monasteries possessed vast domains which yielded
wheat, rye, oats, hay, vegetables, and fruits; and on which were
produced wine, beer, cider, and hydromel; they were tilled by
numerous labourers in bands of tens and hundreds, who while at
work sang hymns and prayers—a veritable religious militia, grouped
beneath the banner of faith in the populous centres and in the
neighbourhood of the towns. These monasteries were generally
schools in which the monks gave gratuitous education, vast
workshops in which they followed and taught every branch of trade
—carving in wood, ivory, bronze, silver, and gold; painting on vellum,
glass, wood, and metal; weaving tapestry, embroidering church
ornaments and vestments; damask work, and enamelling of shrines,
tabernacles, diptychs and triptychs, church furniture, and book-
covers; the cutting of precious stones to prepare them for setting;
the making of arms and instruments of music, illuminating, copying
of manuscripts, &c. The whole life of a monk or nun was passed in
the exercise of one description of art, or perhaps even in executing a
single work which required miraculous patience.
 Fig. 241.—The Abbey of St. Riquier, near Abbeville, founded in
799 by St. Angilbert, who gave it its triangular shape in
honour of the Trinity.—From a Drawing in a very old
Manuscript engraved in the Dissertation of Paul Petau, “De
Nithardo” (4to, 1612).
As the regular associations became permanently settled in the
towns, they began to construct for their use dormitories, cells,
workshops, granaries or sheds for their provisions, and built
handsome churches with long cloisters and vast chapter-rooms. Each
community made a point of having within its own boundary a library,
a study, a lecture-room, schools, a cemetery, some shady walks for
meditation, as well as a fruit and kitchen garden, the cultivation of
which was a healthy and agreeable recreation. In this vast
aggregation of monastic buildings and appurtenances (Fig. 244), we
have a holy city in the heart of the secular town, a retreat for the
peaceful, the devout, and the abstinent, amidst the troubles and
vanities of the world.

Fig. 242.—Abbatial Ring and Cross (front and back) of St. Waudru,
patroness of Mons, who died in 670. The cross is in silver, with
gold relief, and studded with precious stones.—Relics
preserved in the Church of St. Waudru, at Mons.
Fig. 243.—The Offering of a Child to an Abbot.—From a
Miniature in a Manuscript published at the Close of
the Thirteenth Century (Burgundian Library,
Brussels).
Fig. 244.—Priory of the Benedictines at Canterbury (Twelfth
Century), plan in relief drawn by Edwin, a monk, about the year
1530.—A, belfry; B, fountain; C, cemetery; D, reservoir, with
conduit pipes; E, Canterbury Cathedral; F, vestry; G, crypt; H,
chapter-house; I, prior’s house; J, infirmary and annexes; K,
kitchen-garden, with well, pumps, and water-pipes; L, cloister;
M, cellar; N, dormitory; O, refectory; P, kitchens; Q, parlour; R,
house for the guests and the poor; S, water-closets; T, baths;
U, granary; V, bakehouse and brewery; X, the chief entrance; Y,
 Z, fortified wall of the abbey and the city.—From an Engraving
in vol. i. of the “Architecture Monastique,” by M. Albert Lenoir.
The endowment of each monastery was generally made up of the
property belonging to the monks who had fixed their abode there. If
the novice was an adult, he was obliged to distribute all his goods to
the poor, or to make a solemn grant of them to the abbey, before he
could be admitted to the minor orders. If he was a child whose
parents devoted him to the service of God (Fig. 243), the parents
either made no gift to the community which received the young
novice, or they ceded the income of the lands and the property by
deed of transfer to the monastery. Enriched by these successive
donations, the monasteries, especially those which had acquired a
wide renown for learning or piety, acquired still more wealth through
the largesses of princes, great nobles, and bishops, through the
economical management of the abbots, and the annual produce of
the agricultural and commercial labour of the monks. To the various
arts and trades which were at first carried on by the monks with a
view to do honour to the cause of religion, those of the West
afterwards added others of a more lucrative and worldly character.
In the sixth century we find that they spun and wove their own silk;
that they possessed numerous receipts for preparing liqueurs and
drugs; that they practised medicine, surgery, and the veterinary art.
Pepin the Short, suffering from incurable dropsy, went first to the
monastery of St. Martin at Tours, and afterwards to the abbey of St.
Denis, that “the servants of God” might give him relief by means of
their skill as well as by their prayers.
The Church was very sorely tried during the reign of Charles
Martel, and monastic institutions were also exposed to many
difficulties. In order to win back the secular clergy to the habits of
communal life, the wisest of the bishops grouped around them the
clergy who had remained faithful to their cause, and laid down a
code of regulations for their guidance.
Charlemagne, in the Capitularies, added the following excellent
amendments to the rules of monastic institutions:—“Young men
destined to monastic life must first pass their novitiate, and then
remain in the monastery to learn the rules, before they are sent
forth to fulfil their duties outside. Those who give up the world in
order to avoid the king’s service shall be compelled to serve God in
good faith, or else to resume their former occupation. All clerks shall
be required to make their choice between clerical life in conformity
with the canons, and monastic life in conformity with the
regulations. The abbeys shall not receive too large a number of
serfs, so that the villages may not be depopulated; no community
shall have more members than can be properly looked after by one
superior. Young women shall not take the veil until they are of an
age to choose their own career in life. Laymen are to be disqualified
for governing the interior of a monastery, nor shall they fill the post
of archdeacon.” Charlemagne and Louis the Good-natured became
members of the royal monastery of St. Denis under the title of
“conscript brothers” (fratres conscripti)—an academical rather than a
religious title, but one which nevertheless admitted them to certain
liturgical privileges. The Emperor Lothair, in imitation of his father
and ancestor, also got himself invested with this title by the
monastery of St. Martin-lez-Metz.
The Norman invasion, the feudal wars, the encroachment of the
great vassals, and even of the kings, upon ecclesiastical domains
and rights, impoverished the monastic orders, whose lands remained
untilled for want of hands, and their schoolrooms often empty for
want of teachers and scholars. While the Normans burnt and
pillaged the monasteries, fortified though many of them were, in the
country districts, the urban abbeys, nearly always protected by the
diocesan power, preserved some remnants of their former splendour.
Fig. 245.—Foundation of the secular abbeys of Mons, Maubeuge, and Nivelles. The
canonesses meeting at Nivelles, where Walcand, Bishop of Liége (810 to 832
or 836), promises to give them a code of rules.—From the “Chroniques de
Hainaut,” Manuscript of the Fifteenth Century, in the Burgundian Library,
Brussels.
There existed between the principal abbeys of the same order a
spirit of unity, a brotherly zeal to render help and service, and a
reciprocal interchange of learned and skilful clerks, who went from
one community to another to give it the benefit of their learning or
manual ability. It was in this way that the conventual churches and
buildings were erected and kept in repair; that they became rich in
paintings, statues, and mosaics; that the treasury was filled, and the
library founded and maintained. Rupert, a monk of the Abbey of St.
Gall (Switzerland), before his elevation to the bishopric of Metz, a
learned linguist, poet, and man of letters; Tutilo, his contemporary at
St. Gall, a carver, painter, and sculptor; Regino, Abbot of Prüm, an
excellent musician, author of a Treatise on Harmony, are of
themselves a proof that arts and letters were hidden in the cloisters.
At this epoch of barbarism and ignorance, the Church organized
what was good, strengthened the shattered foundations of the social
edifice, established new monastic institutions and reformed the old,
grouped around her the irresolute, lawless, and undisciplined minds
(Fig. 245), selling the principles of order and peace in opposition to
those of violence and disorder engendered by war.
 Fig. 246.—Seal of the Abbey of St. Denis, in the
Twelfth Century, in the National Archives, Paris.—
The saint is clad in his episcopal garb. It is, no
doubt, as the apostle of Gaul that the motto gives
him the title of archbishop.
Never was the monastic order more numerous or better
organized, and at no period, perhaps, were works of mental
intelligence cultivated more ardently or successfully in certain
privileged monasteries, than at this time.
Canterbury, Monte-Cassino, St. Maur, St. Denis (Fig. 246), St.
Martin of Tours, St. Gall, Remiremont, Aix-la-Chapelle, Cologne,
Trèves, St. Trudon, St. Arnulph, St. Clement, and St. Martin of Metz,
the Messinà and the Gorza basilica, were cited as so many foci of
light whence radiated in every direction the good doctrines set forth
in certain remarkable works of art, as well as in learned literary
compositions.

Fig. 247.—The Clergy, with the Cross and the Holy Images, going
in procession before the Emperor.—From a Miniature taken
from a Manuscript of the Fourth Century (Library of M.
Ambroise Firmin-Didot).
In the libraries, which composed the principal wealth of the
religious houses, the cartularies of the diocese were preserved with
great care. The charters of institution, and the patrimonial titles of
the chief abbeys, are both the proof and the reward for the services
rendered to civilisation by the monastic establishments: one abbey
was given a domain on the condition that it put the waste lands into
cultivation; another received its lands with the understanding that it
opened asylums and places of hospitality for the poor and sick, for
pilgrims and for strangers; while a host of documents taken from the
cartularies relate to the instruction of the clerks, the education of the
novices, the splendour of public worship, and the duty of the
ecclesiastical vassals when the suzerain raised the ban and the
arrière-ban, &c., together with all the details of monastic life which
are connected with the various social movements of each territorial
district (Fig. 247).
Outside the abbeys there lived a population whose manual labour
was necessary to their inmates, and profitable to the material
interests of the house. Women, even when doing penance, and
under religious vows, were strictly forbidden to enter the
monasteries. The aged mother of an eminent monk, John of Gorze,
unwilling to separate herself altogether from her son, took up her
abode just outside the walls of his abbey, where she spent her time
in making cloaks for the monks.
It was around the abbatial close, perhaps beneath the shelter of a
second walled enclosure, not so strong nor so high as the first, but
still capable of resisting the attacks of the marauders which were so
frequent in those days of feudal disorder, that were built the shops,
the stalls, and the sheds which served for the sale of the crops, the
cattle, and the agricultural and other produce of the abbatial domain
(Fig. 248). On the anniversary of the festival of the saint to whom
the monastery was dedicated there was a fair—sometimes several—
which attracted large crowds.
St. Romuald, the founder of the Camaldolite order; St. Mayeul,
Abbot of Cluny, and the reformer of the Abbey of St. Denis; St.
Dunstan, the resolute Archbishop of Canterbury, who reformed the
clergy of the British Isles; Adalbert, son of a Duke of Lorraine and
nephew of Hugh Capet, who was elected Bishop of Metz, after
having been monk at Gorze; St. Cadroé, descended from the Kings
of Scotland, Abbot of Vaussey and contemporary of Adalbert, with
whom he was associated in the reformation of the abbeys in the
north-east of France, were among the leading figures who, in the
tenth century, represented the reformed monachism. Unfortunately,
their wholesome influence could not make itself everywhere felt; it
was a period of disorder, of pitiless and bitter wars, of usurpations of
every kind. Upon every side misery reigned supreme; the serfs
attached to the domains of the canonical churches and to the
monasteries left them to find some more certain means of livelihood.
The Cathedral of Metz was in this way deprived of eight hundred
serfs, who were heads of families. The only independent voices
raised on behalf of these victims of oppression came from the great
abbeys, such as Stavelo, St. Arnulph, Cluny, &c., to which monarchs
and popes, under the pretext of dedicating churches (Fig. 249)
which had been recently built or restored, repaired in secret to
consult, with many members of the higher clergy, as to the political
affairs of Christendom.
Fig. 248.—North View of the Abbey of St. Germain-des-Prés, as it still
existed in the Seventeenth Century.—A, outer gates; B, houses in
the enclosure; C, church square; D, church; E, Lady chapel; F,
sacristy; G, small cloister; H, great cloister; I, library; K,
dormitory; L, refectory; M, kitchen; N, dormitory of the Superior;
O, offices; P, inner courtyard; Q, houses for the wine-presses; R,
bakehouse; S, stables; T, garden; V, infirmary; X, infirmary
garden; Y, lavatory; Z, dormitory for the guests. 1, abbey palace;
2, abbey garden; 3, courtyard; 4, outer courtyard; 5, officers’
apartments; 6, stables; 7, barns; 8, houses in the abbatial
enclosure; 9, bailiff’s house; 10, outer gates; 11, bailiwick prisons.
—Fac-simile of an Engraving in the “Histoire de St. Germain-des-
Prés,” by Dom Bouillart, in folio: 1724.
Fig. 249.—Dedication of the Church belonging to the
Monastery of St. Martin-des-Champs, Paris, destroyed
by the Normans and rebuilt by King Henry I. The artist
has represented—1st, the ancient Church of St.
Samson dedicated to St. Martin; 2nd, the counts and
barons who signed the charter for the re-establishment
of the monastery; 3rd, the archbishops and bishops
who were present at the dedication of the new church.
—Fac-simile of an Engraving from Don Meurier’s work,
 “Historia Monasterii regalis Sancti Martini” (4to, Paris,
1636).

Fig. 250.—The Small Cloister of the Chartreuse at Pavia, with the

cupola of the church in the background (close of the Fourteenth
Century).
The two Councils of Rheims and of Mayence (1049), devoted
exclusively to disciplinarian reforms, are characteristic of the state of
monastic institutions at this period, just as the journey of Pope Leo
IX. through France and Germany indicates the exact condition, the
resources, the manners, and the habits of the religious houses. The
illustrious pontiff, when visiting these houses, made them splendid
presents, promised them important privileges, and instituted minute
inquiries into the studies pursued within their walls. At the Abbey of
Gorze, in 1149, he even went so far as to note with his own hand
the nocturn responses in the “Office de Saint-Gorgon.”

Fig. 251.—Saint-Jean des Vignes, an Abbey of Regular Canons at Soissons (1076),

the entrance-gate guarded by a barbican and bastilles.—From an Engraving in
vol. i. of “Architecture Monastique,” by M. Albert Lenoir.
At about the same period, William, Abbot of St. Bénigne de Dijon,
re-established in several dioceses the monastic rules and studies;
Sigebert, a monk in the monastery of Gemblours, came to Metz to
teach the Holy Scriptures, philosophy, and the dead languages; St.
William of Hirsange reformed the cloister discipline in Germany; St.
Robert, Abbot of Molême, founded the Cistercian order; St. Gualbert,
the order of Vallombrosa, in the Apennines; St. Bruno, the
Carthusian order, which he established both in the neighbourhood of
Grenoble and in Calabria.
It is impossible to depict the profound disorder which reigned in
the religious houses during the eleventh century, owing to the social
disturbance which had followed the terrors of the year 1000. There
were but a few solitary monasteries, remote from the troubles and
vanities of the world, which still adhered to the rules (Fig. 251), and
the monastic schools were nearly everywhere closed, and the notes
of song had ceased to be heard in the churches, when, in the year
1095, the inspired voice of a monk, Peter the Hermit, summoned the
Christian peoples to the Holy War. At this voice, which seemed to
come down from heaven, the whole world was stirred up to deeds of
energy; the young were inspired with a current of warlike and
adventurous ideas which converged upon one single object—the
deliverance of the holy places.
The difficulty of managing both the spiritual and temporal affairs
of a monastery or cathedral church had led to the appointment of a
sort of steward or lay administrator, termed an avowee, who was
paid out of the dues which he received from the vassals of the
community. He generally levied on each household a loaf of bread, a
denier, a measure of oats, wheat, or barley, if the land grew cereals;
a measure of wine, beer, or cider if the produce of the domains was
grapes, hops, or apples. The avowee was arbiter in all disputed
cases, and himself fixed the remuneration, before and after giving
his decision, which the two parties to the suit had to pay him. He
presided over judicial duels, and ordeal by boiling water or fire. He
had a right to one head of stock at all cattle fairs, and he also
received a draught or a saddle-horse, according as the district bred
the one kind or the other. The avowee of a cathedral or a monastery
always held a distinguished position in society; barons, dukes, and
counts did not disdain to accept these functions—which they often
abused, it must be added, by keeping for their own use the sums
which they had received for the monastery. The usurpations of every
kind which the avowees committed had been flagrant enough during
the investiture dispute, but they increased enormously during the
Crusades, owing to the absence of so many bishops, archdeacons,
abbots, and priors, who had started for Palestine after loading their
domains with mortgages, and even raising money upon the sacred
vessels of their church.
The Crusades, notwithstanding, had the unquestionable
advantage of sifting the clergy, and of removing from the cloisters a
large number of clerks who were less fitted for study and seclusion
than for the hardships of the battle-field. The monks who remained
in Europe shut up in their cloisters were nearly all acting in
obedience to some special aptitude, and they formed that band of
artists, architects, painters, sculptors, and musicians, calligraphists,
savants, translators, philosophers, rhetoricians, and preachers, which
shed so much lustre upon the monasteries during the twelfth and
thirteenth centuries. By their direct action, as well as by their
example, ecclesiastical architecture made vast progress, and the
wonderful wealth of decoration which accompanied it, suddenly
burst forth in the erection of those holy-chapels which seem like
shrines of chiselled stone, in which the relics brought back from the
Crusades were deposited (Fig. 252). It was under these influences
that most of the great abbeys (Fig. 253) were restored, that painting
upon glass attained its full perfection, that the Roman tongue
reached the solitude of the cloister, and the beautiful literature of the
ancient classics, which had for centuries been relegated to the dust
of the monastic libraries, once more saw the light, and lent the aid
of all its charms to combat the invasion of the vulgar idiom which
the inhabitants of the communes had everywhere substituted for the
Latin language.
Fig. 252.—Reliquary of the Holy
Thorn, preserved in the
Convent of the Augustine
Sisters at Arras.—Carved
brasswork of the Thirteenth
Century.
Fig. 253.—Refectory in the Priory of St. Martin des
Champs, Paris (now part of the Conservatoire des
Arts et Métiers), the work of Pierre de Montereau,
architect to St. Louis (Thirteenth Century).—
Archæological Restoration by M. Alfred Lenoir.
Fig. 254.—St. Bernard taking possession, with the Cistercian
Monks, of the Abbey of Clairvaux. At the foot of the
engraving is inscribed: “St. Bernard, Chaplain of the Virgin
Mary, was descended from the house of the Kings of
Burgundy.” He was, as a matter of fact, related through
his mother, Aleth (diminutive for Elizabeth), to the first
house of the Dukes of Burgundy.—“Chroniques abrégées
de Bourgogne,” a Manuscript of the Fifteenth Century, in
the Library of M. Ambroise Firmin-Didot.