CHAPTER 1

INTRODUCTION

Wireless Sensor Network

The advantages of using local task management with mobile agents, versatile approach,

mutual signal, expandable, scalability and reduces network transmission and delays [1-

3]. Mobile agents do the process of data collection and data filtering. Unwanted packet

flow is managed in data aggregation to reduce the congestion. The method of data

filtration removes the duplicate packets. The base stations use the bridges for the

purpose of communication. In such networks, a mobile individual links with nodes to

their closest base station, so they can communicate with each other[4]. Definitions of

these types of networks are local area networks (WLANs) and wireless offices.

Mobile wireless networks which are mobile networks without infrastructure are also

known as ad hoc networks. In infrastructure-less networks all nodes can switch from

one position to another without fixed routers being limited. Nodes can link up with each

other in a dynamic way. The mobile nodes serve as routers that establish and maintain

routes to other network nodes[5]. The example of such networks are emergency search-

and - rescue operations, meetings or communication apps in which individuals want to

exchange information quickly[6-8].

The wireless design of these networks made them popular in different areas, because

their setup and deployment is easy and cost-effective. Among these wireless networks

is known as the wireless sensor network, which is the one more common. Wireless

Sensor Networks ( WSNs) have recently drawn a great deal of interest from researchers

1
because of their persistent capacity and large use in Cyber Physical Systems[9], the

Internet of Things (IoT)[10] and other areas[11]. Another explanation for the success

of wireless sensor networks is the rapid growth of sensor technology, and the cost of

sensor devices. The advancement in sensor technology allows to track cars, equipment

and structures using sensors in an simple way. In recent years, the sensor technology

has been combined with networks such as wireless networks and ad hoc mobile

networks to monitor the performance of machines, schemes and environment. Wireless

sensor networks ( WSNs) are known as autonomous, spatially distributed sensor

devices[12]. They use sensors to monitor and record environmental physical conditions.

The information gathered from the sensor nodes is stored at the central location.

Wireless sensor networks also rely on wireless connectivity and network spot-forming

such as wireless ad hoc networks so that the sensor node collects the data. The network's

sensor nodes collect data about temperature, sound, pressure etc. The main source of

the wireless sensor network was military for battlefield surveillance purposes[13] but

is used today in many areas such as health[14], customer products monitoring[15] and

control machines[16] etc.

The wireless sensor network comprising hundreds and thousands of "nodes" and

connecting each node to one or more sensor nodes. The key function of sensor nodes is

to senses the activities of the surrounding environment. -- sensor network therefore

consists of four parts a) Localized or distributed assembled nodes b) a network (wired

and wireless) c) a central point d) a central point resource for collecting the information

[17]. So every node in the sensor network performs the computation and senses the

operation of information. Since large data computation is involved in the sensor

2
network system needed for managing them[18]. Figure 1.1 prescribes the wireless

sensor network architecture.

Figure 1.1: Architecture of wireless sensor networks [5]

Architecture

The number of the sensor nodes can be hundreds or thousands, combined with a BS to

create a standard WSN network. They are designed to process, collect and integrate

data along with the communication with other nodes. Creating the sensor node

focuses primarily on energy conservation, cost reduction, complexity reduction, and

robustness[19]. Figure 1.2 shows the basic components of a sensor node that are: unit

of sensing, unit of processing, unit of transmission and unit of power.

• Sensing unit: The sensing unit consists of two parts: sensors and digital

converter (ADC) analogues. Sensors are instruments that track different

parameters of the environment, such as temperature , light, acceleration,

humidity , pressure, motion etc. The output of the sensor is usually an electric

3
 signal which is analog in nature. Thus, an ADC is used to convert the analog

signal generated by the sensor into digital form for processing and storage.

Figure 1.2: Sensor node architecture

• Processing unit: The processing unit contains a small RAM (Random Access

Memory) disk, and a microcontroller for local data processing. It collects data

from various sensors, analyzes the collected data, and stores it to a memory

location.

• Transmission unit: The transmission unit usually consists of communication

medium and transceivers to transmit information either directly or through

certain relay nodes between the sensor nodes and the BS. The transmitting

medium may be optical, radio frequency, or ultrasound depending on the

application criteria the transceivers transmit and receive information from.

4
 Through WSN the communication cycle uses more energy through

transmission, receipt, and idle state than the method of computation. Therefore

the sensor node's transceiver radio must be kept in a sleep state whenever

possible to reduce WSN's energy consumption.

• Power unit : The power unit consists of a battery and tension generator. The

battery is the sensor node energy storage unit that is used to power up the

other devices to track and calculate the external environment's physical

parameters. The sensor node's lifetime depends on the battery that is equipped

with the power unit. The voltage regulator inside the power unit provides the

necessary degree of constant output voltage regardless of input fluctuations

[20,21].

A wireless sensor network is composed of a community of wireless sensor nodes

that work collaboratively to collect and transmit information from the target area to the

sink(s). The sink can be located within the sensing area or outside. The transmission of

data from the sensor nodes to the sink is affected either by long distance single-hop

communication or by short distance multi-hop communication[22]. There are two

styles of architectures respectively – single-hop network architecture and multi-hop

network architecture.

• Single-hop network architecture: For single-hop architecture, the sensor nodes

collect the information periodically and forward it directly to the sink as shown

in Figure 1.3. This architecture is simple to implement and needs minimal

overhead for transmitting data to the sink. When the transfer distance increases,

the amount of energy needed for data transmission increases. The distance

5
 between the sensor nodes and the sink is perfect when the sink is located outside

of the sensing area. So the energy cost of transmitting the data to the sink is

rising. Given that the energy resources of the wireless sensor nodes are limited,

this long-distance transmission is undesirable and should be avoided if the

deployment environment does not allow the wireless batteries to be recharged

or replaced [23].

• Multi-hop network architecture: For multi-hop architecture the wireless sensor

nodes forward their data to the neighboring nodes nearer the sink instead of

sending data directly to the sink. And these nodes forward the data to the sink.

By this way contact over long distances is avoided. In addition, multi-hop

architecture is of two types: flat or hierarchical [24].

Figure 1.3: Single hop network architecture

6
• Flat Architecture: All Wireless Sensor Nodes are viewed as equals in the flat

architecture shown in Figure 1.4. Transmission of data from distant nodes to the

sink is done through multi-hop paths. Here the nodes of the sensors use their

neighbors as relays. Therefore, long distance communication is accomplished

by means of several short hops [25].

Figure 1.4: Flat multi-hop network architecture

• Hierarchical Architecture: Generally the data produced by the sensor nodes

deployed near each other is highly correlated in wireless sensor networks.

Therefore, if the data from the nodes near to each other is aggregated before

being sent to the sink, a significant amount of redundant data would be

eliminated. The aggregated data is then forwarded to the sink. Considering that

7
the energy required for transmission is directly proportional to the number of

bits transmitted, the aggregation of data results in energy efficiency. The

grouping of wireless sensor nodes together into clusters is called clustering for

data aggregation and forwarding. Within every cluster one of the nodes is

elected to be the cluster leader. The head of the cluster collects data from all

members of the cluster, aggregates it and sends the aggregated data to the sink.

Such architecture is referred to as hierarchical architecture because clustering

uses two layers of nodes – cluster leaders and cluster headers. The cluster heads

may use either single-hop or multi-hop transmission[26] for data transmission

into the sink. Figure 1.5 and Figure 1.6 demonstrate the wireless sensor network

architectures that employ single-hop and multi-hop transmission, respectively.

Figure 1.5: Single-hop clustering architecture

8
Through multi-hop transmission, remote cluster heads forward their data to

neighboring cluster heads closer to the sink. The cluster heads near the sink

relay data to the sink [28].

Figure 1.6: Multi-hop clustering architecture

9
 Characteristics of WSN

The various characteristics of WSN are as follows [29,30]:

• Battery resources: The battery resources depend on the growing node in the

sensor network. The node is active inside the network until the available battery

resources. Therefore the network's key concern is the safe and successful usage

of battery energy.

• Mobile nature of nodes: Every node on the network has the feature of mobility.

Without restriction the nodes can switch from one network to another network.

• Heterogeneity and Homogeneity nodes: The wireless sensor network consists

of both the Heterogeneity and Homogeneity node types. Homogeneity nodes

means that the forms of nodes are similar. Node type heterogeneity means that

nodes are of different forms.

• Scalability: Wireless sensor network is important feature. The number of nodes

in the network can easily be increased and diminished. It defines the Wireless

Sensor Network expandability feature.

• Easy to use and Deploy: Wireless sensor network can be implemented quickly

without the need for wired infrastructure. The network of wireless sensors is

wireless in nature so it is easy to use and deploy.

Advantages of WSN

The advantages of wireless sensor network are as follow [31]:

1. Networks can be built without a fixed network requirement.

2. Applicable in those areas where connectivity such as sea, rural area, mountains

etc. is difficult.

10
 3. Instant networks can be built where additional workstations are required.

4. The network deployment cost is smaller than that of the wired network.

5. The cost of buying wires reduced and complicated wired network not needed

for these network types.

6. Easily add the new device or node to the network.

7. It can break large networks into smaller networks.

8. Data is stored and accessed via centralized location.

Disadvantages of WSN

Disadvantage of wireless sensor networks are as follow [31].

1. WSNs are less secure anyone can be part of this network and hack the important

data.

2. Low latency compared with wired networks.

3. Configuration is difficult compared with wired networks.

4. This is influenced by the external environment including vibrations,

microwaves, attenuation, walls etc...

5. Using the radio waves for communication purposes the waves can easily spread

from one location to another.

6. Mobility of nodes also make it more susceptible to various types of attacks.

Applications of WSNs

This section describes the various applications of WSN. Various WSN applications

included in Figure 1.7 are discussed in detail in the following subsections.

11
 Precision Ahriculture

• Animal Habitat Monitoring

Automotive/Vehicular Application

Monitoring health of structures

Military Applications
Applications of WSNs

Area Monitoring

Healthcare monitoring

Polluting of the environment monitoring

Forest fire detection

Landslide detection

Water Quality Monitoring

Industrial Monitoring

Natural Disaster Preventions

Figure 1.7: WSN applications

12
• Precision Agriculture: WSN may be used for monitoring in precision

agriculture: plant, soil , climate, insect, disease, and weed. Greenhouse that can

adapt to different plant species in different seasons is widely used for precision

farming[32]. Research in (Chi1, Chen, and Gao), for example, describes the

implementation of a greenhouse climate monitoring program. Device sensor

node consists of: low-power Texas Instrument MSP430F149 controller (TI-

MSP430-datasheet); Chipcon CC1100 module (TI-CC2500-datasheet) as

transceiver; soil humidity, temperature, and pH sensors. Many energy-saving

methods make up the protocol stack. The agro project Low Frequency Array

tests the microclimate in a potato field to provide information on how

phytophtora, a fungal disease, can be combated. Sensor nodes that calculate

temperature are used for this; atmosphere and soil humidity are deployed in the

field. Also in the field is a weather station that tracks luminosity, air pressure,

precipitation, wind intensity and direction. Each 10 minutes sensor nodes

transmit sensor data to MS. The data is then forwarded for thorough review to

the Internet. Finally, a decision support system lets the farmer take steps such

as varying amounts of fertilizer and pesticides (Baggio)[33].

• Animal Habitat Monitoring: A representative example of animal habitat

monitoring is the installation of a sensor network on Great Duck Island

consisting of 32 nodes[34]. The goal was to track a bird storm petrel's natural

environment and its actions according to climate change. The sensors could feel

the ambient temperature, barometric pressure, and humidity. Furthermore,

passive infrared sensors and photo resistors were mounted to detect their

presence in the burrows of birds and the surrounding environment. Aggregated

13
 data through sensor nodes was transferred to a local MS that maintained the

database. The database was available via the internet and was used to duplicate

environment to another remote location for bird protection. WSN also leads to

a greater understanding of cattle behaviour, such as their feeding patterns, herd

behavior, and contact with the world around them. The information generated

by sensor nodes helps farmers understand pasture status and best ways to use

those tools. Attached to cattle collars in one such application defined in sensor

nodes. The nodes communicated in a peer to peer fashion to the location of the

animal's Global Positioning System (GPS). The cattle's person and herd

behaviors were modelled based on the positioning data of each node and inertial

information. Farmers can use these models to plan grazing areas to avoid

environmental problems, such as overgrazing and soil erosion [35].

• Automotive/Vehicular Application: Vehicle WSN is now another growing

wireless sensor network area[36]. It helps avoid road collisions and keep cars

from crashing, preventing track management from restricting speeds. It supports

vehicle communication and also has huge applications, such as smart parking,

collision avoidance, track information, car theft detection and monitoring [37].

• Monitoring health of structures: The significant feature of the WSNs is bridge

and heritage safety surveillance and historic buildings. Compared with

conventional approaches the program has many advantages [38,39].

• Military Applications: Sensor network 's rapid deployment, self-organization,

and fault tolerance characteristics make them a valuable technology for military

applications [40]. WSNs are used as an integral part of the structures of military

command , control , communications, networking, surveillance , tracking and

14
 targeting. Sensor networks are used for combat operations in hostile areas where

activity by humans is too risky. Some of the applications include tracking

devices, battle field surveillance, identification of opposing forces and terrain,

and gathering enemy motion information [41].

• Area Monitoring: Area monitoring is very widely used in WSNs. The WSN is

deployed in area surveillance covering a region where any phenomenon is

normally to be monitored [42]. A military is an example of where sensors are

used to identify enemies; a civilian example would be gas or oil pipeline geo-

fencing. Surveillance of region is the most critical part [43].

• Healthcare monitoring: There may be two kinds of medical applications:

wearable and implanted [44]. Wearable devices are attached to the human body

surface, or even in near proximity to the user. Those are the implantable medical

devices that are inserted within the body. There are also several other uses, e.g.

measurement of body position and of the person, overall monitoring of the ill

patient in hospitals and also at homes [45].

• Polluting of the environment monitoring: In several cities (Stockholm,

London, and Brisbane) wireless sensor networks have been installed to track the

influence of hazardous gases for citizens [46]. They may use random wireless

connections instead of wired installations to make them more accessible to

check readings in different areas [47].

• Forest fire detection: Normally a Sensor Nodes network is installed in a forest

to detect any time a fire has started. The nodes typically have sensors to measure

temperature , humidity, and gasses generated in the trees or vegetation by

fire[48]. The first identification is crucial to get the fire fighters to act

15
 successfully; As wireless tests as Sensor Networks allow the fire brigade to

know when a fire starts you and when it can spread [49].

• Landslide detection: A landslide warning system uses a wireless sensor

network to track the subtle soil changes and modifications to various parameters

that may occur before or during a landslide. With the data gathered it may be

possible to know the arrival of landslides before it genuinely occurs [50,51].

• Water Quality Monitoring: In measuring water level and water properties in

rivers , seas, reservoirs, oceans, and dams, sensor networks support. With the

implementation of sensor network in the distributed manner water resources can

be easily handled where it is difficult to meet capacity [52,53].

• Industrial Monitoring: In the industrial area the sensors are used to for the

purpose of data logging and condition of the machine. For customized products,

edibility in fracturing units and (iii) broad band wireless networks for

communication across manufacturing. RMS facilitates (i) reconfiguration of the

manufacturing products (ii) monitoring and monitoring of each process through

a corporate network or manufacturing supervisor facilities capable of

reallocating manufacturing resources within and through multiple plants is

required. This is suggested by a Reconfigurable Manufacturing System (RMS)

composed of: (i) wireless sensor and actuator network for machine control and

monitoring (ii) wireless personal area networks for local communication with

specific manufacture [54,55,56].

• Natural Disaster Preventions: The flood conditions can be easily prevented

with the installation of a sensor near the dams. The water level in the dams can

16
 be easily assessed and preventative steps can be taken to avoid flooding

situations in the region [57,58, 59].

Challenges in Wireless Sensor network:

In WSN, various challenges and issues are available which required to be handle

carefully. The challenges are as follow:

• Security: Safety relates primarily to authentication, honesty, privacy and non-

reputation. When the essential information flows from the networks thus rising

the degree of protection by itself. The necessity is to protect the important

information from the hackers by tightening the network level protection.

Various security mechanisms, such as authentication , encryption, decryption,

stenography can be implemented [60].

• Power Consumption: Wireless sensor networks are a perfect solution to those

areas where power supply to the sensor nodes is difficult to access. Sensors are

mounted in those types of position where it is difficult to adjust and charge the

battery periodically. The discussion in WSN needs resources. Without the

power the sensor nodes can't work. So to increase the life span of the WSN there

is need to efficiently utilization of network resources [61].

• Cryptography: For wireless networks, the encryption and decryption are

typically wired authentication mechanisms that can not be applied directly.

Small devices known as sensors which have limited battery resources, small

memory and processor are deployed in the WSN. By implementing these

techniques of encryption and decryption the more computations are needed that

use the more energy resources. The security methods should be managed like

17
 this they provide robustness to the network without using the more network

resources [62].

• Routing: Because of WSN 's wireless nature these networks are unstructured.

Thus the WSN routing protocols are not suitable for wired networks. The WSN

routing protocols should be light weight that requires less resources[63]. The

routing protocol should be data-centric, flood control and energy resource usage

minimal.

• Configuration: Configuring the network devices manual is the tedious work.

The network nodes that would dynamically alter their configurations when there

were some changes in the network[64]. The research becomes more repetitive

as the network increases in size.

Energy Efficiency in WSN

Energy usage is one of the major design problems in the architecture of wireless sensor

networks. The life of each sensor node depends on the dissipation of their energy.

Sensors with depleted batteries can no longer work in applications where the sensors

are not fitted with energy scavenging devices like solar cells [65,66].

Typically energy is absorbed in various operating phases in an arbitrary sensor node.

All steps are sending, detecting, receiving , storing (calculating), idle and sleeping [67]

The different modes of operation are described below:

• Sensing Energy: Energy is dissipated with the intention of activating sensing

circuitry inside the node and collecting environmental data. It is called energy

sensing. The magnitude of this energy depends on which mission the sensor is

assigned. Different sensors need varying energy levels during operation.

18
 • Transmitter Energy: A transmitter circuitry is required to transmit the data

towards the destination. A certain amount of energy is used for this operation

which is known as transmitter power. This depends on the power of the

transmitter, data packet size, distance between nodes and data transmission rate.

• Receiver Energy: A sensor node is also responsible for the forwarding of data

packets produced from other sensor nodes, as a relay node. Sensors must be able

to obtain certain packets of data for this operation. Receiver energy must be

used to access such data packets.

• Computation Energy: For such circuitries to work, the sensor node processing

unit must be triggered. In addition, additional computations must be

encountered if data aggregation or some form of in-network processing is

carried out.

• Idle Energy: A node is disabled in idle mode but the transceiver circuit is on,

and ready to receive. Therefore the same amount of energy is expended as idle

energy compared to receiving electricity.

• Sleep Energy: Active and inactive modes switch between the sensor nodes. A

node may consume more energy for sensing , processing, and transmitting or

receiving, in active state. While a node goes into sleep mode in inactive state

and thus consumes less energy.

Figure 1.8 shows the amount of energy consumption during these modes of operation.

19
 Figure 1.8: Energy consumption during different modes of operations

The cycle of data communication is mode of sending , receiving, idling and sleep. From

the figure it is clear that communication consumes more energy than sensing and

processing a data.

Wireless sensor networks are combinations of autonomous sensor nodes, also known

as motes, installed in a region where data is collected and transmitted to a main

computer in order to observe physical or environmental conditions. The main device is

referred to as a base station or sink node that links directly to an network or the Internet

through a gateway that allows remote users to access the data stored. Gateways are used

to interconnect two separate networks. Wireless sensor network has emerged as a

promising technology for a wide variety of specific needs areas. The WSN still has a

few issues to solve in order to become a mature network. The energy resources are one

20
of the key problems needed to manage them efficiently. Sensor typically has power

constraints in wireless sensor network[68]. The sensors are totally dependent on their

energy resources for carrying out the network activities. Each sensor is fitted with

battery operated units for sensing, data storage , data processing , memory, and

communicating. Radio signals are used with other sensors and the base station to

communicate intent. When performing communication tasks, the power is consumed

more in the sensor node followed by computation and sensing activities.

Another approach known as clustering and another is mobile agent for handling the

energy resources efficiently and effectively.

Clustering

The vast number of distributed sensor nodes had to be divided into small numbers of

manageable clusters to be handled. Every cluster had the head of the cluster fitted. The

head of a cluster may be static and dynamic. In networks where power leakage problem

occurs, nodes have non-zero power and have the static heads unable to communicate

with the sink nodes. The heads inside the selected network are dynamic heads based on

the available energy resources[70]. In short, clustering strategy increases the network's

life span by reducing the number of network transactions. The clustering problem

decreases the delays, calculations to decrease energy resource use.

The heads inside the selected network are dynamic heads based on the available energy

resources[70]. In short, clustering strategy increases the network's life span by reducing

the number of network transactions. The clustering problem decreases the delays,

calculations to decrease energy resource use. Each head of cluster collects data for all

members of its cluster. The CMs are also recognized members of the cluster, or leaf

21
 Figure 1.9: Cluster architecture

nodes[72]. In their sensing range each leaf node is connected to each other. Details

rotated from one leaf node to another leaf node until the destination is not reached. In

cluster, the data are used to move from one cluster head to another cluster head

gateways.

The advantages of splitting the networks into clusters are a) it improves scalability b) it

is simple to handle b) it reduces the amount of data to be transmitted by aggregating

and summing up the data, c) it decreases the number of relay nodes d) load balancing

between clusters e) it improves energy efficiency and f) it increases the lifespan of the

network etc.[73].

Clustering characteristics

In this section we opt for certain clustering characteristics to distinguish various

approaches to clustering. In general, we describe each clustered WSN as having three

key characteristics: cluster properties, CH properties, and process properties for

22
clustering. We'll describe each key section in depth in the following. Figure 1.9 includes

a detailed taxonomy of the clustering characteristics.

Cluster properties

Cluster properties are divided according to cluster requirements, such as cluster

number, cluster size, intra-cluster communication and inter-cluster communication[73].

Below is a brief description of each:

• The number of clusters: The number of clusters that were created may either

be constant (preset) or variable. This number is variable in the approaches

which randomly elect the CHs among the sensor nodes.

• Cluster size: The size of the clusters may be equal or unequal, depending on

the uniform distribution of the load among all the clusters created. The cluster

inequality is based on the distance between the nodes and the BS.

• Intra-cluster communication: The communications within a cluster can be

either direct or mulltihop depending on the clustering algorithm. Some

clustering approaches that involve multi-hop communication between the CH

and members when the number of CHs is low and the size of the clusters high.

• Inter-cluster communication: Because the sensor nodes have short-range

receiver / transmitters, the multi-hop solution is a suitable method for WSNs.

Nonetheless, some WSN applications presume that there is direct contact

between the CHs and the BS (usually in small-scale networks and conventional

approaches).

23
CH properties

Since the CH option is the key part of each clustering algorithm, the elected CHs have

a major impact on the efficiency of the clustering algorithm[74]. They mention the

following characteristics of the CHs:

• Mobility: The CHs can be stationary or mobile, either. Mobile CHs can travel

for a limited distance but mobile CHs' topology management process is more

complex than in a stationary CHs network.

• Node type: Compared to the normal nodes, the distributed CHs around the

network may be rich in resources; that is, the network supports node

heterogeneities. Or, the network may be homogeneous, and regular nodes

choose the CHs.

• Role: The chosen CHs will play different roles in the network, depending on

the algorithm. Those are relay and aggregation / fusion functions. A CH may

act as a simple relay node, such as managing and synchronizing data traffic or

conducting critical data aggregation / fusion operations.

Clustering process

In this section we review some characteristics of a clustering algorithm.

• Method: An algorithm for clustering can be either distributed, or centralized.

Due to the fact that WSNs are networks with a large number of nodes,

distributed approaches have become more common than centralized

approaches.

24
• Objectives: Clustering the nodes inside WSNs may have different goals. As

already stated, certain goals are more important than the others. A full list of

targets for the clustering.

• CH selection: That clustering algorithm has its own mechanism for choosing a

CH. But the CH election algorithms can generally be divided into three

categories: fixed, random, and attribute based methods. The CHs are elected in

preset before nodes are deployed in the sector. The CHs are chosen randomly

in random strategies, and attribute-based algorithms pick them based on some

of their characteristics, like the residual energy and distance to the BS.

• Algorithm complexity: Complexity of the algorithm shows how an algorithm

converges. Some algorithms converge in a variable time , depending on the

specifications of the network such as the number of CHs and some converge in

a constant time regardless of the specifications of the network.

• Clustering nature: Within the literature several clustering algorithms have been

proposed for WSNs. A limited number of such solutions are based on the

datacentric system , known as reactive networks. Some of the solutions

suggested are constructive and do not embrace the reactivity and some of them

use a mixture.

• Clustering dynamism: One approach to clustering can be either dynamic or

static. In dynamic approaches, the CHs are selected on the basis of the

network's current conditions, and most dynamic approaches operate in a real-

time scheme. In static approaches, the CH election and related operations are

performed regardless of the current network conditions [75, 76].

25
 Mobile Agent

A mobile agent (MAs)[77] is a piece of software that is performed on any changes that

have arisen in the network. Alternatively, knowledge rotates within the network and the

message floods the whole network. Mobile agent only transfers certain hops that

required any action to be carried out within the network. The most widely used model

for data aggregation in WSN is mobile agent. Mobile agent performs the function of

aggregating data at node level rather than sink level, reducing the overhead of redundant

network. The mobile agent has identified the entity in the context of WSNs as having

four components: mobile agent Identification (ID), itinerary, storage space, and

services, Where mobile agent ID is used to define the mobile agent uniquely, data space

is the data buffer of the agent carrying partially aggregated information, itinerary is the

order of sensor nodes to be visited during mobile agent migration and service is the

processing code carried with the agent as shown in Figure 1.10. The computing

paradigm based on mobile agents will dramatically reduce bandwidth consumption by

transferring the computational process to the position of the sensed data; otherwise its

raw transmission will consume more node energy[78]. This survey thus examines the

recently proposed energy-efficient clustering schemes and mobile-agent based data

aggregation schemes in WSN.

The mobile agents switch hop to hop before it hits destination. Within the network,

mobile agents carry out their mission autonomously and intelligently. Mobile agents

perform better if nodes are well connected, cluster is low in size. The advantages of

local task management using mobile agents, versatile approach, mutual signal,

expandable, scalability and reduces network transmission and delays. Mobile agents

26
do the process of data collection and data filtering. Unwanted packet flow is managed

in data aggregation to reduce the congestion. The method of data filtration removes the

duplicate packets. The technique of mobile agents also has other inconveniences, such

as protection and code caching[79].

Figure 1.10: Structure of Mobile Agent

Characteristics of a Mobile Agent

An agent is essentially characterized by the following points [80]:

• Nature : The agent can be a physical or virtual object; for instance, the physical

entity operates in the real world: a robot, an airplane or a vehicle. At the other

hand, the virtual object is an object that does not physically exist and that

represents a part of software or a computer module.

• Autonomy: That means it is not powered by the user's or other agent's commands

but by itself. And it's got some freedom of movement. The world in which the

agent can interpret and act reflects the space it occupies.

27
 • Communication: The agent has the ability to communicate directly with other

agents.

• Efficiency: This reflects the pace of the agent 's execution and interference.

Representative capability Agents have only a partial view of their world, that is,

they do not have a broad perception of all that happens to their surroundings.

An agent doesn't know all the information, he just has three kinds of knowledge;

knowledge of area, knowledge of power, communication and interpersonal

skills.

• Own resources: The agent requires a number of resources to be able to function

in the environment: power, CPU, amount of memory, access to certain sources

of knowledge. These tools make the agent dependent on his environment, but

they do give him some flexibility and are able to handle it.

• The objective: Therefore, the agent is a kind of "living organism" whose

activity, which is condensed to interact, act and possibly replicate, aims from

all other elements (perceptions, interpretations, acts, communications and

resources) to fulfill its needs and objectives.

There are other optional features that distinguish agents such as [81]:

• The reasoning: The agent can be connected to an expert system or other

methods of more or less complex reasoning.

• The control: It can be entirely distributed between the agents, but can be

dedicated to a certain class of agents as facilitator agents.

• Anticipation: The agent can more or less have the capability to anticipate events

future.

28
 • Granularity or Complexity: The agent can be very simple like a neuron but also

more difficult.

• The contribution: The agent participates more or less in resolving the problem

or in the activity overall system.

• Intentionality: An intentional agent is an agent driven by his goals. An intention

conveys the will of an agent to reach a goal or to perform an action.

• Rationality: Rational agents have criteria for assessing their actions, and select

according to these criteria the best actions to reach the goal.

• Adaptability: An adaptable agent is an agent able to manage his abilities

(communicational, behavioral, etc.) according to the environment.

• Commitment: One of the significant virtues of cooperative agents is the notion

of engagement. A cooperative agent schedules its actions with the other agents

by cooperation and negotiation. By drawing up a strategy to achieve a goal, the

agent gives himself the means to achieve it and then commits to carrying out

the acts that accomplish that aim; the agent believes he has created, which

causes him to behave in accordance with it .

• Intelligence: An intelligent agent is a cognitive agent, rational, proactive and

adaptive.

Operation of a Mobile Agent

The client gives a purpose to an agent. To achieve this goal, the agent moves in the

network of machines retrieving the services offered by these machines locally. Three

phases can be distinguished [82] as the Figure 1.11 shows:

1. Introduction of the mobile agent with the description of its mission.

29
 2. The execution of the mission by the agent.

3. Possible improvement of the results of the mobile agent.

Figure 1.11: Operation of Mobile Agent

Chapter Outline

The research work present in this thesis has been focused on the energy efficient

techniques in WSN. After a detailed discussion on the areas related to the objectives of

the work in Chapter 1, the rest of the thesis is organized as follows:

Chapter 2 provides the comprehensive literature review of existing protocols employing

the clustering and mobile agent techniques in WSN.

Chapter 3 presents the research gaps in the existing literature and objectives of the

research work that have increased the lifetime to sensor networks by introducing energy

efficient clustering technique assisted by mobile agents in WSN.

30
Chapter 4 presents a systematic anylysis on the energy-efficient clustering schemes and

mobile agent based schemes that have been employed by the protocols for data

aggregation in WSN. Afterwards, it presents a comparative analysis of clustering

schemes and mobile agent based schemes with emphasis on their objectives along with

their strengths and limitations. This comparative analysis facilitates the researchers to

select the appropriate clustering schemes and mobile agent based schemes employed

by protocols for energy-efficient data aggregation in WSN.

Chapter 5 presents Two-Fold Gravitational Search Algorithm (GSA) based approach

for WSN for IoT application that uses Mobile Agents (MA) for the data collection from

the nodes. In this approach first the GSA has been used for optimal selection of cluster

heads and then GSA has been used for optimizing itinerary of the Mobile Agent to

collect data from the sensor nodes. The various parameters used for optimizing the

cluster head selection were the remaining energy of the nodes, their communication

cost with the base station and number of neighbors having emergency data. While in

data transmission phase, the k means clustering was used for allocating source nodes to

MA and GSA was used for itinerary optimization. The proposed approach has been

compared with other clustering protocols in terms of network lifetime and energy

consumption. Also, GSA based itinerary planning of the MA has been compared with

other itinerary planning approaches for the MA in terms of success rate and task energy

consumption. The proposed protocol has shown better network stability, lesser energy

consumption and better success rate for MA as compared to other techniques.

Chapter 6 concludes the thesis by highlighting the main contributions of this research

work. It emphasizes on the contributions of the clustering and mobile agents for data

aggregation in WSN. Later, it provides recommendations for future research to improve

31
the performance of data aggregation protocols by exploring several interesting

directions.

32
33