Computers and Electrical Engineering 68 (2018) 62–75

Contents lists available at ScienceDirect

Computers and Electrical Engineering

journal homepage: www.elsevier.com/locate/compeleceng

Energy–efficient grid–based routing algorithm using intelligent

T
fuzzy rules for wireless sensor networks☆
⁎
R. Logambigaia, , Sannasi Ganapathyb, A. Kannana
a
Department of Information Science and Technology, CEG, Anna University, Chennai, India
b
School of Computing Science and Engineering, VIT University-Chennai Campus, Chennai, India

A R T IC LE I N F O ABS TRA CT

Keywords: In Wireless Sensor Networks, the sensor nodes are energy constrained and hence the design of
Grid-based routing energy- efficient routing protocol is an important challenge to be addressed. Among the different
Wireless sensor networks clustering procedures used for energy-efficient routing, grid-based clustering method is more
Clustering efficient in energy optimization. Therefore, we propose an Energy-Efficient Grid-based Routing
Fuzzy inference system
algorithm for sensor network in this work for preserving the energy in sensor nodes and to
Fuzzy rules
Residual energy
enhance the network lifetime. Moreover, routing is performed through a Grid Coordinator that
applies fuzzy rules to find the optimal route in order to reduce the number of hops in the routing
process. From the simulations carried out in this work, it is observed that the proposed routing
algorithm provides better performance in terms of residual energy and network lifetime when it
is compared with other existing grid and cluster based routing protocols.

1. Introduction

Wireless Sensor Networks (WSNs) comprise of sensor nodes that are communicating with one another through wireless links for
effective data collection and routing. In such networks, each of the sensor nodes are autonomous devices equipped with capabilities
for detecting, preparing and communicating with other nodes. One of the major issues in WSN is the optimization of energy available
in the sensor nodes in order to boost the network lifetime which can be achieved by developing an energy-efficient routing algorithm.
In a WSN, the sensor nodes are densely organized in a geographical region and the nodes are used to sense the environment and to
perform the task of data collection. Moreover, the data collected by the sensor nodes are routed through intermediate nodes in order
to send them to the base station called sink node where it is collected and stored for further analysis. In such a scenario, the sensor
nodes expend more energy in the transmission of data from the sensor nodes to the sink node and hence the energy consumption for
transmission becomes more than the energy required for processing the data at the base station.
In cluster based routing, nodes are grouped into small groups called clusters which help to reduce the energy by sending the data
through the cluster heads. In this scenario, the coordinators or cluster heads can send the data directly to the sink node if they are
present near to the sink node. On the other hand, the coordinators have to send the data to the sink node through multi-hop
communication with other coordinators if the coordinators are not present near to the sink node. If the same node remains as the
coordinator for a longer time, then its energy is drained fast and hence it will lead to reduction in network lifetime. This issue is
addressed earlier in Low Energy Adaptive Clustering Hierarchy (LEACH) [1] and the Hybrid Energy-Efficient Distributed (HEED)
clustering approach [2], where the rotation of coordinator nodes was proposed for achieving uniform energy depletion and load

☆
Reviews processed and recommended for publication to the Editor-in-Chief by Associate Editor Dr. M.H. Rehmani.
⁎
Corresponding author.
E-mail addresses: rlogambigai_14@yahoo.com (R. Logambigai), sganapathy@vit.ac.in (S. Ganapathy), kannan@annauniv.edu (A. Kannan).

https://doi.org/10.1016/j.compeleceng.2018.03.036
Received 28 June 2017; Received in revised form 22 March 2018; Accepted 22 March 2018
0045-7906/ © 2018 Elsevier Ltd. All rights reserved.
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

balancing. Among the different clustering techniques available in the literature to perform cluster based routing, grid-based clus-
tering is more proficient due to its straight forward, feasible and uniform structure handling method leading to reduction in routing
overhead.
In grid-based clustering, a sensing and data collection field is partitioned into smaller grids of certain length. From each grid, one
sensor node is chosen as the Grid Coordinator (GC) which aggregates the data collected from other sensor nodes which are present in
the same grid and routes the aggregated data to the base station. The nodes of the grid closer to the sink node are over-burdened by
subsequent data arrival from the other participating nodes. Therefore, the energy of those nodes is exhausted rapidly. This makes the
networking system to be disconnected from such nodes and this phenomenon is termed as hot spot problem. To address this issue,
many variations of the basic grid-based routing scheme have been proposed in the past by various researchers.
In order to address the energy consumption issue effectively, we propose a new Energy-Efficient Grid-based Routing (EEGBR)
algorithm in this paper which overcomes the hotspot problem and utilizes the energy of nodes more efficiently. This algorithm
consists of three different phases namely Grid formation, Grid Coordinator election and Grid-based routing phases. The GC for each
grid is elected by applying fuzzy logic where the fuzzy variables are used for GC selection based on the residual energy of the nodes,
their mobility pattern and the distance of the nodes from the sink node. As the transmission consumes energy based on the distance,
the distance to the sink is considered as an important parameter in this work to select the GC so that the energy utilization is
monitored perfectly. Moreover, the GC of neighbouring grid is also considered for making routing decisions. The major advantages of
the proposed algorithm are preservation of energy, reduction in transmission delay and an augmentation of the network lifetime. The
major contributions of this work are as follows:

• Proposed a new Energy-Efficient Grid-based Routing (EEGBR) algorithm to overcome the hot spot problem.
• Developed a fuzzy inference system for Grid Coordinator selection with two membership functions and fuzzy rules.
• Proposed a grid-based routing algorithm to enhance the network performance.
• Mobility of nodes is considered in the proposed Grid-based Routing algorithm and delay is reduced.
The remainder of this paper is composed as follows: In Section 2, the literature survey carried out in the areas related to the
proposed methodology is briefly explained. In Section 3, the proposed energy-efficient grid-based routing algorithm has been ex-
plained. In Section 4, assessment of this proposed work is carried out by providing the results and discussions. At last, this paper is
concluded in Section 5 by highlighting the salient features of this work and suggesting some future work.

2. Related works

In the past, numerous works have been carried out by many researchers in the areas of routing and grid-based routing in WSN for
effective communication [3,4]. While outlining the WSN, since energy assumes an imperative part in the system lifetime, more
concentration is to be given on energy utilization of the sensor nodes. In WSN, cluster based approach saves the energy of the sensor
nodes. In all the clustering algorithms, the cluster head election is imperative for the preservation of energy. Moreover in different
works on grid and cluster based routing, decision on next hop is made based on the leftover energy of the individual nodes in a cluster
[5]. In the past, fuzzy logic was used to make decisions on the formation of clusters [6,7]. In [7], a Fuzzy logic Based Unequal
Clustering (FBUC) approach has been utilized for making decisions in the selection of cluster heads in which the member nodes are
provided with a provision to join the cluster by communicating with the cluster head.
In [8], an energy-efficient Multi-Mode Cluster Maintenance (M2CM) algorithm was presented for clusters maintenance to improve
energy utilization rate and also to improve the network lifetime. This M2CM maintains the damaged cluster according to its change in
time and space field, node's remaining energy under threshold, load imbalance of the cluster head, joining in or moving out from
existing cluster. In [9], an energy-aware clustering protocol called Prolong Stable Election Protocol (P-SEP) was presented to max-
imize the network lifetime. In their work, the CH was selected randomly at each round to achieve load balance, energy consumption
and to extend lifetime of the network. In [10], the authors proposed a new routing algorithm in which fuzzy logic was used to
perform decision making based on different parameters namely the residual energy, the energy consumption by the nodes, number of
neighbour nodes, the average distance between the neighbouring nodes and the distance from the sink node. In their model, the
authors used a threshold on energy level for each of the routing technique namely intra-cluster and inter-cluster routings through
multi-hop communication mechanism in order to reduce the energy consumption.
In [11], the authors developed a new energy-efficient cluster-based routing protocol which forms balanced clusters and prolongs
the network lifetime through routing using cluster heads in WSNs. They used fuzzy c-means algorithm for cluster formation and
Sugeno fuzzy system to select the proper cluster head. The authors in [12] presented a multi-clustering algorithm where the sensor
nodes are clustered using different clustering algorithms. Moreover in their work, the cluster heads are not selected for each round in
order to reduce the number of transmission messages. In [13], the authors presented a novel fuzzy clustering scheme to find an
optimal clustering structure of three dimensional WSN in terms of network lifetime and energy consumption. However, the number of
non-CHs connected to CHs and the number of re-clustering rounds of their work are not providing energy efficiency when it is
compared with other related algorithms.
In the past, many grid-based routing protocols were proposed to provide effective routing facility in WSNs [14]. In [15], a grid-
based joint routing protocol was proposed for wireless rechargeable sensor networks. The important features of their algorithm
include the introduction of a technique for charging the local energy balance and the assignment of distinctive charging time at
various charging points. Preetha and Kevin [16] developed an energy-efficient zone clustering algorithm to increase the network

63
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

lifetime and to control the network topology. In [17], the authors proposed a Grid-Based Reliable Routing (GBRR) technique with the
association of clustering and grid-based routing features to solve the problem of burden on cluster heads. Moreover, their work
concentrated on the issue of the consistency of node-to-node link quality in WSN.
Rejina Parvin and Vasanthanayaki [18] proposed a new energy-efficient routing protocol called Enhanced and Optimized Energy-
Efficient Routing Protocol for effective routing in WSNs. In their model, the individual nodes are allowed to join the cluster when the
cluster heads are selected and they provide advertisement for the members to join their clusters. Their model is more useful for
energy-efficient routing since they consider individual node preference in the cluster formation process. In [19], the authors proposed
a reliable routing protocol for effective routing of collected data in wireless sensor networks. In this model, the authors provided
random back-off facility and they used the gradient routing technique in the cluster based routing process in order to reduce the
routing overhead. Moreover, the authors carried out a theoretical analysis on their proposed routing protocol and they also carried
out the simulation of their algorithm. From the experiments and theoretical analysis carried out by the authors, it was proved that
their model reduced the delay and increased the network performance in terms of energy efficiency and packet delivery ratio.
Nayak and Devulapalli [20] proposed anew energy-efficient and cluster based routing algorithm for efficient routing in WSNs. In
their model, the authors used fuzzy rules for selecting two types of cluster heads namely the cluster heads and the super cluster head.
They focused on routing with mobile sensor nodes by performing cluster based routing in order to send the data to the base station.
They have proved that their model increases the network life time by providing the energy-efficient routing process. However, the
cluster structures can be replaced with grid structure in order to enhance the performance further. In [21], the authors proposed a
new protocol using a fuzzy temporal approach for performing energy-efficient routing in WSNs. Their model used temporal con-
straints and fuzzy rules in the formation of clusters, cluster head election and routing. The authors have proved that their routing
protocol provided energy efficiency in the routing process.
In [22], a new rule based and delay constrained energy-efficient routing protocol was proposed for performing effective routing in
WSNs. The main advantage of their model is that the energy required for transmission of data is reduced and at the same time delay is
reduced without increasing the network overhead. In [23], the authors proposed a grid-based clustering algorithm for wireless sensor
networks. In this algorithm, at each grid the CH is elected based on Bollinger Bands. In their work, they used the grid Identity (ID) for
data dissemination without identifying relay nodes that is the grid with smallest ID has to transfer the data to the grid with higher
grid ID. The main drawback of their algorithm is that the data is transmitted based on the grid ID then grid with higher ID has to
gather the entire lower grid ID data in that particular grid row. The grid coordinators are near the sink over burden with the data
which drains energy very soon and results in hot spot problem.
In [24], the authors proposed a new protocol for efficient routing in WSN by applying mobility modelling, energy modelling and
by performing cluster based routing. The authors considered many networking issues such as energy efficiency, scalability of the
network and the load balancing techniques for enhancing the network performance. The authors also developed a trust model in
order to improve the energy efficiency and security. In [25], the authors proposed a new routing protocol called Energy-Efficient
Cluster based Secure Routing Algorithm (CEESRA) which is used not only to address the issue of performance through security, their
model is providing efficient routing paths to the users and hence it is more suitable for sensing, data collection and routing in order to
improve the network performance.
In [26], the authors proposed a Genetic Algorithm based Self-Organizing Network Clustering (GASONeC) method to optimize the
dynamic node clustering. In their work, the cluster head selection is performed after dynamic clusters are formed by the encoding of
genetic algorithm chromosomes. In [27], the authors presented a hybrid approach for controlling communication topology in
wireless sensor networks called Collaborative Topology Control Protocol (CTCP). In their work, clusters are formed through
minimum virtual connected dominating set and the cluster head uses transmission power levels for transmitting the information from
source to destination. In [28], the authors presented a framework which incorporates a dynamic clustering scheme. They presented a
hard or soft decision based threshold which can be tuned for the selection of cluster heads to optimize the frequency of re-clustering
within the network.
In spite of the presence of all these algorithms, the energy problem is still a challenging one due to the methods used for cluster
based selection. To overcome these issues in cluster head selection and routing, an energy-efficient grid-based routing algorithm is
proposed in this paper. In the proposed work, fuzzy logic is applied for the grid coordinator election and also the relay nodes selection
for data transmission. The major advantages of the proposed algorithm over the existing algorithms include the increase in packet
delivery ratio, energy efficiency, network life time and reduction in delay.

3. Energy–efficient grid–based routing algorithm

In this section, the proposed energy aware grid-based routing algorithm is explained. The main objective is to improve the system
lifetime of sensor nodes in a system. The EEGBR algorithm works in three phases. In the first phase, grid is formed in the network
region. In the second phase, the grid coordinator is elected for each grid. In the third phase, the data transmission from source to sink
takes place. The flow of the EEGBR algorithm is given in Algorithm 1.

3.1. Network model

The sensor network model discussed in this paper consists of sensor nodes which are placed in a grid structure and they vary from
100 to 1000 depending upon the experiment. The grid-based network structure which is used in the proposed model is shown in
Fig. 1. The assumptions which are made in this proposed work are as follows:

64
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Algorithm 1
Energy-efficient grid-based routing algorithm.
Phase I: Grid Formation Phase //R – Transmission Range
Step 1: Divide the whole network region into square of equal size grids with side of the grid to be R/2.83 as given in Eq. (2).
Phase II: Grid Coordinator Election Phase
//The nodes present in each grid is responsible for calculating their suitability to become the Grid
//Coordinator where fuzzy rules are applied to make decisions.
Step 2: Using Eqs. (3) and (4) fuzzify the input variables Distance_sink and Residual Energy.
Step 3: Apply the fuzzy rules given in Table 1 to the fuzzified values.
Step 4: The fuzzy rule with maximum value is the value of GC_Choice.
Step 5: For the crisp value apply the Eq. (5) to defuzzify the GC_Choice value.
Step 6: Send GC_Msg(ID, GC_Choice) to its neighbours within the grid.
Step 7: On receiving the GC_Msg each node compares the GC_Choice value of its own.
Step 8: The node with high GC_Choice value will be elected as Grid Coordinator.
Phase III: Grid-based Routing Phase
Step 9: Source node checks for the GC within its grid and transmits the data to its GC.
Step 10: GC in the source Grid calculate the Relay_Choice value among its neighbours.
//Calculate the Relay_choice value using fuzzy if-then rules are given in Table 2.
Step 11: Using Eqs (3) and (4) fuzzify the input variables Distance_sink, GC_Degree and Residual Energy.
Step 12: Apply the fuzzy rules given in Table 2 to the fuzzified values.
Step 13: The fuzzy rule with maximum value is the value of Relay_Choice.
Step 14: For the crisp value apply the Eq. (5) to defuzzify the Relay_Choice value.
Step 15: Node with maximum Relay_Choice value will be selected as the Relay node
Step 16: Relay node selects its next relay node by repeating the steps from step 10 to step 15 till the sink is reached

Fig. 1. Grid-based sensor node deployment.

i.The Wireless Sensor Networks consists of nodes having similar configuration and uniform energy of 2 Joules in each sensor node.
ii.Inside the grid, the sensor nodes have been deployed in a random manner.
iiiThe entire sensor nodes considered in this model and also the base station are static and hence there is no mobility parameter.
iv. All the nodes are energy constrained and hence they will be left after the complete energy is drained. Therefore, there is no way to
recharge the battery. The nodes are deployed in such a way that the distance between any two nodes is decided by the value of the
received signal strength.
v. The sink node or base station is kept inside the grid where the sensor nodes are deployed.

Since most of the sensor nodes use their energy for transmission, energy is an important constraint which is considered in this
work as a challenging issue.

3.2. Energy model

In this work, the existing Order Radio model [29] is used as the energy model for effectively measuring the energy consumption in
the sensor networks considered in our work. According to this model, the sender u sends the message of size 1 bit to a neighbour v
who distance is represented by d(u, v) consumes an energy of c(u, v) Joules. This model considered the energy cs which is used by the
node u for preparing the outgoing signal as the first part of energy consumed. The second part of energy cu is the amount of energy
needed to perform the compensation of the path loss on the signal used for transmission from u to v. The third part of the energy cr is
the amount of energy required by node v for processing the incoming signal which is coming from the node u. According to the
energy model used in this work, the energy consumption c(u,v) is denoted by the formula

c(u, v) = c1 + c2 × d(u, v) × α. (1)

where α is a path loss constant, c1 and c2 are constants whose values are dependent on the type of wireless devices and the
application environment used in the simulation.

65
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Fig. 2. Grid with grid ID.

3.3. Grid formation phase

In this phase, the system area is apportioned essentially into smaller square shaped regions called grids and the size of the grid is
figured as given in Eq. (2). The grids are built in light of the transmission range of the sensor node as shown in Fig. 3. The sensor
nodes are deployed and the sink is located in any grid. Each of the grid cells are uniquely identified by the grid identification number
(GId) as demonstrated in Fig. 2. By using Global Positioning System (GPS), each node determines their GId. The sensor nodes are able
to communicate from one grid to another grid in a single-hop.

R=a 2 +a 2

R = 2a 2

R
a=
2.83 (2)

where R is a transmission range and ‘a’ is the side of the grid.

3.4. Grid coordinator election phase

In a grid topology, sensor nodes are randomly deployed. Each sensor node is associated with the unique identification number and
grid identifier (GId). The sensor nodes in each grid are called grid members. To adjust the load in the grid, a sensor node is chosen as a
grid coordinator (GC). For a span of time, the specific node acts as a GC. At the end of the duration, the GC election process is to be
carried out by the current GC. The current GC broadcasts the message that a new GC is selected to its grid members. In most of the
existing work, the grid coordinator is chosen in view of the remaining energy of the grid members in the grid, but they did not
consider the distance of the coordinator to the sink. The energy of the coordinator will be decreased when the distance for trans-
mission increases.
The GC is selected among its grid members in each grid by fuzzy decision making system. The distance between the sensor node
and the sink node and the residual energy are the input parameters for the system and the output parameter is node elected as a grid
coordinator i.e. GC_Choice. Fuzzification unit of the fuzzy decision making system converts the real numbers also called linguistic
terms to fuzzy sets. The linguistic terms for Distance_Sink are Close, Medium and Far and for Residual Energy are Low, Medium and
High. The linguistic term for the output parameters GC_Choice are Very High, High, Medium High, Rather High, Medium, Medium
Low, Rather Low, Low, and Very Low. The linguistic terms are represented by triangular membership function and trapezoidal
membership function given in Eqs. (3) and (4).

Fig. 3. Grid size.

66
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Table 1
List of fuzzy rules used in grid coordinator selection.
Distance_Sink Residual_Energy Grid Coordinator_Choice

Nearer Higher Much_Higher

Nearer Medium_level Higher
Nearer Lower Rather_Higher
Medium_distance Higher Medium_Higher
Medium_distance Medium_level Medium
Medium_distance Lower Medium_Lower
Farther Higher Rather_Lower
Farther Medium_level Lower
Farther Lower Much_Lower

⎧ 0, x ≤ a1
⎪ x − a1
⎪ b1 − a1 , a1 ≤ x ≤ b1
Triangular (x ; a1, b1, c1) =
⎨ c1 − x , b1 ≤ x ≤ c1
⎪ c1 − b1
⎪ 0, c1 ≤ x (3)
⎩
where x is the member of the universal set, a1 represents lower boundary, c1 represents upper boundary and b1 represents the middle
point of the triangle.

⎧ 0, x ≤ a2
⎪ x − a2 , a2 ≤ x ≤ b2
⎪ b2 − a2
Trapezoidal (x ; a2, b2, c 2) = 1, b2 ≤ x ≤ c 2
⎨ d2 − x
⎪ d2 − c 2 , c 2 ≤ x ≤ d 2
⎪
⎩ 0, d2 ≤ x (4)

Here, x represents the member of the universal set, a2 represents the lower limit, b2 represents lower support limit, c2 represents
an upper support limit and d2 represents upper limit of the trapezoidal.
The relationship between the input parameter and the output parameter in the fuzzy inference system are expressed as IF-THEN
fuzzy rules. The fuzzy rules for the GC selection are given in Table 1.
Two extreme cases of the fuzzy IF-THEN rules for Grid Coordinator_Choiceare given below and all other rules fall within this two
extreme cases.

Case 1: If Residual energy is lower and distance to sink is farther then Grid Coordinator_Choice is much_lower.
Case 2: If Residual energy is higher and distance to sink is nearer then Grid Coordinator_Choice is much_higher.

The proposed system uses the Mamdani inference system since it provides optimal results by using Max-Min composition rea-
soning scheme in order to get the output. The result of the inference system needs defuzzification. The centre of mass technique is
used in the proposed system to calculate the crisp value of the fuzzy output and it is given in Eq. (5).
q
∑i = 1 x i μ (x i )
Z= q
∑i = 1 μ (x i ) (5)

where xi represents each member of the output universe of discourse, µ(xi) represents the membership degree of xi and q indicates the
number of members in the output universe of discourse.

3.5. Grid-based routing phase

Once the GC has been elected, the next phase is the grid-based routing phase. Whenever the source wants to transmit the
information to the sink, the routing takes place between source and sink through single-hop or multi-hop fashion. The individual
members in a grid send their formation to their corresponding GC and goes to the sleep state to save their energy. GC gathers the
information from every one of the nodes in the grid. It then processes the aggregates and sends the collected data to the sink node.
Moreover, we propose a new routing technique in this paper which consumes less energy and the load is balanced in each and every
GC. In this work, the data transmission from the source node GC to the sink node is performed through the relay nodes along the
route. Here, the relay nodes are GC of neighbour grids along the route. The source GC selects its relay node i.e. the next relay GC from
its neighbours.
Initially, source GC applies the fuzzy logic among the neighbours GC and selects the next relay GC that has maximum fuzzy value.
The proposed fuzzy system takes the residual energy of the node, the distance to the sink and the degree of the GC as the factors for
the next relay node selection. Similarly, the next relay node applies fuzzy logic to select its next relay node. The selection process

67
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Fig. 4. Typical grid-based WSN.

continues till the sink is reached. The neighbours of the source and the final route are given in Figs. 4 and 5.
The linguistic terms for Distance-Sink are Nearer, Medium_distance and Farther and the linguistic terms for Residual Energy and
GC_Degree are Lower, Medium_level and Higher. The linguistic terms for the output variable RelayChoice are Much_higher, Higher,
Rather_Higher, Medium_Higher, Medium, Medium_Lower, Rather_Lower, Lowerl and Much_Lower. For the relay node selection, the
membership functions and inference system are used as explained above. The fuzzy rules for relay node selection are given in Table 2.

3.6. Design of grid coordinator election and relay node selection

The main objective of the proposed work is to conserve the energy of the sensor nodes. This is achieved by electing the GC among
its grid members in each grid by using fuzzy decision making system. The input parameters of the system are Distance_Sink and
Residual_Energy and the output parameter is Grid Coordinator_Choice.
Table 3 shows the minimum and maximum values of input parameters given to the fuzzy inference system in order to make
decisions on GC election. The input range for each of the fuzzy input variables are given in Table 4 and their corresponding
membership functions are represented in Figs. 6–8.
For example, consider the values for the Distance_sink = 55 which lies in Medium_distance and Farther and Residual
Energy = 0.38 which lies in Medium_level and Higher.
Fuzzy Rules from the Table 1 are:

Rule 1: If Distance_Sink is Medium_distance and Residual_Energy is Medium_level then Grid Coordinator_Choice is Medium.
Rule 2: If Distance_Sink is Medium_distance and Residual_Energy is Higherthen GridCoordinator_Choice is Medium_Higher
Rule 3: If Distance_Sink is Farther and Residual_Energy is Medium_levelthen Grid Coordinator_Choice is Lower
Rule 4: If Distance_Sink is Farther and Residual_Energy is Higher then Grid Coordinator_Choice is Rather_Lower

Fig. 5. Grid-based routing.

68
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Table 2
Fuzzy rules for Relay_Choice.
Distance-Sink Residual_Energy GC_Degree Relay_Choice

Nearer Higher Lower Much_Higher

Nearer Higher Medium Higher
Nearer Higher Higher Rather_Higher
Nearer Medium_level Lower Higher
Nearer Medium_level Medium Higher
Nearer Medium_level Higher Rather_Higher
Nearer Lower Lower Higher
Nearer Lower Medium Rather_Higher
Nearer Lower Higher Medium_Higher
Medium_distance Higher Lower Rather_Higher
Medium_distance Higher Medium Medium_Higher
Medium_distance Higher Higher Medium
Medium_distance Medium_level Lower Medium_Higher
Medium_distance Medium_level Medium Medium
Medium_distance Medium_level Higher Medium_Lower
Medium_distance Lower Lower Medium
Medium_distance Lower Medium Medium_Lower
Medium_distance Lower Higher Rather_Lower
Farther Higher Lower Medium
Farther Higher Medium Medium_Lower
Farther Higher Higher Lower
Farther Medium_level Lower Medium_Lower
Farther Medium_level Medium Lower
Farther Medium_level Higher Rather Small
Farther Lower Lower Rather Small
Farther Lower Medium Lower
Farther Lower Higher Much_Lower

Table 3
Fuzzy variables used for grid coordinator selection.
Variable_Name Min_Value Max_Value

Sink_Distance 0 150
Residual_Energy 0 1

Table 4
Fuzzy variables and input ranges.
1. Distance_Sink 2. Residual Energy

Fuzzy variables Input ranges Fuzzy variables Input ranges

Nearer 0-40 Lower 0.0 – 0.35

Medium_distance 35-60 Medium_level 0.15 – 0.55
Farther 50-150 Higher 0.35 – 1

Fig. 6. Membership functions for residual energy.

69
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Fig. 7. Membership functions for Distance_Sink.

Fig. 8. Membership functions for GC_Choice.

Now, apply the standard triangular membership function for Distance_Sink which is Medium and Trapezoidal membership
function for Distance_Sink which is farther.
X = 55 (Medium_distance)
x−a c−x
f (x ; a, b, c ) = max ⎛min ⎛ , ⎞, 0⎞
⎝ ⎝b − a c − b⎠ ⎠

55 − 35 60 − 55
f (x ; a, b, c ) = max ⎛min ⎛ , ⎞, 0⎞ = 0.5
⎝ ⎝ 50 − 35 60 − 50 ⎠ ⎠

x−a d − x⎞ ⎞
For X = 55 (Farther), f (x ; a, b, c, d ) = max ⎛min ⎛
⎜ , 1, ,0 ⎟

⎝ ⎝b − a d − c⎠ ⎠

55 − 50 151 − 55
f (x ; a, b, c, d ) = max ⎛min ⎛ , 1, ⎞, 0⎞ = 0.167
⎝ ⎝ 80 − 50 151 − 150 ⎠ ⎠

Similarly, the triangular membership function and trapezoidal membership functions are applied for the residual energy.
0.38 − 0.15 0.55 − 0.38
For X = 0.38 (Mediumlevel), f (x; a, b, c ) = max ⎛min ⎛ , ⎞, 0⎞ = 0.85
⎝ ⎝ 0.35 − 0.15 0.55 − 0.35 ⎠ ⎠

0.38 − 0.35 1.1 − 0.38

For X = 0.38 (Higher), f (x ; a, b, c, d) = max ⎛min ⎛ , 1, ⎞, 0⎞ = 0.15
⎝ ⎝ 0.55 − 0.35 1.1 − 1 ⎠ ⎠

Now apply these values in Rules 1 to 4.

Rule 1: Minimum (0.5, 0.85) = 0.5, Rule 2: Minimum (0.5, 0.15) = 0.15
Rule 3: Minimum (0.167, 0.85) = 0.167, Rule 4: Minimum (0.167, 0.15) = 0.15

Maximum of the rules from 1 to 4 is given as 0.5 and hence it is in Medium. Moreover, its crisp value is coming between 30 and

70
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Table 5
Fuzzy input variables and their minimum and maximum values for Relay_Choice.
Variable name Min. value Max. value

Distance _Sink 0 150

Residual energy 0 1
GC_Degree 0 1

60. In defuzzification process, this value is applied to Z as shown in Eq. (5). Hence, the GC_Choice is 37.8. In this way, fuzzy rules are
applied for all the member nodes present in the grid and the node whose GC_Choice is maximum is elected as GC. Moreover, the relay
GC node for routing is selected by applying the fuzzy logic. Consider the following range values for Distance_Sink, Residual energy,
GC_degree. Considering the input range for each input variables are listed with minimum and maximum values are shown in Table 5
and the values are computed for those variables are shown in Table 6.
For an example, consider a GC in grid 2 as GC_2 and find its relay node from its neighbours. Assume GC_3, GC_4 and GC_6 are its
neighbours. Consider the values of Distance_Sink, Residual Energy and GC_degree for Node GC_3, GC_4 and GC_6 which are given
below:
GC_3 (65, 0.4, 0.3) , GC_4(70, 0.6, 0.8) and GC_6(45, 0.3, 0.4)
Apply the fuzzy rules for GC_3, GC_4 and GC_6. After applying the defuzzification process, the Z values of GC_3, GC_4 and GC_6 are
20, 18.5 and 39.6 respectively. Among these three neighbours GC_6 has maximum value, so it is chosen as relay node for node GC_2.
Similarly, for GC_6 the relay node is selected from its neighbours. This process is continued till the sink is reached. Thus the optimal
route is formed from GC_2 to the sink.

4. Simulation and results

The proposed EEGBR algorithm has been evaluated using NS2. Simulation is carried out with variable number of sensor nodes
deployed over a region of (200x200) m2 and with the starting energy of every sensor node as 0.5 J. The simulation parameters
utilized as a part of the proposed system are given in Table 7.
The proposed algorithm is experimented and the trial results are exhibited. The sink is located within the network region. The
proposed algorithm is compared in two ways. First the proposed EEGBR algorithm is compared with other fuzzy based routing
algorithms such as Multi-clustering algorithm using Fuzzy Logic (MCFL), Fuzzy-Topsis, Grid-Based Reliable Routing (GBRR), Energy-
efficient Zone clustering (EZone), FBUC. Secondly, the proposed EEGBR algorithm is compared with non-fuzzy based routing al-
gorithms such as M2CM, P-SEP and Grid ID based Routing. Experimental results demonstrate that the proposed algorithm performs
superior to any other fuzzy based routing, non-fuzzy based routing algorithms.
The experiment is carried out for more than 1000 rounds with 100 nodes. Table 8, presents the quantity of live nodes for diverse
number of rounds compared with both fuzzy and non-fuzzy based algorithms. From Table 8, it is observed that the proposed al-
gorithm performs superior to the other algorithm. The proposed algorithm gives better performance when it is compared with other
fuzzy and non-fuzzy based routing algorithms. In fuzzy based algorithms after 1600 rounds, FBUC protocol has only 31 nodes are
alive. In EZone 39 nodes, in GBRR 45 nodes, in fuzzy-TOPSIS 48 nodes and in MCFL 60 nodes are alive. But in the proposed EEGBR
algorithm, 65 nodes are alive which is higher than all the other algorithms. The main reason for this performance improvement is due
to the fact that the proposed approach for GC election and relay node selection, the residual energy of the nodes and the distance to
the sink are considered. Thus the energy is balanced among the grids leading to increase in life time of nodes. Similarly in non-fuzzy
based algorithms, the proposed algorithm has more number of nodes alive when compared with other algorithms. The reason behind
this is the proposed algorithm uses fuzzy rules for decision making efficiently. Hence, for long period of time, the number of alive
nodes is more in the proposed work when compared with both fuzzy and non-fuzzy based algorithms.
The system or network lifetime is the time duration until the first node dies (FND) [30] has been examined with various fuzzy and
non-fuzzy algorithms and the outcomes are displayed in Table 9. The system lifetime increases, when the quantity of nodes increases
in all the algorithms. It is clearly identified from the figure that the proposed algorithm performs better than the other algorithms in
terms of network lifetime.
The reason for this performance improvement is that in the proposed algorithm, fuzzy logic is used for effective decision making
in the GC election and relay GC selection and hence it efficiently handles the uncertainty by considering the residual energy and

Table 6
Fuzzy variable ranges for different inputs.
1. Distance_Sink 2. Residual energy 3. GC_Degree

Fuzzy variable Input range Fuzzy variable Input range Fuzzy variable Input range

Nearer 0-40 Lower 0.0–0.35 Low 0.0–0.5

Medium_distance 35-60 Medium_level 0.15–0.55 Medium 0.1–0.9
Farther 50-150 Higher 0.35–1 Higher 0.5–1

71
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Table 7
Simulation parameters.
Parameter Value

Area 200x200 m2
Sensor nodes 100 – 1000
Initial energy 0.5J
Basic routing algorithm LEACH
MAC layer protocol TDMA
No. of grids Static
Packet size 4000 bits

Table 8
Comparison of number of alive nodes in fuzzy based algorithms and non-fuzzy based algorithms.
No. of rounds No. of live nodes in various methods Non fuzzy methods

EEGBR MCFL Fuzzy topsis GBRR EZone FBUC M2CM P-SEP Grid ID based routing

200 100 100 100 100 100 100 100 100 100
400 100 98 100 100 100 97 100 100 92
600 100 95 95 92 90 86 92 91 87
800 97 91 88 83 78 73 88 89 79
1000 92 82 75 73 62 60 75 84 71
1200 85 76 70 62 56 55 70 74 61
1400 78 66 52 51 49 45 62 63 55
1600 65 60 48 45 39 31 53 50 42

Table 9
Comparison of network lifetime between fuzzy based algorithms and non-fuzzy based algorithms.
Network size Network lifetime in rounds

Fuzzy based methods Non fuzzy based methods

EEGBR MCFL Fuzzy topsis GBRR EZone FBUC M2CM P-SEP Grid ID based routing

200 720 650 641 630 602 500 681 654 640
400 800 720 705 700 656 578 732 726 723
600 820 785 700 700 700 700 800 758 731
800 880 820 745 732 719 717 824 812 790
1000 900 880 866 838 772 747 864 801 804

distance to the sink. This results in conservation of more energy and prolongs the network lifetime in the proposed work than the
other algorithms.
In Table 10, the average remaining energy with respect to the number of rounds is exhibited for fuzzy and non-fuzzy based
algorithms. It can be seen from this table that our proposed algorithm has more leftover energy than the other fuzzy and non-fuzzy
algorithms. For the GC election, the residual energy and the distance to the sink are considered. Since transmission consumes more
energy, the relay GC is selected by considering distance to the sink and the GC degree in addition to the residual energy. Because of

Table 10
Comparison of average residual energy for 200 nodes with fuzzy based algorithms and non-fuzzy based algorithms.
No. of rounds Average residual energy for 200 nodes (in Joules)

Fuzzy based methods Non fuzzy methods

EEGBR MCFL Fuzzy topsis GBRR EZone FBUC M2CM P-SEP Grid ID based routing

200 0.43 0.42 0.41 0.40 0.40 0.37 0.42 0.40 0.35
400 0.38 0.38 0.36 0.33 0.33 0.26 0.37 0.33 0.25
600 0.30 0.29 0.25 0.20 0.20 0.20 0.27 0.26 0.20
800 0.27 0.25 0.20 0.19 0.19 0.18 0.25 0.23 0.175
1000 0.22 0.19 0.20 0.18 0.18 0.17 0.20 0.18 0.17
1200 0.21 0.18 0.18 0.17 0.17 0.14 0.19 0.17 0.16
1400 0.20 0.17 0.17 0.13 0.13 0.10 0.18 0.165 0.155
1600 0.20 0.15 0.14 0.10 0.10 0.08 0.18 0.16 0.15

72
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Fig. 9. Comparison of FND and LND with (a) Fuzzy based algorithms and (b) Non-Fuzzy based algorithms.

this more energy is conserved in GC election phase and routing phase in the proposed work. This results in the consumption of energy
uniformly by the individual nodes and this model provides nodes with high residual energy than the nodes considered in related
works. In the proposed EEGBR, only 60.48% of energy is consumed for 1600 rounds, but for other fuzzy based algorithms MCFL
70.32%, for Fuzzy-TOPSIS 72.16%, for GBRR 74.64%, for EZone79.44 and for FBUC 83.28% of energy is utilized. Similarly, non-
fuzzy based algorithms like M2CM, P-SEP and Grid ID based routing consumed 63.26%, 68.64% and 70.52% of energy respectively.
Two measurements namely First Node Dies (FND) and Last Node Dies (LND) are utilized in the proposed work for assessment.
FND is the time period computed from the beginning of the experiment where the system starts its operation. Moreover, LND is the
time taken between system start time and the time at which the last node dies. Fig. 9(a) and (b)depict the FND and LND graphically
for fuzzy and non-fuzzy based algorithms. From Fig. 9(a), it can be observed that EEGBR provides better performance in terms of
number of rounds than other fuzzy based algorithms. As indicated by FND metric, EEGBR is more effective than MCFL about 3.7%,
Fuzzy-Topsis about 14%, GBRR about 28.5%, EZone about 32.8% and FBUC is about 48.5%. Fig. 9(b) shows that EEGBR algorithm is
more effective than M2CM about 22.2%, P-SEP is about 25.2% and Grid ID based routing algorithm is about 28.5%.
In this model, LND metric is viewed as the execution of EEGBR is more prominent than MCFL about 1.2%, Fuzzy-Topsis about
9.1%, GBRR about 17.1%, EZone about 20% and FBUC is about 27%. In addition, EEGBR algorithm is more effective than M2CM
about 33.3%, P-SEP is about 39.9% and Grid ID based routing algorithm is about 50.1%. It is observed from the graph that more
energy is conserved for more number of rounds in the proposed work. This is on account of energy equalization in the grid by
choosing the GC considering the leftover energy and distance to the sink. Moreover, in data transmission phase, the proposed work
finds the way to the sink by selecting the node which is energy proficient, nearest to the sink and has low degree. Hence in the GC
election phase, the grid-based routing phase energy is conserved in the proposed work. The leftover energy of the sensor nodes just
before the first node dies is utilized in the experiment for the assessment of the proposed algorithm with both fuzzy and non-fuzzy
based algorithms. The system was carried out with 100 nodes.
Fig. 10(a) and (b) demonstrates the residual energy of all the nodes just before the FND. The figure clearly indicates that the
leftover energy level of nodes in EEGBR are uniform than other existing algorithms. In the MCFL routing algorithm only 81% of
energy is leftover, in FBUC only 60% of energy is leftover, in GBRR 75% of energy is leftover, in EZone 71% of energy is leftover, in
M2CM 88% of energy is leftover, in P-SEP only 80% of energy is leftover, in Grid ID based routing algorithm 77.9% of energy is
leftover and in the proposed work 93% of energy is leftover. In the proposed work for GC election and for routing, the sensor node is
selected by considering the residual energy. Moreover, sensor nodes consume more energy for transmission than receiving. In the
proposed scheme, distance for transmission of data is considered for energy conservation. The selection of GC and relay node in the
proposed work clearly results in consumption of energy uniformly among the members in the grid. This increases the system lifetime.
Fig. 11(a) and (b) show the amount of data packets received by the sink from 100 nodes. From the Fig. 11, it is clearly observed
that the proposed EEGBR algorithm improves the amount of data packets received effectively compared with fuzzy and non-fuzzy
based routing algorithms. The amount of Packet Delivery Rate (PDR) in the proposed EEGBR is improved by 8%, 20.4%, 34.6%,
33.2% and 36.4% compared with MCFL, Fuzzy-Topsis, GBRR, EZone and FBUC respectively. In addition, EEGBR improves the PDRby
12%, 46.8% and 72.4% compared with M2CM, P-SEP and Grid Id based routing respectively. The reason for this is that the proposed
EEGBR algorithm uses fuzzy rules and grid-based routing which extends the network lifetime.

5. Conclusion and future work

In the proposed Energy-Efficient Grid-Based Routing algorithm, each grid chooses its grid coordinator considering the leftover
energy of the nodes within the grid and separation of the nodes to the sink by utilizing fuzzy logic. Therefore, the energy consumption
is reduced at the sensor nodes. Since the information transmission from source to the sink is performed by the grid coordinator which
acts as a relay node, it minimizes the energy consumption in the routing process which has improved the lifetime of the network. In
this scenario, the design is made in such a way that the problem of energy consumption in majority of the sensor nodes is handled
more effectively in order to avoid the wastage of energy during transmission. In such a scenario, clustering techniques were used in

73
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

Fig. 10. Comparison of Residual energy before FND for 100 nodes with (a) Fuzzy based algorithms and (b) Non-Fuzzy based algorithms.

Fig. 11. Amount of data packets received in (a) Fuzzy based algorithms and (b) Non-Fuzzy based algorithms.

74
R. Logambigai et al. Computers and Electrical Engineering 68 (2018) 62–75

this work in order to perform effective routing through cluster head nodes which helped to save the energy present in the sensor
nodes. In addition, fuzzy rules are applied in this work to perform effective decision making with respect to node deployment, energy
analysis, cluster formation, cluster head selection and routing through the cluster heads. The relay node is chosen in light of the
remaining energy, the separation to the sink and the degree of the node. The main advantage of the proposed algorithm is that it
conserves more energy and it expands the lifetime of the networking system. Through simulations carried out using network si-
mulator for sensor networks, the proposed algorithm has been assessed and it is found to be more efficient than the other related
works in terms of increase in energy efficiency, network lifetime and decrease in delay. Future works in this grid-based routing
approach is that it is possible to extend this algorithm for sensor networks with mobile nodes in order to suit the mobile applications.

References

[1] Heinzelman WR, Chandrakasan A, Balakrishnan H. Energy-efficient communication protocol for wireless Microsensor networks. Proc. IEEE 33rd annual Hawaii
international conference on system sciences. 2000. p. 1–10.
[2] Younis O, Fahmy S. HEED: a hybrid, energy-efficient, distributed clustering approach for ad hoc sensor networks. IEEE Trans Mobile Comput 2004;3(4):366–79.
[3] Bhakare R, Ketki RKK, Bhakare S. An energy-efficient grid-based clustering topology for a wireless sensor network. Int J Comput Appl 2012;39(14):24–8.
[4] Zhuang Y, Pan J, Wu G. Energy-optimal grid-based clustering in wireless microsensor networks. 29th IEEE International Conference on Distributed Computing
Systems Workshops. 2009. p. 96–102.
[5] Song X, Zhang Q, Sun W, Wei W. Energy-efficient data gathering protocol in unequal clustered WSN utilizing fuzzy multiple criteria decision making. J Intell
Fuzzy Syst 2017;32(5):3461–73.
[6] Bagci H, Yazici A. An energy aware fuzzy approach to unequal clustering in wireless sensor networks. Appl Soft Comput 2013;13(4):1741–9.
[7] Logambigai R, Kannan A. Fuzzy logic based unequal clustering for wireless sensor networks. Wireless Netw 2016;22(3):945–57.
[8] Hu X, Liu Z. Energy-efficient multi-mode clusters maintenance (M2CM) for hierarchical wireless sensor networks. China Commun 2017;14(6):1–12.
[9] Naranjo PGV, Shojafar M, Mostafaei H, Pooranian Z, Baccarelli E. P-SEP: a prolong stable election routing algorithm for energy-limited heterogeneous fog-
supported wireless sensor networks. J Supercomput 2017;73(2):733–55.
[10] Khan BM, Bilal R, Young R. “Fuzzy-TOPSIS based cluster head selection in mobile wireless sensor networks”. J Electr Syst Inf Technol 2017:1–16.
[11] Shokouhifar M, Jalali A. Optimized sugeno fuzzy clustering algorithm for wireless sensor networks. Eng Appl Artif Intell 2017;60:16–25.
[12] Mirzaie M, Azinan SMM. MCFL: an energy-efficient multi-clustering algorithm using fuzzy logic in wireless sensor network. Wireless Netw 2017:1–16.
[13] Hai DT, Le Vinh T. Novel fuzzy clustering scheme for 3D wireless sensor networks. Appl Soft Comput 2017;54:141–9.
[14] Chi Y-P, Chang H-P. An energy-aware grid-based routing scheme for wireless sensor networks. Telecommun Syst 2013;54(4):405–15.
[15] Han G, Qian A, Jiang J, Sun N, Liu L. A grid-based joint routing and charging algorithm for industrial wireless rechargeable sensor networks. Comput Netw
2016;101:19–28.
[16] Preetha T, Kevin AW. Topology control of tactical wireless sensor networks using energy-efficient zone routing. Digital Commun Netw 2016;2:1–14.
[17] Meng X, Shi X, ZiWang SW, Li C. A grid-based reliable routing protocol for wireless sensor networks with randomly distributed clusters. Ad Hoc Netw
2016;51:47–61.
[18] RejinaParvin J, Vasanthanayaki C. Particle swarm optimization-based clustering by preventing residual nodes in wireless sensor networks. IEEE Sens J
2015;15(8):4274–624. August.
[19] Xu Z, Chen L, Chen C, Guan X. Joint Clustering and routing design for reliable and efficient data collection in large-scale wireless sensor networks. IEEE Internet
Things J 2016;3(4):520–32. August.
[20] Nayak P, Devulapalli A. A fuzzy logic-based clustering algorithm for WSN to extend the network lifetime. IEEE Sens J 2016;16(1):137–44. January 1.
[21] Selvi M, Logambigai R, Ganapathy S, Ramesh LS, Nehemiah HK, Kannan A. Fuzzy temporal approach for energy-efficient routing in WSN. Proceedings of the
international conference on informatics and analytics. 2016. p. 117–22.
[22] Selvi M, Velvizhy P, Ganapathy S, Nehemiah HK, Kannan A. A rule based delay constrained energy-efficient routing technique for wireless sensor networks. Clust
Comput 2017:1–10.
[23] Thakkar A, Kotecha K. A new bollinger band based energy-efficient routing for clustered wireless sensor network. Appl Soft Comput 2015;32:144–53.
[24] Abuarqoub A, Hammoudeh M, Adebisi B, Jabbar S, Bounceur A, Al-Bashar H. Dynamic clustering and management of mobile wireless sensor networks. Comput
Netw 2017;117:62–75.
[25] Muthurajkumar S, Ganapathy S, Vijayalakshmi M, Kannan A. An intelligent secured and energy- efficient routing algorithm for MANETs. Wireless Pers Commun
2017;96(2):1753–69.
[26] Yuan X, Elhoseny M, El-Minir HK, Riad AM. A genetic algorithm-based, dynamic clustering method towards improved WSN longevity. J Netw Syst Manag
2017;25(1):21–46.
[27] Mir ZH, Young-Bae K. Collaborative topology control for many-to-one communications in wireless sensor networks. IEEE Access 2017;Vol.5:15927–41.
[28] Naeem M, Patwary M, Abdel-Maguid M. Universal and dynamic clustering scheme for energy constrained cooperative wireless sensor networks. IEEE Access
2017;5:12318–37.
[29] Huang H, Zhang J, Zhang X, Yi B, Fan Q, Li F. EMGR: energy-efficient multicast geographic routing in wireless sensor networks. Comput Netw 2017;129:51–63.
[30] Khalil EA, Attea Bara'a A. Energy-aware evolutionary routing protocol for dynamic clustering of wireless sensor networks. Swarm Evolut Comput
2011;1(4):195–203.

R. Logambigai received her ME in Systems Engineering and Operation Research and PhD in the Faculty of Information and Communication Engineering, Anna
University, India. Currently, she is teaching in the College of Engineering, Guindy, Anna University, Chennai, India. Her area of interest is computer networks, wireless
sensor networks and database management system.

Sannasi Ganapathy is an Assistant Professor in the School of Computing Science and Engineering, VIT-Chennai Campus, Chennai, India. He has received his M.E and
Ph.D degrees in Computer Science and Engineering from Anna University, Chennai. He has 10 years of experience in teaching. He has published more than 50 research
papers. His areas of interest are Computer Networks and Security.

A. Kannan is a Professor in the Department of Information Science and Technology, Anna University, Chennai, India. He has completed his M.E. and Ph.D. degrees in
Computer Science and Engineering from Anna University, Chennai. He has more than 25 years of experience in teaching. He has published 400 research papers. His
areas of interest are Computer Networks, Security and Databases.

75