This paper presents energy-efficient clustering protocols for three-tier heterogeneous wireless sensor networks (WSNs) along with their applications to the systems. The performance of a new dynamic, modified, probability-based approach to finding the most energy-efficient cluster head nodes is intended to extend network lifetime. Mathematical modeling and simulation results from using MATLAB-2017b provide a comparative analysis of several heterogeneous and homogeneous WSN variants in terms of energy consumption, active versus dead nodes, and network lifetime based on first node dead (FND) and last node dead (LND). ADV-LEACH2 in a three-tier heterogeneous network performs better than homogeneous and two-tier heterogeneous networks owing to super nodes with higher energy levels.

※ The user interface design of www.jsts.org has been recently revised and updated. Please contact inter@theieie.org for any inquiries regarding paper submission.

### Journal Search

- (Department of Electronics and Communication Engineering, SRM Institute of Science and Technology, NCR Campus, Delhi-NCR Campus, Delhi-Meerut Road, Modinagar, Ghaziabad, UP, India nitindhama85@gmail.com, vinodkur1@srmist.edu.in )
- (Department of Electronics and Communication Engineering, Dr. B. R. Ambedkar National Institute of Technology, Jalandhar, PB, India vermapk@nitj.ac.in )

## 1. Introduction

Networks that contain a large number of sensor nodes (SNs) are known as wireless sensor
networks ^{[1]}. The main task of these networks is collecting data from the environment or the area
in which they are deployed. Sensor nodes are typically placed throughout the preferred
area, where they wake up, self-test, and establish dynamic communications forming
a dynamic network. These networks are designed to support a variety of applications
and have been used in a variety of domains, including scientific exploration ^{[2]}, infrastructure protection ^{[3]}, and habitat monitoring ^{[4]}.

In these applications, the accuracy of individual sensor node readings is critical.
For example, in a surveillance system ^{[5]}, the observations of SNs must be accurate to avoid misleading outcomes, missed detections,
and false alarms. Furthermore, because only a few applications are fault tolerant
to some extent, it is necessary to remove SNs with faulty observations from a network
with built-in redundancy. After that, replacing faulty SNs with good ones can significantly
expand the overall network’s performance in term of network lifetime. Military actions,
natural disasters, and health and home monitoring are some of the most important applications
for WSNs, and WSN technology is rapidly advancing nowadays.

As an effective approach to reducing energy consumption and enhancing network lifetime,
the clustering approach is normally executed in three stages: cluster setup, cluster
head (CH) selection, and data transmission. SNs form various sizes of clusters in
the selected area during the cluster setup phase ^{[6,}^{7]}.

Based on the CH selection approach, a few SNs are selected as CH nodes, and the remaining SNs behave like members of the CH node and generate clusters.

In the third stage, members of the CHs transfer information to their CHs, which then transmit complete data to a base station (BS). Two communication forms in a WSN are sensor-node-to-CH and CH-to-BS. The BS transmits the data to the control center (CC) via Internet, satellite, or a mobile communication network. The architecture of a WSN is presented in Fig. 1.

WSNs are different from ad-hoc networks such as VANETs ^{[8,}^{9]} owing to their mobility. In a homogeneous WSN, all SNs have the same hardware complexity
and energy. Pure static clustering occurs in homogeneous networks. In heterogeneous
networks, multiple types of SNs with varying energy levels are used. This paper proposes
a new, three-tier clustering approach to cluster head selection. The contributions
of the proposed approach are summarized as follows:

· design and development of a threshold-based clus-tering approach to a three-tier, heterogeneous wireless sensor network (HWSN), and

· increasing network lifetime and reducing energy con-sumption by the HWSN.

The rest of this paper is organized as follows. The second section summarizes a literature review. The third section describes the proposed threshold-based cluster head selection approach for the three-tier heterogeneous network. The performance evaluation section includes performance metrics, discusses them, and compares them with existing approaches. The last section presents the conclusion.

## 2. Related Work

Kumarawadu et al. ^{[10]} conducted a survey of clustering protocols and classified them according to CH-selection
and cluster-formation parameters. In the survey, the authors discussed design issues
and performance challenges in clustering protocols using a taxonomy of neighborhood
information-based clustering protocols, identity-based clustering approaches, biologically
encouraged clustering approaches, and probabilistic clustering protocols.

There are several clustering approaches in hierarchical systems, such as Low Energy
Adaptive Clustering Hierarchy (LEACH) ^{[11]}, the Threshold-sensitive Energy Efficient-sensor Network (TEEN) protocol ^{[12]}, the Minimum Energy Communication Network (MECN) ^{[13]}, sensor aggregate routing ^{[14]}, Hierarchical Power-Aware Routing (HPAR) ^{[15]}, and Virtual Grid Architecture (VGA) routing ^{[16]}.

PEGASIS ^{[17]} is a sequence (chain)-based routing protocol that improves on LEACH and is used to
extend the lifetime of WSNs. The impact of energy heterogeneity in hierarchically
clustered SNs is demonstrated using SEP ^{[18,}^{19]} in WSNs. The main goal is to increase the stability period of wireless clustered
sensor networks while decreasing the instability period.

Numerous variations of LEACH are presented in this section, for example, MR-LEACH
^{[20]}, LEACH-B ^{[21]}, Centralized-LEACH (LEACH-C) ^{[11]}, EE-LEACH ^{[22]}, (LEACH-DT) ^{[23]} and ID-LEACH ^{[24]}. LEACH has some drawbacks, such as being a probabilistic method that uses a random
number of cluster head choices that may result in sub-optimum CH nodes and higher
energy consumption.

Various LEACH protocol modifications have appeared in the literature, which differ
from LEACH in terms of node heterogeneity ^{[25-}^{27]}.

## 3. The Proposed Scheme

We assume that a heterogeneous network disperses normal, advanced, and super SNs (based on energy level) in an M x M area. The BS is likewise stationary and can be found either inside or outside the target region.

Phase 1: Heterogeneous SNs are deployed randomly around the $\mathrm{M}\,\mathrm{x}\,\mathrm{M}$
area. The BS’s geographic coordinates are (50, 50) or (50, 150). After that, data
packet sizes of 4000 bits are sent from the nodes to the CH. MATLAB TOOL was used
to test and analyze this technique ^{[18]}.

Phase 2: A 2D elliptical Gaussian distribution approach is used to determine the locations
of the heterogeneous sensor nodes. Because the standard deviation factor has a moderate
effect on these two factors, energy is balanced and network lifetime is extended using
this strategy ^{[27,}^{30]}. The network's Gaussian distribution ^{[28]} is given in Eq. (1):

##### (1)

$ f\left(x,y\right)=\frac{1}{2\pi \sigma _{x}\sigma _{y}}\exp -\left(\frac{\left(x-x_{0}\right)^{2}}{2\sigma _{x}^{2}}-\frac{\left(y-y_{0}\right)^{2}}{2\sigma _{y}^{2}}\right) $where $\left(x_{o},y_{o}\right)$ denotes each sensor node's original location. The standard deviations for the x and y dimensions are $\sigma _{x}$ and $\sigma _{y}$, respectively. Each node has its own unique ID with a specific position; x$_{o}$ and y$_{o}$ will have the same value: 0. Sensor nodes have 0.5, 1, or 2 J of starting energy, depending on whether they are normal, advanced, or super SNs. Because node batteries are not rechargeable, SNs die when the battery's energy is zero.

### 3.1 Energy Model

The radio energy system for energy dissipation is presented in Fig. 2 ^{[29]}. The energy consumption of the SNs totally depends on power transmission in free
space and via multipath transmit amplifier. The free space and multipath distance
variations are given in Eq. (2). The energy model shows energy consumption for $m$-bit packet transmission over distance
$d$:

##### (2)

$ E_{TX}\left(m,d\right)=\left\{\begin{array}{l} m*E_{elec}+m*\varepsilon _{fs}*d^{2},\,\,\,d\leq d_{0}\\ m*E_{elec}+m*\varepsilon _{mp}*d^{4},\,\,\,d>d_{0} \end{array}\right. $where the term $d_{0}=\sqrt{\frac{\varepsilon _{fs}}{\varepsilon _{mp}}}$ represents the threshold distance, and $E_{elec}$ indicates the energy consumed during per-bit transmission. For free space and multipath fading channel propagation models, parameters $\varepsilon _{fs}$ and $\varepsilon _{mp}$, respectively, are free space and multipath energy-dissipation parameters for the power amplifier. While receiving packets, the radio also expends energy, and the energy expended to receive an m-bit packet can be written as seen in Fig. 2.

### 3.2 Methodology

Based on prior research, the suggested technique enhances the SEP approach's clustering
method ^{[18]}, and improves LEACH ^{[29]} by using the residual energy of nodes and the distance among the CHs to redefine
a threshold formula for clustering evenly and decreasing energy loss. LEACH ^{[29]} is the first and the most common self-organizing clustering approach for WSNs. Every
SN chooses whether to be a CH or not during the setup process in LEACH. CH selection
is based on node decisions from a choice of arbitrary numbers between 0 and 1. If
the number chosen is not greater than the default threshold, T(n), the node for the
current round is considered to be the CH. We define T(n) as seen in Eq. (3):

##### (3)

$ T\left(n\right)=\left[\begin{array}{l} \frac{P_{CH}}{1-P_{CH}*\left(r_{c}\mathrm{mod}\left(\frac{1}{P_{CH}}\right)\right)}\\ 0,\textit{otherwise} \end{array}\right.,\,\,\,if\,\,n\in G $Here, $r_{c}$ is a random number between 0 and 1. Probability $P_{CH}$ is the proportion of SNs selected as a CH for data transmission, whereas G is the group of SNs that were not selected as CHs in the current round.

In this section, we design and develop an approach to enhancing the cluster head selection
procedure. The suggested ADV-LEACH2 optimizes node placement and energy distribution
among nodes by using a 2D elliptical Gaussian distribution function ^{[27]}. In this paper, the proposed ADV-LEACH2 is a three-tier, heterogeneous wireless sensor
network having normal, advanced, and super SNs [18, 27, 30]. The advanced nodes have
more initial energy, compared to the normal SNs, and the super SNs have more primary
power than the advanced SNs; m$_{\mathrm{s}}$ is the percentage of N nodes considered
super SNs, which have $\beta $ times more primary energy than the advanced SNs, and
m$_{\mathrm{a}}$ is the percentage of N nodes considered advanced SNs, which initially
have $\alpha $ times more initial power than the normal nodes; the remaining $\left(1-m_{s}-m_{a}\right)$%
of the nodes are normal. N is the number of full SNs. $E_{in}$is the initial energy
of the normal SNs; the initial energy of the super and advanced SNs are $E_{in}*\left(1+\beta
\right)$ and$~ E_{in}*\left(1+\alpha \right)$, respectively. $P_{sup}$, $P_{adv},$
and $P_{nrm}$ stand for the probability of super, advanced, and normal SNs, respectively,
being selected as a CH by using Eqs. (4) to (6).

The combination of current energy ($E_{current}$) and initial energy factor ($E_{initial}$),
as well as the distance of the current SN ($D_{current}$) from the BS are used to
design the new probability formulation. $D_{max}$ is the highest value for the node-to-BS
distance used to design the improved formulation of T(n) for CH selection. The main
purpose in designing a new formulation is to reduce the CH nodes' energy consumption.
Eqs. (7) to (9) are calculated in order to deploy the SNs the WSNs. The new formulation of T(n) is
given below for normal and advanced nodes ^{[27]}:

For normal sensor nodes (heterogeneous WSN):

##### (7)

$ \begin{array}{l} T\left(n\right)_{mnrm}\\ =\left[\begin{array}{l} \frac{P_{bnrm}*\left(u\left(\frac{E_{\textit{current}}}{E_{\textit{start}}}\right)+v\left(\frac{d_{\textit{current}}}{d_{\max }}\right)+\left(\frac{1}{d_{\textit{basestation}}}\right)\right)}{1-P_{nrm}*\left(r\mathrm{mod}\left(\frac{1}{P_{nrm}}\right)\right)}\\ 0,\textit{otherwis} \end{array}\right.,\,\,n\in G \end{array} $For advanced sensor nodes (heterogeneous WSN):

##### (8)

$ \begin{array}{l} T\left(n\right)_{madv}\\ =\left[\begin{array}{l} \frac{P_{badv}*\left(u\left(\frac{E_{\textit{current}}}{E_{\textit{start}}}\right)+v\left(\frac{d_{\textit{current}}}{d_{\max }}\right)+\left(\frac{1}{d_{\textit{basestation}}}\right)\right)}{1-P_{adv}*\left(r\mathrm{mod}\left(\frac{1}{P_{adv}}\right)\right)}\\ 0,\textit{otherwise} \end{array}\right.,\,\,n\in G \end{array} $For super sensor nodes (heterogeneous WSN):

##### (9)

$ \begin{array}{l} T\left(n\right)_{m\sup }\\ =\left[\begin{array}{l} \frac{P_{b\sup }*\left(u\left(\frac{E_{\textit{current}}}{E_{\textit{start}}}\right)+v\left(\frac{d_{\textit{current}}}{d_{\max }}\right)+\left(\frac{1}{d_{\textit{basestation}}}\right)\right)}{1-P_{\sup }*\left(r\mathrm{mod}\left(\frac{1}{P_{\sup }}\right)\right)}\\ 0,\textit{otherwise} \end{array}\right.,\,\,n\in G \end{array} $According to Eqs. (7) to (9), $P_{bnrm}=b*P_{nrm},$ $P_{badv}=$ $a\ast P_{adv},$ and $P_{bsup}=a\ast P_{sup}$ are the weights of $P_{nrm},\,\,P_{adv}$, and $P_{bsup}$, respectively. The value of $b$ is a proportional constraint of $a$ according to network size; u, v is the ratio coefficient having a value that varies from 0 to 1, and u+ v= 1.

For normal and advanced SNs, $T(n)_{mnrm}$ & $T(n)_{madv}$, and $(n)_{msup}$ are used as multiple factors for residual and initial energy of the SNs per round, as shown in Eqs. (10) to (12). Our simulations show that the CH threshold can be changed. $T\left(n\right)_{nrm1},\,\,T\left(n\right)_{adv1},$ and $T\left(n\right)_{sup1}$ can improve the lifetime of the nodes:

##### (10)

$ T\left(n\right)_{nrm1}=\left\{\begin{array}{l} T\left(n\right)_{mnrm}\times \frac{E_{re}}{E_{in}}\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} \\ $##### (11)

$ T\left(n\right)_{adv1}=\left\{\begin{array}{l} T\left(n\right)_{madv}\times \frac{E_{re}}{E_{in}}\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} \\ $##### (12)

$ T\left(n\right)_{sup1}=\left\{\begin{array}{l} T\left(n\right)_{msup}\times \frac{E_{re}}{E_{in}}\\ 0 \end{array}\right.\begin{array}{l} ifn\mathit{\int }G\\ \textit{elsewhere} \end{array} $In Eqs. (10) to (12), where$~ E_{re}~ $is the residual energy of the SNs per round, the system stops after a few rounds because the threshold value is a lot less. To fix this issue, it is essential to solve it by prolonging $T\left(n\right)_{nrm1}$, $T\left(n\right)_{adv1}$, and $T\left(n\right)_{sup1}$ with average energy $\left(E_{avg}\right)$ that increases the CH threshold ($T\left(n\right)nrm_{2},\,\,T\left(n\right)adv_{2},$ and $T\left(n\right)sup_{2}$) for all SNs. When compared to other nodes in the network, nodes with more residual energy have a far greater chance of being picked as CH nodes. The improved thresholds, $T\left(n\right)nrm_{2},$ $T\left(n\right)adv_{2}$, and $T\left(n\right)_{sup2}~ $, are given in Eqs. (13) to (15):

##### (13)

$ T\left(n\right)nrm_{2}=\left\{\begin{array}{l} T\left(n\right)nrm_{1}\times \left(E_{avg}\right)\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} \\ $##### (14)

$ T\left(n\right)adv_{2}=\left\{\begin{array}{l} T\left(n\right)adv_{1}\times \left(E_{avg}\right)\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} \\ $##### (15)

$ T\left(n\right)sup_{2}=\left\{\begin{array}{l} T\left(n\right)sup_{1}\times \left(E_{avg}\right)\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} $Another aspect that affects the CH count is the distance. The longer the distance between the SN and the BS, the more energy is required for data communication.

##### (16)

$ T\left(n\right)_{fnrm}=\left\{\begin{array}{l} T\left(n\right)nrm_{2}\times \left(~ \frac{dtobs_{av}}{dtobs_{n}}\right)\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} \\ $##### (17)

$ T\left(n\right)_{fadv}=\left\{\begin{array}{l} T\left(n\right)adv_{2}\times \left(\frac{dtobs_{av}}{dtobs_{n}}\right)\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} \\ $##### (18)

$ T\left(n\right)_{fsup}=\left\{\begin{array}{l} T\left(n\right)sup_{2}\times \left(\frac{dtobs_{av}}{dtobs_{n}}\right)\\ 0 \end{array}\right.\begin{array}{l} ifn\in G\\ \textit{elsewhere} \end{array} $The fresh CH threshold, $T\left(n\right)\,,$ is presented in Eqs. (16) to (18) where$~ dtobs_{av}$ is the average distance among SNs to the BS; $~ dtobs_{n}$ is the distance among the SNs to the BS. Measure the threshold value from the formulas, $T\left(n\right)_{fsup}$, $T\left(n\right)_{fadv}$, and$~ T\left(n\right)_{fnrm}$, for the super, advanced, and normal SNs. The threshold formula supports only those SNs that have greater energy and shorter distances to the BS. That sensor node will have the opportunity to be selected as the CH for the round.

## 4. Performance Evaluation

The network terminology and the initial parameters used are in Table 1. Of the total nodes in the network, 25% were super nodes, 25% were advanced nodes, and 50% were normal SNs. The BSs were located at either (50, 50) or (50, 150).

Both algorithms were implemented using MATLAB (R2017b) on Windows 10. The line plot was also designed in MATLAB with the respective color code. The outcome shows that ADV-LEACH2 (three-tier-hetero) performed better, based on active/dead node-analysis metrics, throughput, network lifetime, residual energy, and optimized CH selection in the network. ADV-LEACH2 (two-tier-hetero) further improved the efficiency; however, ADV-LEACH2 (three-tier-hetero) performed best, compared to the other homogeneous and heterogeneous approaches.

Fig. 3 shows that after only 1600 rounds, the average energy of the ADV-LEACH2 (HOMO) approach reaches 0, whereas ADV-LEACH2 (hetero) continues till \textasciitilde{}2000 rounds. With ADV-LEACH2 (two-tier-hetero) and ADV-LEACH2 (three-tier-hetero), both algorithms had at least 20% energy remaining after 2500 rounds. This is possible because energy was saved for those SNs with more energy in ADV-LEACH2 (three-tier-hetero) so that nodes can be chosen as the CH a few more times. ADV-LEACH2 (homo) did not have any higher-energy nodes left; all nodes had equal amounts of energy, so as the SNs were selected as CHs for a few rounds, their energy was depleted faster, and the SN died. As seen in Fig. 3, ADV-LEACH2 worked better in the three-tier heterogeneous sensor network. ADV-LEACH2 (three-tier-hetero) worked better than ADV-LEACH2 (two-tier-hetero) and ADV-LEACH2 (homo).

LEACH2 (three-tier-hetero), LEACH2 (two-tier-hetero), and LEACH2 (homo) methods still had active nodes for up to 2500 rounds, as shown in Fig. 4. Note that the number of active nodes in ADV-LEACH2 (homo) dropped significantly at around the 1600$^{\mathrm{th}}$round and continued to decrease, compared to the other LEACH2 (three-tier-hetero) approaches. In contrast, the number of active nodes in LEACH2 (three-tier-hetero) was much higher than ADV-LEACH2 (homo) and ADV-LEACH2 (two-tier-hetero).

Fig. 4 (top) compares the number of dead nodes every round when both ADV-LEACH2 (three-tier-hetero) and ADV-LEACH2 (two-tier-hetero) approaches ran on the network with similar parameters. The results demonstrate that SNs could not remain active after 2000 rounds under ADV-LEACH2 (homo). However, the last node died after 2000 rounds under ADV-LEACH2 (homo), but the ADV-LEACH2 (three-tier-hetero) approach still had active nodes after 2500 rounds. ADV-LEACH2 (three-tier-hetero) and ADV-LEACH2 (two-tier-hetero) approaches had better results from the dead/active node analysis than ADV-LEACH2 (homo) and ADV-LEACH2 (two-tier-hetero) approaches.

In addition to network lifetime, performance is another metric for determining an approach's effectiveness. The efficiency of the proposed solution was verified when a CH had more data packets. In a sense, but not always, throughput relies on network lifetime. Fig. 5 clearly demonstrates that the throughput of ADV-LEACH2 (two-tier-hetero) is quite similar to the approach under ADV-LEACH2 (homo). The performance increase is attained by limiting the data transmissions by the CH. In ADV-LEACH2 (two-tier-hetero),\-\- the number of packets received at the CH was equivalent to ADV-LEACH2 (homo) till after 500 rounds. ADV-LEACH2 (three-tier-hetero) had higher average throughput, compared to the other two variants of ADV-LEACH2.

The energy efficiency of WSNs is significantly influenced by the selection of CHs. Thus, these CHs are going to die sooner. Stability from an optimum number of CHs is needed in successive rounds to obtain balanced energy consumption. The number of CHs in each round for ADV-LEACH2 (hetero and homo variants) is shown in Fig. 5 (bottom). The experiments show that in both cases, the optimal number of CHs is mismatched. In this case, better performance than the others cannot be explained by the approach used. This enhancement is based on the modified CH-selection approach that also increases the number of rounds. ADV-LEACH2 (hetero) had balanced clusters, compared to ADV-LEACH2 (homo).

By considering the network lifetime of ADV-LEACH2 (three-tier-hetero), ADV-LEACH2 (two-tier-hetero), and ADV-LEACH2 (homo), and their performance with regard to a differently positioned BS, network lifetime analysis was based on three metrics: first node dead (FND), half the nodes dead (HND), and last node dead (LND). As shown in Fig. 6, ADV-LEACH2 (three-tier-hetero) outperformed ADV-LEACH2 (two-tier-hetero) and ADV-LEACH2 (homo) in the FND metric, and it had stable performance based on the number of rounds when changing the position of the base station.

Fig. 6 (bottom) illustrates the distribution of dead nodes versus the number of rounds for every approach. The number of dead nodes in ADV-LEACH2 (three-tier-hetero) changes over the rounds more slowly than ADV-LEACH2 (homo). The comparative analysis for ADV-LEACH2 (three-tier-hetero), ADV-LEACH2 (two-tier-hetero), and ADV-LEACH2 (homo) was based on network lifetime until FND for 2000 nodes, as seen in Table 2.

##### Fig. 3. Comparative analysis of average remaining energy for homogeneous and heterogeneous sensor networks based on the number of rounds.

##### Fig. 4. Comparative analysis of the number of dead nodes (top) and active nodes (bottom) for homogeneous and heterogeneous sensor networks based on the number of rounds.

##### Fig. 5. Comparative analysis of average throughput (top) from the number of cluster heads (bottom) for homogeneous and heterogeneous sensor networks based on the number of rounds.

##### Fig. 6. Comparative analysis of network lifetime (top) and number of dead nodes (bottom) for homogeneous and heterogeneous sensor networks for different base station (BS) positions.

##### Table 1 Experiment Parameters.

##### Table 2 Comparative Analysis.

## 5. Conclusion

In this paper, the proposed approach, ADV-LEACH2 (three-tier-hetero), extended the network lifetime and reduced energy consumption. This approach increases cluster head lifetime and saves energy for data transmission. Based on this approach, cluster heads never die soon. The ADV-LEACH2 (three-tier-hetero) energy-efficient approach was presented in order to enhance cluster head selection and node distribution. In a heterogeneous WSN, the first strategy tries to choose the best CH for every cluster per round. The 2D Gaussian distribution was designed compensate for some SNs that are farther from the BS and the CHs. In the simulation, the ADV-LEACH2 (three-tier-hetero) approach performed better, compared to existing protocols based on network lifetime, the number of active/dead SNs, network energy consumption, and throughput.

### REFERENCES

## Author

Nitin Kumar received his Master’s in Electronics and Communication Engineering from Maharshi Dayanand University Rohtak, Haryana, India in 2010. He joined as an Assistant Professor in the Department of Electronics and Communication Engineering in SRM Institute of Science and Technology (SRMIST), NCR Campus, Ghaziabad in the year 2010. He registered for the Ph.D. research program in 2016, under the guidance of Dr. Vinod Kumar, Associate Professor, in the Department of Electronics and Communication Engineering, SRMIST. He has published more than ten research articles in refereed indexed journals and conferences. His interested areas of research are Wireless Sensor Network, Communication Network Protocols, and Network Security.

Vinod Kumar earned his M. Tech and a Ph.D. degree in Electronics Engineering from the Indian Institute of Technology, BHU, Varanasi, India, in 2008 and 2015. He has more than 15 years of experience in teaching, research, and administration. He is currently working as an Associate Professor at SRM IST Delhi-NCR Campus, Ghaziabad (U.P.). His areas of interest include MOS Sensors, WSN, GaN HEMT, Radio over Fiber, etc. He has published more than 40 research papers in international journals and conferences like IEEE Sensors, IEEE Transactions, JAP, Elsevier, Springer, etc.

Pawan Kumar Verma received his B.E. degree in Electronics and Communication from Agra University, India, and M. Tech. Degree in VLSI Design from the C-DAC, Mohali, India, in 2005 and 2009, respectively. He has worked as a consultant for 2 years at Cadence Design Systems, Noida, India. He was a Visiting Research Scholar at the University of Waterloo, Canada, from April 2012 to October 2012. He has completed his Ph.D. from the Motilal Nehru National Institute of Technology, Allahabad, India, in December 2016. Currently, he is working as an Assistant Professor at the National Institute of Technology, Jalandhar, India. His main research interests are M2M communications, MANETs, VANETs, wireless networks, and mobile computing.