Content uploaded by Nadeem Javaid
Author content
All content in this area was uploaded by Nadeem Javaid on Feb 09, 2019
Content may be subject to copyright.
Multiple Criteria Decision Making based
Clustering Technique for WSNs
By
Mansoor Mustafa
Registration Number: -REE-044/ISB
MS Thesis
In
Electrical Engineering
COMSATS Institute of Information Technology
Islamabad – Pakistan
Spring, 2013
CIIT/FA11
Multiple Criteria Decision Making based
Clustering Technique for WSNs
A Thesis presented to
COMSATS Institute of Information Technology
In partial fulfillment
of the requirement for the degree of
MS (Electrical Engineering)
By
Mansoor Mustafa
CIIT/FA11-REE-044/ISB
SPRING, 2013
COMSATS Institute of Information Technology
ii
Engineering).
Multiple Criteria Decision Making based
Clustering Technique for WSNs
A post Graduate Thesis submitted to Department of Electrical Engineering as
partial fulfillment of the requirement for the award of Degree of M.S
(Electrical
Name Registration Number
Mansoor Mustafa CIIT/FA11-REE-044/ISB
Supervisor
Dr. Safdar H. Bouk
Assistant Professor, Department of Electrical Engineering
Islamabad Campus
COMSATS Institute of Information and Technology (CIIT)
Islamabad
Co-Supervisor
Dr. Nadeem Javaid
Assistant Professor,
Center for Advanced Studies in Telecommunications (CAST),
COMSATS Institute of Information and Technology (CIIT)
Islamabad
Spring, 2013
iii
Final Approval
This thesis titled
Multiple Criteria Decision Making based
Clustering Technique for WSNs
By
Mansoor Mustafa
CIIT/FA11-REE-044/ISB
Has been approved
For the COMSATS Institute of Information Technology, Islamabad
External Examiner: __________________________________
(To be decided)
Supervisor: ________________________
Dr. Safdar H. Bouk/Assistant Professor,
Department of Electrical Engineering CIIT, Islamabad
Co-Supervisor: ________________________
Dr. Nadeem Javaid/Assistant professor,
Center for Advanced Studies in Telecommunications (CAST),
CIIT, Islamabad.
Head of Department:________________________
Dr. Raja Ali Riaz / Associate professor,
Department of Electrical Engineering,
CIIT, Islamabad.
iv
Declaration
I Mansoor Mustafa , CIIT/FA11-REE-044/ISB hereby declare that I have
produced the work presented in this thesis, during the schedule period of study. I
also declare that I have not taken any material from any source except referred to
wherever due that amount of plagiarism is within acceptable ra nge. If a violation
of HEC rules on research has occurred in this thesis, I shall be liable to punishable
action under the plagiarism rule of the HEC.
Date: ___________
_____________
Mansoor Mustafa
CIIT/FA11-REE-044/ISB
v
Certificate
Date: _________________
Supervisor: ________________________
Dr. Safdar H. Bouk/Assistant Professor,
Department of Electrical Engineering CIIT, Islamabad
Co-Supervisor: ________________________
Dr. Nadeem Javaid/Assistant professor,
Center for Advanced Studies in Telecommunications (CAST),
CIIT, Islamabad.
Head of Department:________________________
Dr. Raja Ali Riaz / Associate professor,
Department of Electrical Engineering,
CIIT, Islamabad.
It is certified that Mansoor Mustafa, CIIT/FA11-REE-044/ISB has carried out
all the work related to th is thesis under my supervision at the Department of
Electrical Engineering, COMSATS Institute of Information and Technology,
Islamabad and the work fulfills the requirements for award of MS degree.
vi
DEDICATION
Thisthesisisdedicatedtomyparents
Fortheirendlesslove,supportandencouragement
.
vii
ACKNOWLEDGMENT
Mansoor Mustafa
CIIT/FA11-REE-044/ISB
I thank to Almighty ALLAH for His Blessing and Guidance in completing this project
in time.
.
I would like to admit the great and unconditional academic support and
encouragement from my family. This success is all because of their prayers and
help in my university career.
I would also like to specially thank my highly regarded teacher and supervisor of
thesis Dr. Safdear Hussain Bouk and co-supervisor Dr. Nadeem Javaid for their
utmost help and precious guidance and expert advices during execution of the
work and throughout my studies at COMSATS Institute of Information Technology,
Islamabad.
I am extremely obliged to my research fellows Tauseef Shah , Imran Israr , Aziz-ur-
Rehman, Mohammad Mateen Yaqoob and Haad Akmal, who helped me in some
way or the other in completing my thesis.
At the end my gratitude is for my parents and friends who supported me morally
to reach this stage.
.
viii
ABSTRACT
Multiple Criteria Decision Making based Clustering
Technique for WSNs
En
!
"
# $ %
"&#'(%")
' '
'
$
ix
Table of Contents
1 Introduction 2
1.1 BriefOverview ............................. 2
1.2 Motivation................................ 3
1.3 ProblemStatement ........................... 4
1.4 Research Methodology . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Wireless Sensor Networks (WSNs) 8
2.1 Architecture of WSN Node . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Sensing Subsystem . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.2 Processing Subsystem . . . . . . . . . . . . . . . . . . . . . 10
2.1.3 Communication Subsystem . . . . . . . . . . . . . . . . . . . 10
2.1.4 Power Subsystem . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Challenges in WSNs . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.1 Power Consumption/ Network Lifetime . . . . . . . . . . . . 12
2.2.2 Fault Tolerance . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.3 Scalability ........................... 12
2.2.4 Throughput........................... 12
2.2.5 Accuracy/Latency . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.6 Node Deployment . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.7 Data Aggregation . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.8 Hardware Constraints . . . . . . . . . . . . . . . . . . . . . 13
2.2.9 Security Issues . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Applications of WSNs . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.1 Military Applications . . . . . . . . . . . . . . . . . . . . . . 14
2.3.2 Environmental Monitoring . . . . . . . . . . . . . . . . . . . 14
2.3.3 Medical Applications . . . . . . . . . . . . . . . . . . . . . . 14
2.3.4 Other Applications . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 Summary ................................ 16
x
3 WSN Network Layer 18
3.1 Introduction............................... 18
3.2 Routing Challenges and Design Issues in WSNs . . . . . . . . . . . 19
3.3 RoutingIProtocols in WSNs . . . . . . . . . . . . . . . . . . . . . . 20
3.3.1 Protocols based on Network Organization . . . . . . . . . . 20
3.3.2 ProtocolsIbased onIProtocol Operation . . . . . . . . . . . . 23
3.3.3 Protocols based on Route Discovery . . . . . . . . . . . . . . 24
3.4 Clustering in WSNs . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.4.1 Energy Saving Schemes in WSN Clustering . . . . . . . . . . 25
3.4.1.1 Cluster Formation and Rotation . . . . . . . . . . 25
3.4.1.2 Cluster Head Election and Rotation . . . . . . . . 27
3.4.1.3 Efficient Intra-cluster Communication . . . . . . . 28
3.5 Clustering protocols in WSNs . . . . . . . . . . . . . . . . . . . . . 29
3.5.1 Low-EnergyIAdaptiveIClusteringIHierarchy (LEACH) . . . . 29
3.5.2 Centralized LEACH (C-LEACH) . . . . . . . . . . . . . . . 31
4 Proposed Clustering Scheme 35
4.1 Energy Model for WirelessISensor Node . . . . . . . . . . . . . . . . 35
4.2 ProposedScheme ............................ 36
4.2.1 Network Deployment Phase . . . . . . . . . . . . . . . . . . 37
4.2.2 Neighbor Discovery Phase . . . . . . . . . . . . . . . . . . . 38
4.2.3 CH selection and Cluster Formation Phase . . . . . . . . . . 38
4.2.4 Communication Phase . . . . . . . . . . . . . . . . . . . . . 42
5 Simulation and Results 45
5.1 Simulation Environment . . . . . . . . . . . . . . . . . . . . . . . . 45
5.2 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.2.1 Number of Dead Nodes . . . . . . . . . . . . . . . . . . . . . 46
5.2.2 Number of Alive Nodes . . . . . . . . . . . . . . . . . . . . . 47
5.2.3 Energy Consumption . . . . . . . . . . . . . . . . . . . . . . 47
5.2.4 Change of Cluster Head . . . . . . . . . . . . . . . . . . . . 48
5.2.5 Control Overhead (Hello) Packets . . . . . . . . . . . . . . . 49
5.2.6 Packets Sent to Base Station . . . . . . . . . . . . . . . . . . 49
6 Conclusion 52
References 52
xi
List of Figures
2.1 Wireless Sensor Networks . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 WSN Node Architecture . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Environmental Applications of WSNs . . . . . . . . . . . . . . . . 15
2.4 Medical Application of WSNs . . . . . . . . . . . . . . . . . . . . . 15
3.1 Single-hop routing versus multi-hop routing model . . . . . . . . . . 18
3.2 ClassificationIof WSN RoutingIProtocols . . . . . . . . . . . . . . . 21
3.3 Architecture of LEACH . . . . . . . . . . . . . . . . . . . . . . . . 30
4.1 Wireless Sensor Node Energy Model . . . . . . . . . . . . . . . . . 35
4.2 Sensor Nodes Deployed in Field . . . . . . . . . . . . . . . . . . . . 37
4.3 Procedure for CH Change . . . . . . . . . . . . . . . . . . . . . . . 41
4.4 Inter and Intra-cluster Communication . . . . . . . . . . . . . . . . 42
4.5 Flow Diagram of Proposed Scheme . . . . . . . . . . . . . . . . . . 43
5.1 Number of Dead Nodes . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.2 Number of Alive Nodes . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.3 Energy Consumption of Network . . . . . . . . . . . . . . . . . . . 48
5.4 ChangeofCHs ............................. 49
5.5 Control Overhead Packets . . . . . . . . . . . . . . . . . . . . . . . 50
5.6 PacketssenttoBS ........................... 50
xii
List of Tables
3.1 Difference between flat and hierarchical routing . . . . . . . . . . . 22
3.2 Detailed Comparison of WSN Clustering Protocols [30] . . . . . . . 33
4.1 CriteriaWeights............................. 39
4.2 Fuzzy Membership Functions . . . . . . . . . . . . . . . . . . . . . 39
5.1 Simulation Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 45
xiii
Chapter 1
Introduction
1
Chapter 1
Introduction
Rapid advances in Wireless Sensor Networks (WSNs) have enabled densely deploy-
ment of nodes. WSNs is an emerging technology that consists of large number of
low cost, low power sensor nodes; a sensor node is an electronic device that is ca-
pable of detecting environmental conditions. Those sensor nodes can be deployed
randomly to perform many applications such as monitoring physical events, for
example environmental monitoring, battlefield surveillance, disaster relief, target
tracking, etc. and they work together to form a wireless sensor network.
1.1 Brief Overview
A typical node of a WSN is equipped with four components [1]: a sensor that
performs the sensing of required events in a specific field, a radio transceiver that
performs radio transmission and reception, a microcontroller, which is used for
data processing and a battery that is a power unit providing energy for operation.
The limited energy given to each node, supplied from non-rechargeable batteries,
with no form of recharging after deployment is one of the most crucial problems
in WSNs. Many routing protocols have been proposed for WSNs. Most of the
hierarchical algorithms proposed for WSNs concentrate mainly on maximizing the
lifetime of the network by trying to minimize the energy consumption [1].
Researchers agreed that clustering of nodes in wireless sensor networks is an effec-
tive program of energy conservation [2]. Clustering is defined as the grouping of
similar objects or the process of finding a natural association among some specific
objects or data. In WSNs it is used to minimize the number of nodes that take
part in long distance data transmission to a Base Station (BS), what leads to
lowering of total energy consumption of the system.
2
Clustering reduces the amount of transmitted data by grouping similar nodes
together and electing one node as a Cluster Head (CH), where aggregation of data
is performed to avoid redundancy and communication load caused by multiple
adjacent nodes, then sending the aggregated data to the next CH or to the BS,
where it is processed, stored and retrieved.
In any clustering organization intra-cluster communication as well as inter-cluster
communication can be single hop or multi-hop. However, the hot spot and network
partitioning arises when using multi-hop routing in inter-cluster communications.
Because the CHs close to the BS, are burdened with heavy relayed traffic that will
make them die faster than other CHs, resulting in loss of coverage of sensing.
CH selection in any routing protocol is most challenging issue, because network
lifetime and stability strongly depends on selected CH. Research has proved that
CH selected on single criterion is to not much energy efficient. Hence an ideal CH
is one which is selected on multiple criteria. Solution of using multiple criteria can
be solved using Multiple Criteria Decision Making (MCDM)[3] technique.
MCDM methods are used to solve the decision making problem in eld of engineer-
ing and sciences, with multiple attributes. MCDM techniques compare and rank
multiple alternatives based on degree of desirability of their respective attributes
[3]. There are different types of MCDM approach. Fuzzy logic and fuzzy set the-
ory is applied to decision making process. TOPSIS is a solution for multi-criteria
optimization problem. TOPSIS was initially proposed by Hwang and Yoon [3].
Fuzzy-TOPSIS consists of decision matrix with m number of a alternatives and n
number of attributes for each alternative. This technique is applied is scientic and
engineering problem solving. Fuzzy-TOPSIS uses relative importance of attributes
instead of using precise values, because is some situations precise assignment is
not possible due to any reason.
1.2 Motivation
Low-power electronic devices have grown interest in recent years. WSNs use these
many low-power devices along with communication capabilities for sensing and
monitoring various fields. Major areas of WSNs include environmental sensing
of temperature and humidity, earthquake monitoring, healthcare monitoring and
battle field surveillance [4]. In some applications sensor nodes are deployed in
strategic manner but in most of the applications like battle field, these nodes are
dispersed randomly.
3
After deployment in the field, sensor nodes have to be self organized without hu-
man interference. These sensor nodes consist of battery operated radio devices,
which have limited memory and processing capabilities. Their batteries cannot
be charged or replaced after deployment. Hence, energy is one of the major is-
sues in WSNs. The sensor nodes not only sense data but they also process data
and communicate with the BS. The processing and communication also consume
energy and it is major requirement of WSNs, therefore, it has to is to be energy
efficient. lots of research have been conducted in WSNs on energy efficiency to
prolong network life time and stability. This research includes design of various
energy efficient MAC and routing protocols. Routing protocols may be either be
flat or hierarchical. In flat protocols, each sensor node sends data to BS directly or
in a multi-hop fashion. On the other hand, in hierarchical architecture, WSNs are
divided into optimum number of groups or clusters. Inside each cluster a cluster
head (CH) is selected to perform management and routing tasks for that cluster.
Research has proved that hierarchical protocols perform much better than the flat
protocols in terms of energy efficiency and stability. The CH is selected either
randomly or based on some criteria [5]. After selection of a CH, other nodes join
that CH and act as member nodes. These member nodes send their data to CH,
which then aggregates the data and sends to the BS. Hence, the selection of a CH
largely affects whole network’s performance and stability. In most of the previous
clustering schemes, CH selection is based on single criterion. In single criterion,
CH is mostly selected randomly or based on residual energy, node density or
distance form BS. If CH is selected on the basis of residual energy, then problem
arises when a node with higher residual energy but located far away form BS is
selected as CH. That node consumes more energy to forward aggregated data to
BS. Similarly if CH selected no the bases of shortest distance form BS, then similar
type of problem arises if node near to BS is selected as CH, but with not sufficient
residual energy to communicate with BS. Hence single criterion is not sufficient
for CH selection. So the motivation behind this work is to design a clustering
technique which is based on multiple criteria rather than single criterion, also
which overcomes the bottlenecks in previous proposed clustering protocols.
1.3 Problem Statement
In most of the existing clustering techniques CH is selected randomly and based
on single criterion. They use centralized scheme, means BS performs CH selection
process. Also they use single hop communication model. In previous protocols CH
4
is changing in every round. The bottlenecks of previous protocols are discussed
below.
•Problems with single criterion:
–Mostly based on residual energy
–Dont consider other information, like location of nodes, number of
neighbor nodes etc
–So normal nodes consumes more energy to send their data to CHs.
•Problems with centralize algorithm:
–Increased processing over head. Results in shorter lifetime of nodes.
•Problems with single hop communication:
–Data from nodes away form CHs/BS have to travel longer distance as
compared to nearer nodes, hence they die earlier.
•Problems with frequent CH change:
–Frequent CH change increases processing over head. Results in shorter
lifetime of nodes.
To overcome these problems, we proposed a technique based on multiple crite-
ria, in which we use distributed algorithm of CH selection i-e nodes themselves
decide whether to become CH or not. Our proposed scheme is based of MCDM
fuzzy-TOPSIS method, we consider four criteria for CH selection which are resid-
ual energy, number of neighbors, average distance from neighbors and distance
between node and BS. In our proposed scheme we avoid quick deformation of CH,
which in result reduces control overhead. Because of reduction in control over-
head energy consumption is minimized. In our proposed scheme we use realistic
communication model by introducing multi-hop communication model in both
intra-cluster (communication between normal nodes and CH) and inter-cluster
(communication between CH and BS)
1.4 Research Methodology
In this work we comprehensively analyse the bottlenecks in existing clustering pro-
tocols, and design a new clustering scheme based on multiple criteria. Different
weights are assigned to each criteria according to their importance. Each node
5
share all four criteria value with all its neighbors, then rank index value is calcu-
lated for each node. The node with highest rank, elect itself as CH. The elected
CHs remain CHs until their rank index value differ from any other node value
by 0.1. (threshold value). In intra cluster, nodes with 5mrange of CH directly
communicate to CH, other nodes perform multi-hoping with neighbor nodes. Sim-
ilarly, in inter cluster CHs within 20mrange of BS directly communicate with BS,
remaining CHs perform multi-hoping with neighbor CHs.
1.5 Thesis Organization
Remaining thesis is organized as follows:
Chapter 2 provides the background of Wireless Sensor Networks (WSNs). This
include basic architecture of wireless sensor node, current challenges in WSNs, like
power consumption, network life time, scalability etc. and common applications
of WSNs, like military, environment and currently merging healthier applications.
Chapter 3 discusses WSNs network layer. This includes brief introduction, routing
challenges in WSNs, and Types of routing protocols. In this chapter we also
discuss concept of clustering in WSNs, and provide overview of existing clustering
protocols along with their working principle.
Chapter 4 describes our proposed clustering scheme. This include basic energy
model of wireless sensor nodes and different phases of our proposed scheme, along
with complete mathematical equations, explanations and flow charts.
Chapter 5 discusses simulation parameters and simulation results of our proposed
scheme. In this chapter we compare the results of proposed scheme with previous
clustering techniques, and analyze the improvements in our proposed scheme in
terms of network lifetime and stability.
Chapter 6 concludes the manuscript of our research work
6
Chapter 2
Wireless Sensor Networks
(WSNs)
7
Chapter 2
Wireless Sensor Networks
(WSNs)
Wireless sensorInetworks areIgaining lot ofIattention in researchIareas as wellIas in
theIdevelopment of variousIapplications. It has becomeIone of the leadingIand ef-
ficientItechnologies in wirelessIcommunication. TheseInetworks are usedIto mon-
itor differentIapplications suchIas snow monitoring, homeIand industry automa-
tion,Iand most importantlyIin military applications to monitorIthe information.
WSNIis a new network technologyIwhich integrates low-power communication,
sensor and Micro-Electro-Mechanics [MEMS] [6]. It is collection of many number
of sensor nodes which communicate themselves to acquire monitored information.
In WSNs sensor nodes can be deployed in either random (adhoc fashion) or in a
manual way depending upon the application. These networks which are grouped
with sensors are linked through a wireless medium to perform their required tasks.
Communication between these sensors is occurred with the help of infrared devices
or base stations or radios [7]. This radio network helps user to access informa-
tion from any remote location and allows to visualizing and analyzing the sensor
data. WSNs which consists of many number of sensor nodes are able to commu-
nicate within themselves and as well as with the BS. Each sensor device consists
of transceiver, a micro controller and is equipped with a power source which is
usually an AA and AAA batteries. The specifics of each WSN depend on the
nature of the application.
8
Figure 2.1: Wireless Sensor Networks
2.1 Architecture of WSN Node
Sensor nodes are small devices which are battery powered and are a part of this
typical wireless sensor network. A typical sensor node consists of four basic com-
ponents: sensing subsystem, processing subsystem, power subsystem and com-
munication subsystem [8]. The fig. 2.2 shown at the end of this section, is the
architecture of a single node. The explanation of each subsystem is given as fol-
lows:
2.1.1 Sensing Subsystem
Sensors play a crucial role in WSNs architecture as they establish a link between
the real-time world and the computational environment. Sensors are the hardware
devices which are used to monitor the data for required applications and to react
to the environmental changes. After sensing the environment, the function of the
sensor is to collect the sensed data and send it to further system for processing.
The energy in sensor nodes is transformed from one form to another form using
transducers. Sensor nodes normally include analog, digital and A/D converters
and a microcontroller [9]. Sensors are categorized depending on the application
and these can act according to the requirements of each application. Also, the
factors to choose in a sensor are size and battery consumption.
9
2.1.2 Processing Subsystem
Sensor nodes also consist of a processing unit along with memory units and con-
verters. The communication interface processes the data. Later the collected data
can be analyzed to verify the performance of the network. Here, the unit is re-
sponsible for adapting the routing information and align the topology if needed.
Also performs data gathering, data acquisition and however processes the received
data (incoming and outgoing) [10]. This subsystem also involves data fusion where
the different packets arriving from the sensor nodes are gathered to form a sin-
gle packet, thereby reducing the transmission energy between the sensor and user
(observer).
2.1.3 Communication Subsystem
This subsystem is responsible for the transmission of data. Sensor nodes use radio
frequencies to carry the signals from sensors through the BS to the required end
user. The role of the BS is to maintain the communication between sensor network
and external source (user). In a network, there can be a single or multiple base
stations depending upon the requirement, area and number of sensors to monitor.
In a network, each individual node communicates and co-ordinates with other
nodes. There are two types of communications: infrastructure and application
[11].Communication which is required to build, maintain, optimize a network is
referred as infrastructure. Due to environmental changes in the network there can
be a varying topology and sometimes nodes can fail. Therefore, these situations
can be managed by conventional protocols [11]. Hence even in a static sensor
network, there is a need of infrastructure communication and external commu-
nication which is required to re-configure the topology [11]. The data which is
gathered should be transferred further to the monitoring end and is referred as
application [11]. The amount of energy required to transmit a packet to sink is
depended on distance and more over the energy required for a node to transmit is
fixed. But, if the distance is far then that requires high amount of energy. Hence,
this can be eliminated by choosing the shortest path for transmission of data.
Also this communication refers to application based. For example, when there is
a necessity to communicate, nodes should communicate and data should be sent
continuously. Another example is when the application depends on event driven,
sensor suppose to act when the event or environment change occurs. Therefore, it
is good to decrease communication cost in order to increase life time of a network.
10
2.1.4 Power Subsystem
All the above mentioned subsystems require a power unit to function and perform
their individual tasks. Power subsystem provides the supply voltage and the
requirements of the power are strict due to energy consumption constraints. It also
supplies sufficient levels of current during the radio transmission and reception.
A battery can act as energy storage which is generally AA or AAA size and
voltage regulator is also included in the power subsystem. In most of the hardware
platforms there is a possibility to allow switching of the states i.e. between on, off
and idle for each device to minimize the power usage. These sensors collaborate
with each other at certain interval of time to carry out the required task. Data
from each sensor is collected and analyzed by a data processor (computer) outside
the network [12]. These nodes can be self organized and self healed depending
upon the routing topology that is used for the communication between them.
Also, it is very difficult to replace a sensor node if they are placed in extreme
geographical areas.
Figure 2.2: WSN Node Architecture
2.2 Challenges in WSNs
Before formation of the sensor network and deployment of sensor nodes the prior
and fundamental understanding about connecting and managing the network in
needed to achieve beneficial scalability and efficiency. Even though sensor net-
works are grouped under the class of ad-hoc networks but these differs with their
characteristics. Both ad-hoc and sensor networks share the challenges of energy
constraints and routing techniques [13]. Generally, in an ad-hoc networks nodes
are considered as mobile where as in the sensor networks nodes are static for most
11
of the applications such as military. Hence, these networks may differ in their traf-
fic patterns [13]. These are some of the most important aspects that the wireless
sensor networks should overcome, and they are described below.
2.2.1 Power Consumption/ Network Lifetime
Power consumption is one of the crucial challenges required to manage in sensor
networks. Many researchers are focusing their efforts to improve energy efficiency
in these networks [13]. As many of the sensors are battery powered, energy con-
sumption is a very crucial metric and should be managed wisely in order to extend
the network lifetime. For example in the military applications, it is difficult to
replace the batteries in the battle field. Hence the sensors may fail and might not
function if the batteries are exhausted. So, efficient routing may overcome this
issue and extend the network lifetime.
2.2.2 Fault Tolerance
While processing and communication between the sensors, some sensors may fail
to communicate because of link failures, lack of power supply or due to any phys-
ical damage or even by environmental interventions. In order to overcome these
mentioned problems, accommodation of new links is required. Also, maintaining
the transmission power and signaling rates; rerouting of packets and redundancy
is necessary to establish a robust and fault tolerant network.
2.2.3 Scalability
Scalability is a critical factor especially for sensor networks which contains many
number of nodes and can be responsible for degradation of network performance as
well. Topological changes in network such as network size and node density should
not affect the performance of the network. Hence, routing protocols employed in
WSN must be scalable enough to maintain the sensor states when it changes its
state from sleep to ideal or vice versa.
2.2.4 Throughput
Most of the times sensor must transmit its data to the BS, the required number
of successful packet transmission of a given node per time slot is determined as
12
throughput.
2.2.5 Accuracy/Latency
Acquiring the exact information without any distortions is the most primary objec-
tive in a WSNs. Also, there should not be any sort of delay. The routing protocols
and network topology will ensure the delivery of the data with minimum delay.
2.2.6 Node Deployment
The sensor nodes are placed manually in a random fashion and are deployed
depending upon the required application. Another way of deployment is self or-
ganizing systems, where the sensor nodes are scattered and topology is formed in
an ad-hoc manner. Uniform distribution of nodes and optical clustering schemes
can efficiently maintain the network [14].
2.2.7 Data Aggregation
Data aggregation is the combination of data arriving from different sources by
using some functions such as suppression (finding and eliminating duplicates),
minimum, maximum and average [14]. As sensor node generates the meaningful
data, data from multiple nodes can be aggregated in order to reduce the num-
ber of transmissions. This aggregation technique is used to reduce the energy
consumption and achieve data transfer optimization in the routing protocols.
2.2.8 Hardware Constraints
Since sensor nodes are very small in size and are operated under low power. These
have limited energy capacity, low storage and in addition to these, sensors have low
computational capability. Therefore, there is a need of adequate network design
for routing protocols that can overcome mentioned challenges.
2.2.9 Security Issues
As the routing protocols have limited capability, some of these protocols cannot
accommodate all the crucial information acquired by the sensor, challenging the
security of data. Data is sent to the end users by getting direct access to the
13
messages present in the sensors through internet services. Hence, there is a need
to prevent the data from unauthorized parties or from any malicious actions.
2.3 Applications of WSNs
WSNs have a capability to monitor wide range of applications including physi-
cal conditions [15] such as temperature, humidity, light, pressure, noise intensity
level, object movement and its characteristics etc. WSN node promises many new
applications by implementing concept of micro-sensing and wireless communica-
tion. There are many application related to WSNs and some of these are explored
below;
2.3.1 Military Applications
Wireless sensor network helps in surveillance and tracking of information in mil-
itary command control. The Ad-Hoc deployment of the sensor nodes, self orga-
nization and fault tolerance characteristics of WSNs, improves the firm sensing
capability of this application. Some of the other military applications are mon-
itoring the friendly forces, ammunition and equipment, attack detection, battle
surveillance and targeting etc. [16]
2.3.2 Environmental Monitoring
Applications like snow monitoring which is used to monitor the snow conditions
and avalanche forecasting; habitat monitoring which helps to deliver the informa-
tion about localized environmental conditions of each individual habitat, such as
issues affecting animals, plants and humans [17]; humidity and temperature mon-
itoring, wild life monitoring, traffic control, fire detection, flood detection etc also
utilizes WSNs. Also another important example that comes under environmental
monitoring is disaster management. Sensor networks help in detection of location
that could be useful for rescue operations, also used for prevention of potential
hazards. Figure 2.3 shows environmental applications of WSNs.
2.3.3 Medical Applications
Sensor networks have also focused its attention on medical application. These are
used to monitor the patient’s physiological condition, also used to administrate the
14
Figure 2.3: Environmental Applications of WSNs
drug section, monitor the patients and the doctors within the hospital [18]. These
are also used to detect the different types of viruses by monitoring the infected
area. Figure 2.4 shows medical application of WSNs. This is called Wireless Body
Area Networks (WBANs).
Figure 2.4: Medical Application of WSNs
2.3.4 Other Applications
For commercial purposes, sensors are widely used in home and industry automa-
tons. Also, the commercial buildings and offices are equipped with sensors and
15
actuators to monitor the room temperatures and air flow thereby improving the
living conditions. In home automation these applications are used for remote me-
tering and for smart intelligence purposes. Vehicle tracking and detecting is also
an application of WSNs that can help avoid car thefts.
2.4 Summary
In this chapter, a brief overview of WSNs is presented. Starting with introduc-
tion, then basic basic architecture of wireless sensor node, current challenges in
WSNs, like power consumption, network life time, scalability etc. and common
applications of WSNs, like military, environment and currently merging health-
ier applications. Next chapter gives overview of WSN network layer, concept of
clustering, and some well known clustering protocols.
16
Chapter 3
WSN Network Layer
17
Chapter 3
WSN Network Layer
3.1 Introduction
Data collected by sensor nodes in a WSNs is typically propagated toward a BS or
gateway that links the WSNs with other networks where the data can be visualized,
analyzed, and acted upon. In small sensor networks where sensor nodes and
a gateway are inclose proximity, direct (single-hop) communication between all
sensor nodes and the gateway may be feasible. However, most WSNs applications
require large numbers of sensor nodes that cover large areas, necessitating an
indirect (multi-hop) communication approach. That is, sensor nodes must not
only generate and disseminate their own information, but also serve as relays or
forwarding nodes for other sensor nodes. Difference between single-hop and multi-
hop is shown if fig. 3.1. The process of establishing paths from a source to a sink
(e.g., a gateway device) across one or more relays is called routing and is a key
responsibility of the network layer of the communication protocol stack [19].
Figure 3.1: Single-hop routing versus multi-hop routing model
18
The key responsibility of the network layer is to nd paths from data sources to
sink devices (e.g., gateways). In the single-hop routing model, all sensor nodes
are able to communicate directly with the sink device. This direct communication
model is the simplest approach, where all data travels a single hop to reach the
destination. However, in practical settings, this single-hop approach is unrealistic
and a multi-hop communication model must be used. In this case, the critical
task of the network layer of all sensor nodes is to identify a path from the sensor
to the sink across multiple other sensor nodes acting as relays. This design of a
routing protocol is challenging due to the unique characteristics o WSNs, includ-
ing resource scarcity or the unreliability of the wireless medium. For example,
the limited processing, storage, bandwidth, and energy capacities require routing
solutions that are lightweight, while the frequent dynamic changes in a WSN (e.g.,
topology changes due to node failures) require routing solutions that are adaptive
and exible. Further, unlike traditional routing protocols for wired networks, proto-
cols for sensor networks may not be able to rely on global addressing schemes(e.g.,
IP addresses on the Internet) [19].
ToIminimize energyIconsumption, routingItechniques proposed in theIliterature
for WSNs employIsome well-known routingItactics as well asItactics special to
WSNs,Isuch as data aggregationIand in-networkIprocessing, clustering,Idifferent
node roleIassignment, andIdata-centric methods.
3.2 Routing Challenges and Design Issues in WSNs
The challenges and characteristics of WSNs are different than the conventional
Wireless Ad-hoc Networks (WANETs) due to their specific requirements. There-
fore, the designing task of a routing protocol in WSNs requires more careful con-
siderations than the other wireless ad-hoc networks (MANETs or WMNs). As
outlined below, the issues to consider for an efficient and reliable communication
in WSNs include network topology, data reporting methods, node and link het-
erogeneity, mobile adaptability, energy efficiency, coverage, data aggregation, and
quality of service [20]. These challanges are discussed below:
•Mobile adaptability: Most of the WSNs use the fixed nodes and base
stations. However, sensing the node or sink mobility can be a demand of
an application in different scenarios like vital sign monitoring of a mobile
patient in the hospital.
•Energy efficiency: Routing protocols need to maintain the connectivity
between the nodes and the base station with minimum energy consumption.
19
The periodicrouting updates help the nodes to refresh the status of neighbor
nodes. The flooding of these updates can shorten the nodes lifetime due to
the additional energy required.
•Coverage: Each sensor node can sense the environment within a certain
range of area coverage. The design of WSNs requires the deployment of the
nodes in a way that can get the maximum coverage. The routing protocol
needs to choose another node from the same sensor area where a node fails
to ensure the proper coverage of the whole sensor area.
•Data aggregation: Several methods, such as duplicate suppression, me-
dian, and minima-maxima, are used in routing protocols for data aggre-
gation to avoid the redundant transmissions and enhance the energy effi-
ciency. These techniques also help to reduce the traffic load and increase the
throughput.
•Quality of Service: Reliability-sensitive and delay-control algorithms are
used for routing protocols to fulfill the QoS demand of different WSN appli-
cations. These protocols help to monitor the sensor areas during a critical
situation.
3.3 RoutingIProtocols in WSNs
Several WSN routing protocols have been proposed by researchers in the last few
years. The WSN routing protocols can be classified in three ways: by its protocol
operations, by network structure, and by packet destinations [20]. Figure 3.2 shows
the classification of WSN routing protocols. The details of this classification are
given below.
3.3.1 Protocols based on Network Organization
The underlying network architecture plays an important role in the operations of
the routing protocol in WSNs. Routing Protocols based on network organization
can be divided into three categories. In this section the review of routing protocols
with respect to network structure is provided.
•Flat-Based:
The multi-hop routing approach is used by flat routing protocols. AllIthe nodes
inIthe network play the sameIrole. The BS generates the queries to the nodes and
20
Routing Protocol
Network Organization Protocol Operation Route Discovery
Flat- Based
Hierarchical-
Based
Location- Based
Negotiation-
Base
Multi-Path-
Based
Query-
Based
QoS-
Based
Coherent-
Based
Reactive
Proactive
Hybrid
Figure 3.2: ClassificationIof WSN RoutingIProtocols
in response nodes transmit data towards a base station. Scalability and simplicity
are the two major advantages of this kind of routingIprotocols. Because allIthe
nodes playIthe sameIrole, these protocols can easily accommodate a large number
of nodes or can add more nodes. Simplicity emerges from not choosing any cluster
head. All the nodes only send the data to the next hop or base station. The
complexity involved for electing the cluster head is not required. The disadvantage
of these routing protocols is the hot-spots. Every node is capable of sending its
own sensed data and of forwarding the other nodes dataIto the BS.IThe energy in
nodes around the sink drains quickly due to forwarding a lot of other nodes data
to the BS.
•Location-Based:
The location informationIof the sensor nodes in WSNs is used to calculateIthe
distanceIbetween the nodes, which helps to choose the next hop. The deploy-
ment of nodes in WSNs is spatial in nature. The sensor nodes are addressed by
using their location information since there is no addressing scheme for WSNs
such as IP-addressing. The nodes locationIinformation assists the location-based
routing protocols to send the data in a desired region instead of the whole net-
work. Some of the examples of location-based routing protocols are Minimum
21
Energy CommunicationINetwork (MECN), Trajectory-BasedIForwarding (TBF),
GeographicIAdaptive FidelityI(GAF), and GeographicIand EnergyIAware Rout-
ingI(GEAR).
•Hierarchal-Based:
InIhierarchical routing,Ithe networkIis dividedIintoIclusters to achieve efficiency.
The selection of cluster heads and formation of clusters are the two important
concerns of hierarchical routing protocols. The advantage of hierarchical routing
is the aggregation of data. The data from all member nodes are sent to the cluster
head and then the cluster head forwards these data towards the sink after applying
compression techniques. The aggregated data are easy to handle and simple to
process. One of the major drawbacks of hierarchical routing is the increase in
energy consumption of cluster heads due to their additional functions. The nodes
are selected as cluster heads in rotation manner, which overcomes this issue. The
selection of cluster heads and the formation of clusters in each round require more
computations, which also causes more energy consumption.
Table 3.1: Difference between flat and hierarchical routing
Flat Hierarchical
Contention-basedIscheduling Reservation-basedIscheduling
CollisionIoverhead present CollisionsIavoided
VariableIduty cycle byIcontrolling
sleep timeIof nodes
ReducedIduty cycleIdue to peri-
odicIsleeping
NodeIon multihopIpath aggregates
incomingIdata fromIneighbors
DataIaggregationIby clusterIhead
RoutingIcan be madeIoptimal butI
with an added complexity.I
SimpleIbut non-optimalIrouting
LinksIformed onIthe flyIwithout
synchronizationI
Requires globalIand localIsynchro-
nization
RoutesIformed onlyIin regionsIthat
haveIdata forItransmission
OverheadIof clusterIformation
throughoutIthe network
LatencyIin wakingIup intermediateI
nodes and settingIup the multi-path
Lower latencyIas multipleIhops net-
work formedIby cluster heads al-
waysIavailable
Energy dissipationIdepends onItraf-
fic patterns
EnergyIdissipation isIuniform
EnergyIdissipation adaptsIto traffic
pattern
EnergyIdissipation cannotIbe con-
trolled
FairnessInot guaranteed FairIchannel allocationI
22
3.3.2 ProtocolsIbased onIProtocol Operation
In this classification, different routing functionalities of the routing protocols are
considered. The protocol operation based routing protocols are divided into dif-
ferent types given below [20]:
•Negotiation-Based:
InIthese protocols, a high level of descriptors is used for the negotiation between
the nodes to prevent redundant data and reduce duplicate information. Generally
this negotiation is done between the source and the next node or base station
before real data transmission. SPIN [21] is an exampleIof thisItype of routing.
•Multipath-Based:
The routing protocols in this category use the multiple pathsIbetween a sourceI
and destination for data transmission to enhance the network performance. The
directedIdiffusionIprotocol [22] is an example of multipath-based routing.
•Query-BasedI:
These protocols depend upon the queries from a destination. The source node
sends its sensed data in response to a query generated by the destination node.
A natural language or high level query language is used to generate these queries.
An example of these protocols is the rumor routing protocol [23]. The directed
diffusion protocol [22] is also counted in query-based routing protocols.
•QoS-Based:
The algorithm used by these protocols ensures the QoS requirements of the data.
Some of the QoS metrics are reliability, delay, and bandwidth. Balancing energy
consumption while satisfying QoS conditions is an important task for these routing
protocols. SPEED protocol is a good example of QoS-based routing.
•Coherent-Based:
Different data processing techniques are used to reduce the processing computa-
tions, which help to reduceIthe energy consumptionIof the node. Coherent and
non-coherent are the two major data processing methods used for this purpose.
In the non-coherent data processing technique, the sensed raw data are processed
locally by the node and then the node transfers it to the aggregator. Aggregator
is a node which receives the data from many sensor nodes and sends these data
to the sink or base station after aggregation. In the coherent method, the mini-
mum processing is done locally by the sensor node. After receiving the data, the
aggregator is responsible for the major and complex part of processing. Exam-
23
ples of coherent and non-coherent data processing techniques are Multiple Winner
Algorithm (MWE) [24] and Single Winner Algorithm (SWE) [24], respectively.
3.3.3 Protocols based on Route Discovery
These routing protocols are responsible for identifying or discovering routes from a
source or sender to the intended receiver. This route discovery process can also be
used to distinguish between different types of routing protocols. There are three
protocols based on this types.
•Reactive:
Reactive protocols discover routes on-demand, that is, whenever a sourceIwants
toIsend data to a receiver and does not already have a route established.
•Proactive:
While reactive route discovery incurs delays before actual data transmission can
occur, proactive routing protocols establish routes before they are actually needed.
This category of protocols is also often described as table-driven, because local
forwarding decisions are based on the contents of a routing table that contains
a list of destinations, combined with one or more next-hop neighbors that lead
toward these destinations and costs associated with each next hop option. While
table-driven protocols eliminate the route discovery delays, they may be overly
aggressive in that routes are established that may never be needed. Further, the
time interval between route discovery and actual use of the route can be very
large, potentially leading to out-dated routes (e.g., a link along the route may
have broken in the meantime).
•Hybrid:
The cost of establishing a routing table can be signicant, for example, in some
protocols it involves propagating a nodes local information (such as its list of
neighbors) to all other nodes in the network. Some protocols exhibit characteris-
tics of both reactive and proactive protocols and belong to the category of hybrid
routing protocols.
24
3.4 Clustering in WSNs
Clustering is a technique used in hierarchical based routing. It is a technique to
divide whole network intoIsmall blocks, calledIclusters. With each cluster hav-
ing a managing node, called cluster head (CH) and rest act as members. CH
is responsible to provide communication bridge between members and the base
station. The selectionIof cluster headsIand formation of clusters are the two im-
portant concerns of clustering protocols. The advantage of clustered routing is
the aggregation of data. The data from all member nodes are sent to the cluster
head and then the cluster head forwards these data towards the sink after apply-
ing compression techniques. The aggregated data are easy to handle and simple
to process. One of the major drawbacks of hierarchical routing is the increase in
energyIconsumption of cluster heads due to their additional functions. The nodesI
are selectedIas cluster heads in rotation manner, which overcomes this issue. The
selection of cluster heads and the formation of clusters in each round require more
computations, which also causes more energy consumption. There are many clus-
tering protocols developed by researchers in past few years. Some well known
clustering protocols are discussed in next section.
3.4.1 Energy Saving Schemes in WSN Clustering
As discussed above, CH selection and cluster formation are important constrains
in WSNs clustering. Hence a large amount of energy can be preserved by efficiently
electing cluster head and forming clusters. This is discussed in detail in following
sections.
3.4.1.1 Cluster Formation and Rotation
With the evolving trend in application and management of WSNs, clustering pro-
vides an efficient means of managing sensor nodes in order to prolong its lifetime.
Several clustering formation technique have been develop in the past such as ran-
dom competition based clustering (RCC)[25]. RCC algorithm uses random timer
and node identification for cluster formation is based on First Declaration Wins
Rule. This rule assigns governorship position to any node which declares itself
first as being a CH to other nodes in its radio network.
In direct broadcasting technique, cluster advertisement message is sent to all sen-
sors within a selected region. For instance, two clusters formation requires two
random nodes are selected for broadcasting. This randomly selected node is known
25
as an initiator. All initiators broadcast a cluster advertisement message to all sen-
sor nodes in the network. If any node in the network that is not an initiator
receives an advertisement message within the cluster, it sends a message to the
initiator from which the message was received. It will not only send a reply but
also refrain from accepting any other cluster advertisement message for that sim-
ulation round. Such a sensor node will however become a part of this initiators
cluster [26].
The technique of direct broadcasting is very simple when it comes to its implemen-
tation but it is not cost effective in termsIof energy consumption.IThis is due to
the fact that all sensor nodes receive a broadcast from the CH. The sensor nodes
that are very far to the CH will still need to receive broadcast but it does not mean
that the sensor will respond to the message. But, in a situation where a sensor
node receives a broadcast from an initiator, the subsequent broadcast message will
be dropped, and energy which is used in transmission will be underutilized.
Multi-hop broadcasting on the other hand uses specific transmission range to
transmit a cluster advertisement message to the sensor nodes. It is the duty of
the receiving node to proceed in sending the cluster advertised message to all
the sensor nodes in its transmission range. This technique works very closely to
direct broadcasting technique for the fact that it also selects an initiator node
that sends cluster advertisement messages at the start of cluster formation. These
techniques use a concept which is known as minimum communication energy which
means that the sensor node that is easiest to reach will form part of the initiator
cluster. Also, when cluster are formed dynamically, the reorganization is done on
a periodic basis. The initiator is selected at the beginning of every period and
broadcasted messages are sent out using one of the above-mentioned methods for
cluster organization.
The multi-hop broadcasting minimizes the problem of energy usage. This is due
to the fact that there is a limit for transmission because the highest amount of
energy that can be wasted is the minimum transmission energy of neighboring
sensor nodes. This will create no need for the sensor nodes which are far away
from each other to transmit directly. It has a disadvantage in the sense that it
has more delay when compared to the former technique of broadcasting. This
is because, in multi-hop broadcasting, the data are required to be processed by
each sensor node along the multi-hop path, which creates delay in the formation
of cluster. However, the multi-hop is much better than the direct broadcast if the
problem of delay is taken care of.
26
3.4.1.2 Cluster Head Election and Rotation
After cluster formation, CHs are designated which act as a leader in each clus-
ters. ClusterIheads areIprovided with the responsibility for data aggregation and
performing routing for its cluster members information to the base station. Also,
the clusters that consist of many nodes have a higher burden than clusters with
fewer nodes as the CHs for those large-sized clusters have to receive, aggregate
and transmit more data.
A CH can be elected randomly or pre-assigned by the designer of the network. A
CH can also be elected byItaking into considerationIthe residual energy of nodes
in the cluster. The CHs are known to have higherIburdens than member nodes;
therefore, the role of CH is rotated to share the burden and thus improving the
useful lifetime of those clusters.
In random selection, a node is selected randomly as CH, based on the probability
that it has never being selected during the entire lifetime of the network. This
reduces the traffic burden on a CH since the role of CH is spread throughout the
sensor nodes. The rotation is done at a periodic interval.
Whereas, in the residual energy selection, the sensor node that hasIthe highest
amountIof energy in the cluster is selected as the cluster head. It will continue
to remain the CH until the energy drops below the average energy of the entire
cluster. So, rotation of CH is done at every instance when its energy level drops
below the average cluster energy. This rotation of CHs will lead to the overall
energy of the sensor network being evenly distributed. This technique eventually
improves the lifetime of the network.
Another approach to cluster headIselection isIbased on minimizing the distance
to cluster nodes as this offers reduction in energy usage during data transmission
to the BS. With this method of minimizing sum of distances to CH, the cluster
formation is better enhance to reduce energy usage as transmission takes place. It
helps in reducing the unnecessary energy which the sensor node uses in commu-
nicating with the CH by minimizing the transmission distances from sensor node
to any CH [27].
Since communication energy is an important concept to consider in wireless trans-
mission, it is known that energy greatly depends on distance [7]. Therefore, it is
very good idea to minimize the distance in transmitting data from sensor nodes
to base station via the CH, as this helps to reduce the communication energy in
wireless sensor network.
27
3.4.1.3 Efficient Intra-cluster Communication
When addressing the problem of energyIconsumption inIwireless sensorInetwork,
the size of a cluster is an important factor subject to analysis in hierarchical
network. Clusters of small size save power in intra-cluster communication but it
will also increase the complexity of the backbone network. Also, smaller cluster
size means less load in the backbone and thus a less complicated communication,
but the intra-cluster communication consumes more power and reduces the lifetime
of the sensor network. These necessitate a tradeoff in clusters formation [27].
One of the trade-off is mentioned in where a method of limiting the number of
hops which a sensor node takes in communicating with its CH. Bandyopadhyays
and Coyles approach makes a use of the K-tree cluster framework and optimize
the framework and optimize the value K to minimize power consumption within
a cluster. The K-tree clustering algorithm can be described as a technique that
ensures that every sensors node within a cluster can carry out communication
with its corresponding cluster head by using a maximum of K- hops. Therefore, a
sensor node needs to use other sensor nodes to relay its transmission to the base
station via its routing table. And, the data of any sensor node will not be relayed
within a cluster for more than K times.
Moreover, in, Bandyopadhyay and Coyle utilize a method of selecting cluster head
with the probability P. The message will then be forwarded to all nodes which are
K-hops away. The optimal K value is a predetermined number that is set according
to the size of the network. Any sensors that receive the advertisement message
from the elected CH are considered a cluster member of the cluster from which it
received this message. The message sent by the sensor node would be ignored if
it is received by another volunteer cluster head. Also, a node can be mandated
to become a cluster head if it does not receive a CHs advertisement in a specific
time t which is defined as the time required to send K hops data away [27].
With the K-hops approach, communication in wireless sensor network can be
either of the single-hop or multi-hop. Single-hop simply refers to direct commu-
nication from sensor node to cluster head while multi-hop does not require direct
communication from all sensors to the base station. But, it can send data to the
next cluster head which is closer to the base station. Therefore, multi-hop com-
munication has higher energy efficiency than the single hop within the clusters.
When the sensor node is at a far distance from the CH, much energy is expended
thereby reducing the lifetime of sensor network.
For instance, when two sensor nodes are placed at a far distance away from each
28
other in the same cluster but one of them is closer to the base station, it is observed
that the energy consumed by the node closer to the base station is lower compared
to that which is far away. And with the help of multi-hop communication, an
intermediary node is used between a source environment and BS to relay the
data, a great deal of energy is conserved in the network.
Furthermore, taking computational complexity into consideration when designing
cluster with uneven data traffic in each clusters when data are being transmitted,
the size of clusters seem to be random at the time of formation. Although, in some
circumstances, the cluster sizes are equal which denotes equal number of nodes,
and in other scenario, the size is randomly sized.
3.5 Clustering protocols in WSNs
Proper clustering and CH selection largely affects network lifetime and stability
in WSNs. Lots of research have been devoted in this regard and many clustering
protocols have been purposed. This section reviews few of the previously proposed
clustering schemes.
3.5.1 Low-EnergyIAdaptiveIClusteringIHierarchy (LEACH)
The LEACH [28] protocol is the basic clustering-basedIenergy-efficient routing
protocol. The clustering techniques proved to be very useful to reduceItheIenergy
consumptionIand increase the networkIlifetime. The entire network is divided into
clusters in the LEACH routing protocol. One sensor node in eachIcluster must act
as a cluster head and all remaining sensor nodes are member nodes of that cluster.
Communication between the member nodes and sink is only possible via the cluster
head. From each cluster, only the cluster headIcan directlyIcommunicateIwith the
sink. The clusterIheads collect, aggregate,IandIforward theIdata from member
nodes to the sink. The cluster head consumes more energy due to the additional
functions and this node can die quickly if it continuously plays the role of a cluster
head. LEACH resolved this problem by changing dynamically the role of nodes
as cluster heads.
LEACH works in rounds. The operations that are carried out inIeach round
consist of two phases known as setup and steady state phases. The organization
ofIclusters andIselectionIof clusterIheadsI(CHs) are done in the setup phase of the
LEACH. The data are sent to the sink during the second or steady state phase.
29
InItheIsetupIphase, the formation of clustersIand the election process of cluster
heads are performed. First of all, the whole network is divided into clusters. Now
the cluster head election process starts in each cluster. There are many ways to
elect the CH. Some of the wireless nodes in the network ignore the negotiation
process with other nodes and elect themselves autonomously as CHs. The CH
selection criteria of a member node are the recommended percentageIPand the
earlier recordIas a CH. If a node is not a CH in preceding 1/P rounds, it produces
a numberIbetween zero and one (0-1). Only nodes with a generated number less
than threshold T(n) are eligible to become CHs. The formula used to calculate
the value of threshold is given in eq. (3.1).
T(n) =
p
1−p(r∗mod(1/p)) if n∈G
0 otherwise
(3.1)
where
G = Group ofInodesInot selectedIas CHs in preceding 1/pIrounds;
P = Recommended percentageIofICH;
r = Current round.
Figure 3.3: Architecture of LEACH
The basic architecture of LEACH isIshown in fig. 3.3.IA node cannot be selected as
a CH if it has already performed a CH role in the lastI1/p rounds but all nodes that
were CHs before 1/p rounds will again be candidates for the selection of CHs [28].
30
The uniform service of each node as a CH prevents the uneven energy consumption
of the member nodes. The CSMA/CA protocol is used by the CH to broadcast
its status. After receiving the broadcast messages from CHs, the non-cluster head
nodes useIthe ReceivedISignal Strength Indication (RSSI) as a parameter to select
their CH. Each CH creates a TimeIDivisionIMultiple AccessI(TDMA) schedule for
their cluster members. The CH and member nodes communicate with each other
during their assigned time slots in the steady state phase. The member node is
only in active mode during its communication with a CH. Otherwise the mem-
ber node goes to sleep mode during an unallocated time slot. The management
of the member node in this way reduces theIenergy consumptionIandIincreases
theIbattery life of the node. The CH collects the data from all the cluster member
nodes. The CH transmits that dataItoIthe baseIstation after compression. The
time duration of setup phase is lower than the steady state phase.
3.5.2 Centralized LEACH (C-LEACH)
C-LEACH [29] is proposed by Heinzelman et al. The conventional LEACH proto-
col does not guarantee the best possible number of CHs and their effectual loca-
tions. The problem is due to the clusters formation method used by the LEACH
algorithm. LEACH-C is therefore proposed to enhance the cluster creation part
of the LEACH protocol.
All nodes are required to send their ID, location, and energy information to the
base station during the setup phase of C-LEACH [29]. The base station is re-
sponsible for assigning the role of CH to any member node by using its central
control algorithm. The central control algorithm first specifies the average energy
level and then compares that energy level to the energy level of the received signal
energy. The base station picks the optimal number of CHs from the nodes with an
energy level greater than the average energy level. A list of IDs of these selected
nodes is transmitted by the base station to all nodes. From this list, a node having
minimum distance from its member nodes is elected as CH of that cluster. The
approach used in C-LEACH reduces the energy consumption of CH and member
nodes. The following assumptions are taken in this protocol:
•Every node can calculate its energy.
•Every node knows its location.
•Every node can communicate to the base station.
The successful data transmission during the steady state phase of LEACH-C in-
31
creases. Some of the disadvantages of using the LEACH-C protocols are given
below.
•During the setup phase, all the nodes in the network are required to send
their information to the base station. This process causes additional energy
consumption from each node.
•The central control algorithm runs on the base station to select the CHs. The
IDs of selected CHs are passed to all nodes. Every node needs to compare
its ID with the IDs of CHs to determine its role as a CH. These additional
computations consume more energy from the nodes.
At the end of this chapter we present the detailed comparison of some well known
routing protocols in table 3.2 on next page.
32
Table 3.2: Detailed Comparison of WSN Clustering Protocols [30]
Protocol Energy
Homo-
geneity
Scalability Energy
Effi-
ciency
Network
Lifetime
Data
Delivery
Model
CH Selection based on
Initial
Energy
Residual
Energy
Average
Energy of
Network
LEACH Yes Very
limited
scalability
Lower Poor Cluster
head
YesINoINoI
TEEN Yes Very
limited
scalability
Very high Best Active
Thresh-
old
YesINoINoI
SEP No (Two
levels
heteroge-
neous)
MoreIscalableLowIGood Cluster
Ihead
No Yes No
DEEC No
(Multi
levels
heteroge-
neous)
Most Scal-
able
High Better Cluster
head
No Yes Yes
33
Chapter 4
Proposed Clustering Scheme
34
Chapter 4
Proposed Clustering Scheme
Various techniques can be found in clustering protocols. The basic idea of these
techniquesIisIto efficiently utilize the energy in orderIto prolongInetwork lifetime
and stability. In this chapter we discuss the details of my proposed clustering
scheme, which is based on MultipleICriteria DecisionIMaking (MCDM).
4.1 Energy Model for WirelessISensor Node
In WSNs most of the energy is consumed in communication process. Therefore
for estimation of WSNs energy utilization, it is necessary to calculate energy cost
for transmission and reception. This cost can be evaluated using simple model for
radio hardware consumption. Figure 4.1 shows energy model for wireless sensor
node.
Transmitter Electronics Transmitter Amplifier
Receiver Electronics
K-bit packet
K-bit packet
Eelect *k Eamp*d2*k
Eelect *k
Tx Antenna
Rx Antenna
d
Figure 4.1: Wireless Sensor Node Energy Model
35
DependingIonIthe distance dbetweenItransmitterIand receiver, the required trans-
mitting and receiving energy for a k-bitIpacket can beIexpressed as following equa-
tions. Eq. (4.1) for transmitting energy, where as eq. (4.2) for receiving. Both
free space and multi-path fading channel models are used in this energy model.
ET x(k, d) = kEelect +Eamp (k, d)
=
kEelect +Kef sd2, d ≤d0
kEelect +kEmpd4, d > d0
(4.1)
ERx =kEelect (4.2)
where:
ET x ⇒energyIdissipated per bitIatItransmitter
ERx ⇒energyIdissipatedIper bit atIreceiver
Eamp ⇒amplification factor
Eelect ⇒costIofIcircuitenergyIwhen transmittingIor receivingIone bit of data
efs ⇒free space coefficient
emp ⇒multi path coefficient
k⇒number ofItransmittedIdata bits
d⇒distanceIbetween a sensor node and its respective cluster head or between a
CH to another CH nearer to the BS or between CH and BS
d0⇒distance threshold value obtained by d0=qef s
emp
For scalability purpose, we assume that the intra-cluster transmission range must
satisfy d > d0and inter-cluster transmission range must satisfy the bound d≤d0.
An error free communication and an ideal MAC layer are also assumed so that
transmission is perfect and there is no collision and retransmission.
4.2 Proposed Scheme
We propose a multi-criteria based distributed CH selection technique based on
fuzzy-TOPSIS method. We improve deficiencies in previous fuzzy based CH se-
lection technique. Due to using distributed algorithm, nodes themselves take
decision to be selected as CH, hence nodes join CH with maximum resources be-
cause all nodes have index value of their neighbouring nodes (which is a rank value
obtained using multi-criteria, final CH selection is based on this value).
36
We define a threshold value for change of CH, so in our proposed scheme CHs
are not changing in every round, due to this, control overhead is much reduce as
compared to previous scheme. We consider four criteria includingIresidual energy,
numberIof neighbor nodes, distance fromIBS and average distance form neighbor
nodes. Our proposed scheme consists of four phases, i.e. network deployment,
neighbor discovery, CH selection and cluster formation and last one is communi-
cation. These are three phases are describedIin detailIin this section.
4.2.1 Network Deployment Phase
The basic architecture of Wireless Sensor Networks used in our protocol is shown in
Fig. 4.2. We assume that the sensor nodes are deployed randomly and uniformly,
are not movable, and are homogenous initially with respect to their antenna gain.
It is also assumed that the dimensions of the sensor field are given and the coor-
dinates of the base station are known. The base station is capable of receiving,
aggregating, and then forwarding the data from the cluster heads to the desired
destinations.
Figure 4.2: Sensor Nodes Deployed in Field
37
4.2.2 Neighbor Discovery Phase
The initial step of our proposed distributed clustering scheme is to perform neigh-
bor discovery. Initially, all nodes broadcast a Hello packet, which contains node’s
ID, residual energy, C1, node density C2, distance to the BS, C3, average distance
between this node and its neighbors, C4and location information. Initially, C2and
C4fields in the Hello packet will be empty because each node has no information
about its neighbors. However, after sharing Node ID and location information
with its neighbors, each node can easily compute C2and C4and exchange it in
the next Hello packet. All the other nodes in the transmission range Trof that
node, receive Hello packet. After receiving hello packet from all neighbors, a node
updates its neighborhood table (T) with neighboring node’s ID, C1,C2,C3,C4
as well as its own information. Suppose, there are nneighbors of node k, then Tk
will be an (n+ 1) ×4 matrix, as shown below:
Tk=
C1C2C3C4
a1v1,1v1,2v1,3v1,4
a2v2,1v2,2v2,3v2,4
a3v3,1v3,2v3,3v3,4
: ::::
an+1 vn+1,1vn+1,2vn+1,3vn+1,4
(4.3)
After updating packet form all neighbors, the nodes perform multi-criteria tech-
nique to calculate their respective rank index, and share it with all neighbor nodes
through hello packet. Following are steps to calculate rank index value based on
fuzzy-TOPSIS.
4.2.3 CH selection and Cluster Formation Phase
Comprehensive explanation of CH selection process of our proposed scheme is
given in this section. Following are steps to calculate rank index value based on
fuzzy-TOPSIS:
Step 1: It is evident that the values of all criteria, Ci, do not lie in the same
range, e.g. range of C1is not similar to the C2. Therefore, these criteria must be
normalized to the similar range [0 −1] to fairly select a CH. Note that there are
some criteria whose larger value is suitable for a node to be selected as a CH e.g.
C1and C2. These criteria are called Positive criteria, Benefit criteria or Positive
Ideal Solution (PIS) and are normalized as in eq. (4.4). On the other hand, the
38
criteria with smaller values are appropriate for a node to be selected as a CH,
e.g. C3and C4. This type of criteria are called Negative criteria, Cost criteria or
Negative Ideal Solution (NIS) and are normalized as in eq. (4.5).
Ni,j =vi,j −min∀i(vi,j )
[max∀i(vi,j)−min∀i(vi,j)] (4.4)
Ni,j =max∀i(vj)−vi,j
[max∀i(vi,j)−min∀i(vi,j)] (4.5)
Each element of the Tk, is normalized using eq.(4.4) and (4.5) and the normalized
matrix at node k,Nk, will be:
Nk=
N1,1N1,2N1,3N1,4
N2,1N2,2N2,3N2,4
N3,1N3,2N3,3N3,4
: : : :
Nn+1,1Nn+1,2Nn+1,3Nn+1,4
(4.6)
The preference or weight ’wi’ are assigned to each criterion. These, weights are
application specific, however, for our proposed scheme, the weights assigned to the
selected criteria are shown in Table 4.1.
Table 4.1: Criteria Weights
Criteria Weight
Residual Energy (w1) 0.4
Node density (w2) 0.2
Distance form BS (w3) 0.2
Avg. Distance between Neighbors (w4) 0.2
After normalization, Fuzzy membership function is used to categorize these nor-
malized value of each criteria and their respective weights for every node. For our
proposed scheme Fuzzy member function is given in Table 4.2.
Table 4.2: Fuzzy Membership Functions
Very LowI(VL) (0.00, 0.08, 0.15, 0.25)
LowI(L) (0.15, 0.28, 0.35, 0.45)
MediumI(M) (0.35, 0.48, 0.55, 0.65)
HighI(H) (0.55, 0.68, 0.75, 0.85)
Very HighI(VH) (0.75, 0.88, 0.95, 1.00)
39
Step 2 : On the basis of these Fuzzy membership functions, weighted decision ma-
trix is determined for each criteria and their respective weight. Weighted decision
matric Vkis given as:
Vk=
V1,1V,12 V1,3V1,4
V2,1V2,2V2,3V2,4
V3,1V3,2V3,3V3,4
: : ... :
Vn+1,1Vn+1,2Vn+1,3Vn+1,4
(4.7)
Step 3 : After that, PIS and NIS are determined from Vkmatrix, using following
equations:
P I S = (V+
1, ..., V +
n) = [(maxiVij |i= 1, ..m), j = 1, ...., n] (4.8)
NIS = (V−
1, ..., V −
n) = [(miniVij|i= 1, ..m), j = 1, ...., n] (4.9)
Step 4 : Separation measure is also determined form Vkmatrix, using the n-
dimensional Euclidean distance. This distance can be calculated as:
D+
j=v
u
u
t
n
X
i=1
(Vij −V+
i)2, j = 1,2, ...m (4.10)
D−
j=v
u
u
t
n
X
i=1
(Vij −V−
i)2, j = 1,2, ...m (4.11)
Step 5 : Finally Rank Index (R.I) is determined according to following formula:
R.I =D−
j
D+
j+D−
j
(4.12)
The node with highest value in this rank index announces itself as CH in that
region. Other nodes in that region, send join request to associate with the CH
and act as member nodes. After successful reception of the join request message,
the CH a acknowledges to all of its members. In this manner, wholeInetwork is
dividedIinto clustersIand inside each cluster, a potential node is selected as a CH.
40
R = 0
Diff. b/w CH and any
other node index value
> 0.1
Send data to BS
End
Yes
No
Start
R= R + 1
Re-election
All nodes die ?
Yes
No
R= Round
Figure 4.3: Procedure for CH Change
After successful clustering round, all member nodes in clusters, start normal com-
munication through their respective CHs. Along with the normal communication,
they will also compare their index value in their neighborhood table. If the index
value of any node is greater than index value of CH plus specific threshold (in our
proposed scheme it is 0.1), then the CH will no more be eligible to act as CH,
and nodes will perform re-election process within the cluster only by following
the steps discussed above. The significance of using threshold value is to avoid
CH change in every round. This process will continue till last node dies in the
network. The flow diagram of CH change is shown in fig. 4.3
41
4.2.4 Communication Phase
After CH selection and cluster formation, communication phase starts. The multi-
hoping communication model is considered by our proposed scheme because it is
the more realistic and practical one. The nodes within five meters range of CH,
send their data directly to CH, however, the nodes that are at a distance of
greater than five meter from CH, perform multi-hoping with other nodes coming
in their way to communicate with the CH. Same condition is also applied on CHs
when they communicate with the BS. The CHs within twenty meters range of BS,
communicates directly to BS, where as the remaining CHs perform multi-hoping
via other CHs. The purpose of using multi-hoping is to increase network stability
and life time. Figure 4.4 shows intra and inter-cluster communication.
Figure 4.4: Inter and Intra-cluster Communication
Complete flow diagram of our proposed scheme, including neighbor discovery, CH
selection and both inter and intra-cluster communication, is shown in fig. 4.5 on
next page.
42
Node Deployment
Neighbor Discovery
Sharing of Information
(four criteria values
among neighbors)
CH
Send Join Request
Cluster Formation
Node within
5m of CH
Broadcast CH
advertisement
Receive CH
advertisement
Receive Join
Request
Send Data to CH
End
CH selection based on
MCDM
Yes No
No
Perform Multi-hoping with
Neighbor nodes
Start
Yes
CH within
20m of BS
Send aggregated
data to BS
Perform Multi-hoping with
Neighbor CHs
No
Yes
(Normal Node)
Intra-cluster
communication
Inter-cluster
communication
Aggregate & compress
data
Figure 4.5: Flow Diagram of Proposed Scheme
43
Chapter 5
Simulation and Results
44
Chapter 5
Simulation and Results
In this chapter, we compare the performance of our proposed scheme with LEACH
[28] and previous fuzzy based CH selection scheme [31], through simulation using
MATLAB.
5.1 Simulation Environment
In our simulation n number of sensor nodes is randomly dispersed in a field of
100m×100m. The BS is located at corner of the field. In our simulation we
made some assumptions. First is that sensor nodes are continuously monitoring
the environment and always have data to be sent to the BS. Second is that wireless
channel is free of signal collision and interference. Simulation parameters are given
in Table 5.1.
Table 5.1: Simulation Parameters
Parameter Value
Network Area 100m x 100m
Number of Nodes (n) 100
BS Position (50,100)
Initial Energy 0.5 J
Data Aggregation Energy 50pj/bit/report
Data Packet Size 4000 bits
Hello Packet Size 200 bits
Transmitter Electronics (EelectT x) 50 nJ/bit
Receiver Electronics (EelecRx) 50 nJ/bit
Transmit Amplifier (Eamp) 100 pJ/bit/m2
Transmission Frequency Band 2.4 GHz
MAC Protocol CSMA/CA
45
5.2 Simulation Results
We evaluate the stability of the network by examining the numbers of rounds until
first node dies. Following graphs show simulation results of our purposed scheme
compared with previous schemes.
5.2.1 Number of Dead Nodes
Number of dead nodes per round in the network are shown in fig.5.1 .It is clear
from graph that in LEACH, first node dies around 170 rounds and previous fuzzy
model first node dies around 530 rounds where as our proposed scheme first node
dies 1600 rounds It shows that network lifetime and stability in our purposed
scheme is much better than previous clustering techniques. Reason for is that
LEACH is a single criteria based technique, where as previous fuzzy-TOPSIS
method , BS is performing CH selection process, which does not depends on
geographical conditions of nodes, where as in our purposed scheme every node
itself take decision for CH, considering the knowledge of neighbor nodes.
0 500 1000 1500 2000 2500
0
10
20
30
40
50
60
70
80
90
100
Number of rounds
Number of dead nodes
Network Stability
Proposed
previous Fuzzy
LEACH
Figure 5.1: Number of Dead Nodes
46
5.2.2 Number of Alive Nodes
Fig. 5.2 shows network lifetime. from graph, in LEACH last nodes dies around
1000 rounds, in previous fuzzy model it dies 1100 rounds, where is in our proposed
scheme network dies after around 2400 rounds. Network lifetime in our proposed
scheme is much better than previous schemes, due to proper CH selection and
using multi-hop communication.
0 500 1000 1500 2000 2500
0
10
20
30
40
50
60
70
80
90
100
Number of rounds
Nmber of alive nodes
Network Lifetime
LEACH
Previous Fuzzy
Proposed
Figure 5.2: Number of Alive Nodes
5.2.3 Energy Consumption
Energy consumption of network per round is shown in Fig. 5.3. It is observed
that our proposed scheme consumes less energy than previous schemes. A major
constituent of energy consumption is communication process. Almost 70 percent
of whole network’s energy is consumed in communication. So the proper commu-
nication model is very much necessary for any energy efficient clustering protocol.
In our proposed scheme we use multi-hop communication in both inter and intra-
cluster communication, this is the main reason of lower energy consumption in
our proposed scheme is much less than LEACH and previous fuzzy based scheme.
In previous schemes, energy variation at different points is observed and it due to
re-selection of CHs in each round. This CH re-selection is avoided in our proposed
47
scheme by introducing the CH changing threshold and that is also the main reason
of less energy consumption in our proposed scheme.
0 500 1000 1500 2000 2500
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Number of rounds
Energy consumed per round
Network Energy Consumption
LEACH
Previous Fuzzy
Proposed
Figure 5.3: Energy Consumption of Network
5.2.4 Change of Cluster Head
One major drawback of LEACH is that it performs CH election process in every
round or communication cycle. Hence in every round, all nodes send their complete
information to BS, if it is distributed. In centralized algorithm, like C-LEACH,
nodes share their information with all neighbor nodes in every communication
cycle. This shearing of information or sending information to BS in every round
require control overhead packets. Hence protocols based on re-election of CHs in
every round have large number of control overhead packets. Resulting in more
energy consumption.
In our proposed scheme we define a threshold, CH will only change if difference
between index value of CH and any other node in that cluster exceed that thresh-
old. Due to small number of variations in CH, there are very small number of
control overhead packets. Figure 5.4 shows the number of CH change during each
round.
48
0 500 1000 1500 2000 2500
0
2
4
6
8
10
12
14
16
18
20
Number of rounds
Change of CHs per round
Cluster Head Stability Ratio
LEACH
Previous Fuzzy
Proposed
Figure 5.4: Change of CHs
5.2.5 Control Overhead (Hello) Packets
The control overhead or Hello packets are the control signals required for any type
of data processing in WSNs. Lager the control overhead packets, greater will be
the energy consumption, Hence for any energy efficient clustering protocol it is
necessary that these packets should be minimized. In our proposed scheme CH
change is very rare and we use distributed algorithm for selection of CH. These
are main reasons for minimization of control overhead packets. Figure 5.5 shows
comparison of our proposed scheme with LEACH and previous fuzzy model. It
can clearly observed that control overhead packets in our proposed scheme are
very small as compared to other two schemes.
5.2.6 Packets Sent to Base Station
Number of packets to BS is are shown in fig. 5.6. Due to proper CH selection
and communication model, network throughput of our proposed scheme is much
greater than previous schemes.
49
0 500 1000 1500 2000 2500
0
1
2
3
4
5
6
7
8
9x 104
Number of rounds
Number of Hello packets
Network Control Overheads
LEACH
Previous Fuzzy
Proposed
Figure 5.5: Control Overhead Packets
0 500 1000 1500 2000 2500
0
0.5
1
1.5
2
2.5
3
3.5
4x 104
Number of rounds
Number of packets to BS
Network Throughput
LEACH
Previous Fuzzy
Proposed
Figure 5.6: Packets sent to BS
50
Chapter 6
Conclusion
51
Chapter 6
Conclusion
In this thesis, we present a new clustering technique based on multiple criteria.
We consider four criteria, including residual energy, number of neighbor nodes,
distance form BS and average distance between a node and its neighbors. We use
distributed algorithm for selection of CHs, means nodes themselves take decision
to become CHs or not. We control frequent change of CH in our proposed scheme.
If the index value of any node is greater than index value of CH plus specific
threshold (in our proposed scheme it is 0.1), then the CH will no more be eligible
to act as CH, and nodes will perform re-election process within the cluster. The
significance of using threshold value is to avoid CH change in every round. This
process will continue till last node dies in the network. We also improve both
intra and inter-cluster communication by using multi-hop communication. The
nodes within five meters range of CH, send their data directly to CH, however,
the nodes that are at a distance of greater than five meter from CH, perform
multi-hoping with other nodes coming in their way to communicate with the CH.
Same condition is also applied on CHs when they communicate with the BS. The
CHs within twenty meters range of BS, communicates directly to BS, where as
the remaining CHs perform multi-hoping via other CHs. The purpose of using
multi-hoping is to increase network stability and life time.
We perform MATLAB simulation to compare the results of our proposed scheme
with LEACH and previously proposed fuzzy based centralized clustering model.
All results show that the network performance of our proposed scheme is much bet-
ter than perviously proposed clustering techniques. This improvement is achieved
using distributed algorithm, using multi criteria for CH selection and using multi-
hop communication model in both intra and inter-cluster communication.
52
References
[1] Chaurasiya, S.K.; Pal, T.; Bit, S.D.; ”An Enhanced Energy-Efficient Proto-
col with Static Clustering for WS”. International Conference on Information
Networking (ICOIN), 2011 Kuala Lumpur, Malaysia, on page(s): 58 63.
[2] Ismail Olalekan Eleburuike; Soetan Simeon Adekunle, ”EnergyEfficient Wire-
less Sensor Network Using Hierarchical Routing Technique”. Blekinge Institute
of Technology, Karlskrona, Blekinge, Sweden, 2010.
[3] Opricovic, Seram, and Gwo-Hshiung Tzeng. ”Compromise solution by MCDM
methods: A comparative analysis of VIKOR and TOPSIS.” European Journal
of Operational Research 156.2 (2004): 445-455.
[4] Shah, T., N. Javaid, and T. N. Qureshi. ”Energy Efcient Sleep Awake
Aware (EESAA) Intelligent Sensor Network Routing Protocol.” arXiv preprint
arXiv:1212.4106 (2012).
[5] W. Heinzelman, A. Chandrakasan, and H. Balakrishnan, ”Energy-efcient rout-
ing protocols for wireless microsensor networks,” in Proc. 33rdHawaii Int. Conf.
SystemSciences(HICSS), Maui, HI,Jan. 2000.
[6] Ding Yong and Qin Xiao-Ming, ”EERM: Energy Efficient Routing Metric for
Wireless Sensor Networks,” in First International Conference on Pervasive
Computing Signal Processing and Applications (PCSPA), 2010, pp. 171174.
[7] K. Sohrabi, J. Gao, V. Ailawadhi, and G. J. Pottie, ”Protocols for self-
organization of a wireless sensor network,” in IEEE Personal Communications,
vol. 7, no. 5, pp. 1627, Oct. 2000.
[8] M. N. Elshakankiri, M. N. Moustafa, and Y. H. Dakroury, ”Energy Efficient
Routing protocol For Wireless Sensor Networks,” in International Conference
on Intelligent Sensors, Sensor Networks and Information Processing, 2008,
pp.393 398.
[9] J. Feng, F. Koushanfar, and M. Potkonjak, ”Sensor Network Architecture,”
Handbook of Sensor Networks, 2004.
53
[10] ”Evaluation of Routing Protocols in Wireless Sensor Networks” - Arkiv EX
- Blekinge Tekniska Hgskola.
[11] L. B. Ruiz, J. M. S. Nogueira, and A. A. F. Loureiro, ”Sensor Network Man-
agement,” Handbook of Sensor Networks: Compact Wireless and Wired Sens-
ing Systems, M. Ilyas and I. Mahgoub, Eds., Ch. 3. CRC Press, 2005.
[12] H. K. Kalita and K. Avijit, ”A New Algorithm of Self Organization in Wireless
Sensor Network,” algorithms, vol. 3, p. 4, 2010.
[13] Martin Haenggi, ”Sensor Network Management,” Handbook of Sensor Net-
works: Compact Wireless and Wired Sensing Systems, M. Ilyas and I. Mah-
goub, Eds., Ch.1.CRC Press, 2005.
[14] K. Akkaya and M. Younis, ”A survey on routing protocols for wireless sensor
networks,” Ad Hoc Networks., vol. 3, no. 3, pp. 325349, May 2005.
[15] H. Hassanein, K.Xu and Q.Wang, ”Sensor Network Management,” Handbook
of Sensor Networks: Compact Wireless and Wired Sensing Systems, M. Ilyas
and I. Mahgoub, Eds., Ch. 9. CRC Press, 2005
[16] C. S. Raghavendra, K. M. Sivalingam, and T. F. Znati, ”Wireless sensor
networks” Boston: Kluwer Academic Publishers, 2004.
[17] R. Szewczyk, E. Osterweil, J. Polastre, M. Hamilton, A. Mainwaring, and D.
Estrin, ”Habitat monitoring with sensor networks,” Commun. ACM, vol. 47,
no. 6, pp. 3440, Jun. 2004.
[18] Y. Miao, ”Applications of sensor networks,” in Seminar Wireless Self-
Organization Networks Published, 2005, pp. 0502.
[19] W. Dargie, C. Poellabauer, ”Fundamentals of Wireless Sensor Networks”, A
John Wiley and Sons, Ltd., Publication, 2010.
[20] J. N. Al-Karaki and A. E. Kamal, ”Routing Techniques in Wireless Sensor
Networks: A Survey,” IEEE Wireless Communications, vol. 11, no. 6, pp. 6-28,
2004.
[21] R. C. Shah and J. Rabaey, ”Energy Aware Routing for Low Energy Ad Hoc
Sensor Networks,” in Proceedings of the IEEE Wireless Communications and
Networking Conference, Orlando, USA, 2002.
[22] C. Intanagonwiwat, R. Govindan and D. Estrin, ”Directed diffusion: A scal-
able and robust communication paradigm for sensor networks,” in Proceedings
of the 6th International Conference on Mobile Computing and Networking
(Mobicom), Boston, USA, 2000.
54
[23] D. Braginsky and D. Estrin, ”Rumor Routing Algorithm for Sensor Net-
works,” in Proceedings of the first Workshop on Sensor Networks and Appli-
cations, Atlanta, USA, 2002
[24] K. Sohrabi and J. Pottie, ”Protocols for Self-Organization of a Wireless Sensor
Network,” IEEE Pers. Commun., vol. 7, no. 5, pp. 16-27, 2000.
[25] A. A. Abbasi and M.Younis. (2007,Oct.), ”A survey on clustering algorithms
for wireless sensor networks”, in Computer Communications[online]. Vol.30(14-
15), pp. 2861- 2841
[26] D. Wei and A. Chan, ”Clustering A Hoc Networks: Schemes and Classifica-
tions” Proc. of IEEE International Workshop on Wireless Ad Hoc and Sensor
Networks (IEEE IWWAN 2006), New York, June 28-30, 2006.
[27] S.Ghiasi, A. Srivatava, X.Yang, and M.Sanafzadeh, ”Optimal Energy Aware
Clustering in Clustering in Sensor Networks”, in Sensors, volume 2, pp. 258-
269, Feb. 2002.
[28] W. Heinzelman, A. Chandrakasan and H. Balakrishnan, ”Energy-efficient
routing protocols for wireless microsensor networks,” in Proc. 33rd Hawaii
International Conference System Sciences (HICSS), Maui, HI, 2000.
[29] B. P. Heinzelman, A. P. Chandrakasan and H. Balakrishnan, ”An applica-
tion specific protocol architecture for wireless microsensor networks,” IEEE
Transactions on Wireless Communications, vol. 1, no. 4, pp. 660-670, 2002.
[30] Latif. K, Jaffar. M, Javaid. N, Saqib. M. N, Qasim. U, Khan. Z. A, Per-
formance Analysis of Hierar- chical Routing Protocols in Wireless Sensor Net-
works, 5th International Workshop on Next Generation of Wireless and Mobile
Networks (NGWMN 2012).
[31] Azada, Puneet, and Vidushi Sharmab. “Cluster Head Selection in Wireless
Sensor Networks under Fuzzy Environment.” ISRN Sensor Networks, vol. 2013,
Article ID 909086, 8 pages, 2013. doi:10.1155/2013/909086.
55
