Measurement: Sensors 36 (2024) 101300

Contents lists available at ScienceDirect

Measurement: Sensors
journal homepage: www.sciencedirect.com/journal/measurement-sensors

Enhancing Wireless Sensor Network lifetime through hierarchical

chain-based routing and horizontal network partitioning techniques
V.Rama Krishna * , Vuppala Sukanya, Mohd Abdul Hameed
Department of Computer Science and Engineering, Osmania University Hyderabad, India

A R T I C L E I N F O A B S T R A C T

Keywords: The efficient use of energy is a critical aspect in Wireless Sensor Networks (WSNs) to prolong the network’s
Wireless sensor networks lifespan. Investigating routing in WSNs requires considering the effective establishment of chains or clustering
Energy efficiency methods to optimize the network. The proposed routing technique aims to enhance sensor longevity through
Network life time
different network partitioning methods. The strategy employs the PEGASIS (Power Efficient Gathering in Sensor
Residual energy horizontal network
partitioning and minimum spanning tree
Information Systems) protocol, which uses Prim’s Algorithm to adjust the chain structure and is based on hi­
erarchical chain-based routing. To transmit data from working nodes to the base station (BS), horizontal network
partitioning techniques called EEPEG-PA-H and EEPEG-PA-V are utilized. The transition occurs when a node’s
remaining energy is low. This method has the potential to significantly increase the average network lifespan
compared to existing routing methods. For example, EEPEG-PA-H improves longevity by 22.7763 % compared to
PEGASIS and by 1.259 % compared to EEPEG-PA across various network sizes using senor.

1. Introduction The design of existing routing protocols for WSNs has predominantly
been driven by the dual objectives of minimizing energy consumption
Wireless Sensor Networks (WSNs) play a crucial role in enabling the and maximizing network lifespan. However, these designs frequently
Internet of Things (IoT) to be applied across various sectors such as fail to strike an effective balance with the network’s scalability and
environmental monitoring, urban development, agriculture, healthcare, reliability demands. For example, flat routing strategies, despite their
and industrial automation. These networks consist of numerous inde­ simplicity, can precipitate rapid energy depletion in nodes engaged in
pendent sensors spread out in different locations to monitor a range of long-distance transmission activities [5]. Similarly, traditional clus­
physical and environmental factors like temperature, sound, and pres­ tering approaches enhance scalability but can lead to disproportionate
sure. Working together, these sensors send the data they collect to a energy consumption among nodes, precipitating premature network
central hub for analysis and processing. Despite their usefulness, WSNs partitioning and a reduction in the overall network lifespan. Further­
face challenges, especially in terms of energy usage, which requires the more, many of these protocols do not adequately address the dynamic
development of efficient communication protocols to extend their nature of WSNs, wherein sensor nodes may experience failures or
operational lifespan [1]. environmental conditions may shift, necessitating the adoption of
Due to the inherent limitations of resources such as battery life, adaptive and resilient routing mechanisms [6].
computational power, and memory, Wireless Sensor Networks (WSNs) A WSN is characterized by its infrastructure, which encompasses
primarily consume energy for communication purposes, surpassing the small, locally distributed sensor nodes (SNs) that gather environmental
energy used for data processing [2]. Therefore, the development of data and convey it to a base station (BS) positioned at a predetermined
energy-efficient routing protocols is essential to prolong the network’s location. These SNs collaborate to fulfill objectives typically specific to a
lifespan. While early techniques like direct communication, minimum given application, equipped with sensing, processing, and communica­
transmission energy strategies [3], and basic clustering mechanisms tion capabilities. Owing to their compact, battery-powered, and cost-
have laid the groundwork for addressing these issues, they may fall short efficient design, these SNs are ideally suited for a diverse range of ap­
in extensive deployment scenarios or settings demanding prolonged plications. The components of a sensor node are illustrated in Fig. 1 [7].
network functionality [4]. The sensor nodes are strategically placed in specific locations on land

* Corresponding author.
E-mail address: vrama10001@gmail.com (V.Rama Krishna).

https://doi.org/10.1016/j.measen.2024.101300
Received 25 May 2024; Received in revised form 9 August 2024; Accepted 5 September 2024
Available online 16 September 2024
2665-9174/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-
nc/4.0/).
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

6. Adapt to Geographical Variations: Tailor clustering and resource

use to environmental conditions within layers.
7. Support Application-Specific Needs: Facilitate detailed environ­
mental monitoring, precise agricultural management, and targeted
urban sensing.

This approach aims to improve the overall performance, reliability,

and sustainability of WSNs across various applications.
The key contributions of this research paper are delineated as
follows:

1. To develop a routing strategy that combines the advantages of tree-

based and chain-based architectures, aiming to significantly decrease
both transmission distances and chain lengths, thereby optimizing
network efficiency.
Fig. 1. A typical architecture of a Sensor node.
2. To introduce a method that markedly lowers network latency by
strategically employing clusters operating in parallel, thereby
to collect important data, which is then sent to a central base station. accelerating data processing and transmission to enhance network
The data can be sent to the base station using either a direct route or by performance.
passing through multiple nodes, as explained in Ref. [8]. These nodes 3. To develop and implement a novel network partitioning technique
use wireless technology to transmit data and are powered by small, for optimal load balancing, addressing key network management
non-rechargeable batteries that cannot be replaced. Due to these limi­ challenges by ensuring a more equitable distribution of resources
tations, it is crucial to create effective plans that make efficient use of and enhancing system resilience and reliability.
these resources to achieve the network’s main goals for a long period.
Sensor nodes consume a significant amount of energy during data The manuscript is structured into four main sections, detailed as
transmission in network operations, emphasizing the importance of follows: The initial sections focus on the groundwork of related litera­
energy efficiency and network longevity. Implementing well-designed ture, energy models, and radio models. Specifically, Section 1 elaborates
routing protocols is suggested as an effective approach to address on pertinent previous studies, setting the stage for the subsequent dis­
these issues. Hierarchical routing algorithms, such as chain-based, cussion. Section 2 examines the radio model in depth and introduces the
cluster-based, or hybrid methods, have been proposed to enhance both proposed horizontal schemes of the EEPEG-PA. Following this, Section 3
energy efficiency and network lifespan. In cluster-based routing, sensor is dedicated to the simulation and comprehensive analysis of the results
nodes are grouped into clusters, with a Cluster Head (CH) responsible for derived from these schemes. The paper culminates in Section 4, which
collecting and forwarding cluster data to the sink. On the other hand, provides a conclusion, summarizing the key findings and contributions
chain-based routing involves a leader node transmitting chain data to of the study.
the sink, with subsequent nodes relaying data through the chain to the
leader node. This study introduces power-aware routing techniques 2. Related work
specifically designed for wireless sensor networks (WSNs), integrating
clustering, data aggregation, and multi-hop transmission strategies to In the quest to optimize energy efficiency within Wireless Sensor
overcome the energy constraints of sensor nodes [9]. Networks (WSNs), significant scholarly attention has been directed to­
Notably, the energy efficiency of sensor nodes can be compromised wards the development of hierarchical protocols and chain-structured
by frequent alternations between active and sleep states. This phe­ algorithms. These strategies are fundamentally aimed at addressing
nomenon is often precipitated by the practice of selecting active nodes the critical limitations imposed by the finite energy resources of WSNs,
based on superior residual energy levels in successive rounds, a process which considerably restrict their operational lifespan. This section
that can lead to expedited energy depletion. To counteract this chal­ critically examines seminal contributions and pivotal advancements in
lenge, the implementation of Sleep Scheduling and Tree-Based Clus­ this field, thereby contextualizing the proposed innovations in hierar­
tering (SSTBC) routing protocols has been suggested in Refs. [10,11]. chical chain-based routing and network partitioning techniques.
Despite the potential of these protocols to conserve energy, the opera­
tional cycle of constant activation and dormancy still contributes to 2.1. Hierarchical protocols
substantial energy consumption [12].
The primary aim of the proposed technique in Wireless Sensor Net­ Hierarchical protocols play a crucial role in improving energy effi­
works (WSNs) is to enhance network efficiency, reliability, and scal­ ciency and prolonging network lifetime within Wireless Sensor Net­
ability by dividing the network into manageable horizontal layers. Key works (WSNs). These protocols are known for their hierarchical
objectives include: structure, offering effective and scalable network management solu­
tions. One notable protocol, the Low Energy Adaptive Clustering Hier­
1. Optimize Energy Consumption: Reduce energy usage through archy (LEACH) [14], stands out for its cluster-based approach, where
localized communication and balanced load distribution, extending cluster heads are elected to manage data aggregation and transmission
node and network lifespan. to the base station. Despite LEACH’s energy-saving capabilities, its
2. Improve Data Management: Enable efficient data aggregation and random selection of cluster heads may lead to uneven energy distribu­
localized processing within layers, minimizing redundant trans­ tion across the network. To address this issue, enhancements like
missions and conserving energy. LEACH-C [15] have been introduced, utilizing a centralized algorithm
3. Enhance Scalability: Create manageable segments for easier for cluster head selection to promote more balanced energy usage
administration and seamless network expansion. among nodes. By maintaining a hierarchical network structure, these
4. Increase Fault Tolerance: Isolate failures within layers and provide protocols not only enhance scalability and reduce energy consumption
redundant paths, ensuring continuous operation. but also optimize the network’s longevity [6,16].
5. Simplify Routing: Use layer-based routing to reduce complexity, LEACH (Low-Energy Adaptive Clustering Hierarchy): LEACH is
overhead, and latency. one of the earliest and most well-known clustering protocols designed to

2
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

minimize energy consumption in WSNs. The latest literature highlights energy consumption. These algorithms often utilize machine
several enhancements to LEACH: learning to predict optimal paths for data transmission.
• Dynamic Chain Reconfiguration: Dynamic chain adjustment
1. Machine Learning Integration: mechanisms have been proposed to adapt to changes in network
• Predictive Algorithms: Recent studies integrate machine learning topology and node energy levels, ensuring balanced energy con­
techniques to predict optimal cluster heads based on historical sumption and prolonged network lifetime.
data and real-time metrics. This reduces the likelihood of selecting 2. Fault Tolerance and Reliability:
nodes that might deplete their energy quickly. • Redundant Data Transmission: To enhance fault tolerance, multi-
• Adaptive Mechanisms: Machine learning models are employed to path strategies and redundant data transmission methods have
adapt the cluster formation process dynamically, considering the been introduced. These ensure data integrity even in the case of
network’s changing conditions and node mobility. node failures.
2. Energy Harvesting Techniques: • Adaptive Fault Tolerance Mechanisms: Adaptive mechanisms have
• Solar Energy Utilization: Researchers have explored integrating been developed to detect and respond to node failures in real-time,
solar energy harvesting into LEACH. Solar-powered nodes can act maintaining reliable communication within the network.
as cluster heads, extending the network’s lifetime significantly by
reducing reliance on battery power alone. Comparative Analysis: While LEACH and LEACH-C focus on clus­
• Energy Harvesting Models: New models have been proposed to tering to manage energy consumption, PEGASIS offers a different
optimize energy harvesting and consumption, making LEACH approach with its chain-based data transmission strategy. Each protocol
more sustainable in long-term deployments. has unique strengths and is suitable for different types of WSN
3. Security Enhancements: applications:
• Lightweight Encryption: Enhancements include lightweight
encryption algorithms to secure data transmission within clusters • LEACH and LEACH-C: Best suited for networks where nodes are
without significantly impacting energy consumption. relatively static and the overhead of cluster management is
• Trust-Based Mechanisms: Trust-based mechanisms have been manageable.
introduced to mitigate the risk of malicious nodes and ensure • PEGASIS: More effective in scenarios requiring significant data ag­
reliable communication within the network. gregation and where minimizing transmission distance is crucial.

LEACH-C (Centralized Low-Energy Adaptive Clustering Hierar­ 2.3. Network partioning technique
chy): LEACH-C improves upon LEACH by centralizing the cluster for­
mation process, which can lead to more optimized clustering. Recent Network partitioning in Wireless Sensor Networks (WSNs) is a crit­
advancements in LEACH-C include: ical technique used to enhance the efficiency, reliability, and lifespan of
the network. Partitioning can help in managing the network’s
1. Optimization Algorithms: complexity, optimizing resource usage, and ensuring better performance
• Genetic Algorithms and Swarm Optimization: Advanced optimi­ under various conditions.
zation techniques such as genetic algorithms and particle swarm Horizontal partitioning in Wireless Sensor Networks (WSNs) offers
optimization have been used to enhance the central cluster for­ numerous advantages:
mation process. These techniques help in achieving a more
balanced load distribution across the network. 1. Enhanced Energy Efficiency: Localized communication within
• Hybrid Approaches: Hybrid LEACH-C protocols combine central­ layers reduces energy consumption. Balanced load distribution pre­
ized and decentralized methods to benefit from both approaches, vents overburdening individual nodes, extending network lifespan.
improving overall efficiency and robustness. 2. Improved Data Management: Efficient data aggregation within
2. Improved Energy Models: layers minimizes redundant transmissions and lowers network
• Accurate Energy Consumption Models: Researchers have devel­ traffic. Localized processing reduces the need for constant data relay
oped more accurate energy consumption models that account for to the central base station, conserving energy.
node mobility, environmental factors, and varying energy har­ 3. Better Scalability: The network can be divided into manageable
vesting rates, leading to better performance and longer network segments, simplifying expansion and maintenance. New layers can
lifespans. be added as the network grows without significantly altering existing
3. Scalability and Adaptability: structures.
• Scalable Architectures: New scalable architectures for LEACH-C 4. Increased Fault Tolerance: Failures can be contained within a
have been proposed to handle larger networks efficiently. These single layer, preventing widespread network disruption. Redundant
architectures ensure that the central controller does not become a paths ensure continuous operation despite node or link failures.
bottleneck as the network size increases. 5. Simplified Routing: Layer-based routing reduces complexity and
overhead, leading to more efficient data transmission and lower
2.2. Chain-structured algorithms latency.
6. Geographical Adaptation: Tailored clustering and optimized
Exploring the realm of hierarchical protocols, innovative algorithms resource use within layers enhance network performance, especially
like PEGASIS (Power-Efficient GAthering in Sensor Information Sys­ in varying environmental conditions.
tems) have brought forth a new strategy where nodes create a chain 7. Application-Specific Benefits: Enables detailed environmental
[17]. PEGASIS differs from LEACH and LEACH-C by organizing nodes monitoring, precise agricultural management, and targeted urban
into chains rather than clusters, which can significantly reduce the sensing, enhancing overall network effectiveness and sustainability.
overall transmission distance. Recent studies on PEGASIS focus on the
following improvements: These advantages make horizontal partitioning a valuable technique
for optimizing WSN performance and longevity [13,17,22,25–27].
1. Chain Formation Algorithms:
• Efficient Algorithms: New algorithms have been developed to form
more efficient chains, reducing the transmission distance and

3
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

Fig. 2 illustrates the first-order radio model, providing a visual rep­

resentation of the energy consumption mechanisms at play during data
transmission and reception within WSNs. The mathematical formula­
tions derived from this model serve as the foundation for analyzing the
energy efficiency of the proposed routing methodologies, ensuring that
our strategies are grounded in realistic operational parameters. Energy
is required to operate the receiver and transmitter circuits. Here is the
equation for the total amount of bits (k).
ЕTx/Rx− elec = Еelec × k (1)

HereEelec is energy exhausted pr bit to function transmit or receive.

During transmission, the energy spent by the amplifiers is dependent on
both the transmission distance (∂) and the number of bits being
broadcast.
Fig. 2. First order radio model. {
εfs × k × ∂2 , if ∂ < d0
ЕTx− amp = (2)
2.4. Network and radio model εmp × k × ∂4 , if ∂ > d0

where“d0” is the should given as

Prior to discussing the proposed routing strategies and their empir­
√̅̅̅̅̅̅̅
ical validations, it is crucial to establish the fundamental network and εfs
d0 = (3)
radio models that serve as the foundation for our simulations and εmp
theoretical assumptions in Wireless Sensor Networks (WSNs). These
models define the operational constraints and capabilities of WSNs, The constants εfs and εmp determine the fading model utilized, either
playing a key role in accurately evaluating energy efficiency and the free-space or multipath model. This is illustrated in Fig. 2.
network performance. The key assumptions and models are as follows: Then the total energy essential for communication [18] is:
Base Station (BS): The BS is stationary, located at a distance from the {
Еelec × k + εfs × k × ∂2 , if ∂ < d0
sensing area, equipped with sufficient processing power, and an un­ ЕTx = (4)
Еelec × k + εmp × k × ∂4 , if ∂ > d0
limited energy supply.
Energy Constraints: All nodes start with the same energy levels, Reception energy is comprised of the energy necessary to run the
emphasizing the uniform distribution of initial energy reserves across circuitry and transmit the signal.
the network. ЕRx = Еelec × k (5)
Node Mobility: Once deployed, nodes remain stationary, ensuring
that mobility does not affect network dynamics and energy consumption hence energy required for data aggregation is
patterns.
Sensing Approach: The network adopts a “Time-Driven” sensing ЕDA− tot = s × k × Еda (6)
mechanism, where sensors continuously monitor their surroundings and Here, “s" stands for the number of signals actuality aggregated, while
transmit data to the BS, showcasing the network’s continuous opera­ “EDA” is for the energy used per bit during aggregation.
tional mode.
Energy Usage in Transmission: Data transmission consumes the most
3. Methodology
energy for a node, with energy consumption levels directly linked to the
transmission distance.
The proposed algorithm is articulated through a comprehensive
Node Placement: Nodes are randomly dispersed throughout the
methodology, systematically divided into four distinct phases: setup,
network coverage area, mirroring common deployment scenarios in
chain leader selection, minimum spanning tree (MST) formation, and
practical applications [28–32].
data transmission. Each phase plays a crucial role in optimizing the
energy efficiency and enhancing the longevity of Wireless Sensor Net­
works (WSNs).

3.1. Set-up phase

3.1.1. Area partioning

The proposed algorithm begins with a network-wide broadcast,
where each sensor node sends a HELLO packet to the Base Station (BS)
containing essential identifiers like the node ID and energy level [19].
This enables a detailed energy assessment of the network. The BS uses
localization methods, particularly the Received Signal Strength Indica­
tor (RSSI) [20], to determine the exact locations of the nodes, laying the
groundwork for cluster creation. After pinpointing the nodes’ spatial
layout, the BS starts the clustering process, which involves two strate­
gies: horizontal and vertical clustering based on y-coordinates align­
ment. This approach allows for flexible cluster formation to suit different
deployment scenarios. The clustering process can be mathematically
represented as:
Let N denote the set of all nodes within the network, where each node
( )
ni ∈ N broadcasts its position xi , yi and energy level Ei to the BS. The
Fig. 3. Horizontal partioning in 100 × 100 m2 network area. BS, leveraging RSSI, maps these nodes onto a coordinate system,

4
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

Fig. 4. Pseudocode of one-hop neighbour node pairing.

establishing a spatial framework for cluster formation. transfer, while those outsides are labelled “unpaired,” meaning no direct
communication is possible. The process of pairing nodes is illustrated in
3.1.1.1. Horizontal partioning. For horizontal partitioning, the parti­ Fig. 4, showing how the algorithm categorizes nodes. This method not
tioning criterion segregates nodes based on their y-coordinate values, only improves data transmission efficiency by leveraging node prox­
maintaining a constant x-coordinate across the cluster. Conversely, imity but also reduces redundant data transmission, thus conserving
vertical clustering segregates based on x-coordinate values, maintaining energy resources in the network.
a constant y-coordinate. The objective function for cluster formation, This strategic combination, based on the concepts of physical prox­
whether horizontal or vertical, seeks to minimize intra-cluster distances, imity and shared sensory functions, serves as the foundation of the
thereby optimizing the energy efficiency of intra-cluster communica­ network’s system for collecting and sharing data, enhancing both the
tions. Mathematically, this can be represented as minimizing the sum of effectiveness of operations and the sustainability of energy usage within
the squared distances between each node and the cluster centroid, which the sensor network.
is a function of either the x or y coordinates, depending on the clustering
orientation. EEPEG-PA’s horizontal clustering involves dividing the 3.2. Chain leader selection phase
network equally based on the y-coordinates of the nodes, while main­
taining a constant x-coordinate. The size of a cluster in a (h X h) m2 Within the context of each operational round, the determination of a
network would be (h X h/k) m2, here“k”denotes the clusters that are Chain Leader (CL) for every designated region is executed through a
required. Fig. 3 demonstrates the chain and cluster method of Horizontal meticulous evaluation process conducted by the Base Station (BS). This
Clustering. process employs a weighted metric, symbolized as Q, which integrally
considers two paramount factors: the residual energy of the active nodes
3.1.2. One-hop neighbour node pairing (Eresidual ) and their geographical proximity to the BS, denoted by the
In accordance with the methodologies delineated in prior studies [8, distance (d). The algorithmic formulation for Q is represented as follows,
12], and [21], the process of one-hop neighbour node pairing is predi­ where a higher Q value signifies a node’s suitability as a CL due to its
cated on the spatial positioning and sensing capabilities of the nodes optimal energy reserve and strategic location relative to the BS:
within the network. For a given node, denoted as " sii , positioned at Erssidmal (si )
( )
coordinates xsi , ysi and possessing a sensing range denoted by " Rsi ", Q(si ) = (9)
d(si , BS)
the Euclidean distance to an adjacent node “sj” is calculated using the
following formula: The node, identified as " siʹʹ, that exhibits the maximal Q value within its
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ respective region is accorded the status of CL for the ensuing commu­
( ) ( )2 ( )2
d si , sj = xsi − xsj + ysi − ysj (7) nication round. This selection criterion ensures that the chosen CL
possesses not only a sufficient energy reserve to sustain its enhanced
A node " sj " is deemed to be within the sensing range of node “si”, communicative role but also a favorable position that minimizes the
and thereby considered a neighbouring node capable of detecting the energy expenditure associated with data transmission to the BS. This
same environmental target, if the following condition is satisfied: phase is pivotal in optimizing the network’s energy utilization and
( ) maintaining the efficacy of data relay processes. By judiciously selecting
d si , sj ≤ Rsi (8) CLs based on a balanced consideration of energy and proximity, the
The study suggests using a greedy algorithm to help the Base Station network is poised to achieve a significant augmentation in its opera­
(BS) identify nearby nodes for each node in the network. This algorithm tional lifespan and efficiency.
sorts nodes into “paired” and “unpaired” categories based on set dis­
tance and sensing range thresholds. Nodes within the specified range are
considered “paired,” indicating a direct communication link for data

5
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

Fig. 5. Flow chart of EEPEG-PA-H.

3.3. Minimum spanning tree formation phase using Prim’s algorithm mathematically defined by the criterion of minimizing the distance (d),
or equivalently, the communication energy between nodes, while pro­
In this critical phase of network optimization, the Base Station (BS) gressively expanding the tree until it encompasses all active nodes
conceptualizes each region within the network as a connected undi­ within the region [23,24].
rected graph, symbolically represented as G (A,B). Here A delineates the
set of active nodes within the specified region, and B encompasses the 3.3.1. Algorithm: construct minimum spanning tree using Prim’s algorithm
set of edges, each signifying a potential link between pairs of active Input: A graph G = (A,B), where A is the set of nodes and B is the set
nodes. The objective is to construct a Minimum Spanning Tree (MST) of edges between the nodes. Each edge has an associated weight, rep­
that connects all active nodes (A) via the optimal subset of edges (Bʹ), resenting the cost (or distance) between two nodes.
minimizing the aggregate edge weight and, consequently, the total en­ Output: A minimum spanning tree T(A, Bʹ) for the graph G , where Bʹ
ergy expenditure for intra-network communications. is a subset of B that connects all nodes in A with the minimum total edge
Employing Prim’s algorithm, renowned for its efficacy in establish­ weight.
ing an MST [22], the BS initiates the process with the Chain Leader (CL) Steps:
of each region as the foundational node of the tree. The algorithm
iteratively explores the edges, selecting at each step the edge of minimal 1 Initialize:
weight that extends the tree to an unincorporated node. This selection is • Start with an empty tree T, i.e., T = ∅.

6
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

• Select the Chain Leader (CL) as the starting node and add it to the Table 1
set C.C = {CL}. Simulation parameters.
2 Tree Expansion: No. Simulation Parameter Values
• While C, the set of nodes included in T, does not contain all nodes
1. Simulation Area 100 x100 m2
in A : 2. Number of sensor nodes considered 100,160,200
• Identify the edge (a, b) with the lowest weight such that a is in C 3. Energy required for transmission and receiving 50 nJ/bit
and b is not in C (i.e., a ∈ C and ∈ A − C). (ETx&ERx)
• Add this edge to the tree T, i.e., T = T ∪ {(a, b)}. 4. Free space constant (Efs) 10 pJ/bit/m2
5. Multipath fading constant (Emp) 0.0013 pJ/bit/
• Include node b into the set C, i.e., C = C ∪ {b}. m4
3 Termination: 6. Initial Energy of nodes (En) 0.5 J
• The algorithm terminates when all nodes are included in C, 7. Packet size of HELLO-bit 30
meaning the tree T now connects all nodes in A with the minimum 8. Size of K- bit data packet 4000
9. Simulation rounds 5000
possible total edge weight.
10. Sensing range (Rs) 10 m
• The output T(A, Bʹ) represents the minimum spanning tree of the
graph G .
2 Selection: At each step, select the edge (u, v) with the minimum
End Algorithm weight such that u ∈ VT and v ∕ ∈ VT .
The algorithmic method guarantees that every node is smoothly 3 Inclusion: Add v to VT and (u, v) to ET .
incorporated into the network’s communication system, focusing on 4 Iteration: Repeat the selection and inclusion steps until all nodes are
conserving energy effectively. Specifically, the CL nodes play a crucial included in VT , i.e., VT = V
role as central points for relaying data. Fig. 5 visually displays how an
MST is created through Prim’s algorithm, showcasing the successful The weight function Q(si) for selecting CLs, as mentioned earlier,
connectivity between the operational nodes [25]. Implementing Prim’s combines the residual energy and distance considerations into a single
algorithm at this stage is fundamental for optimizing the network’s metric, which influences the structure of the MST by prioritizing nodes
structure. It facilitates the establishment of an energy-efficient routing with higher residual energy and closer proximity to the BS:
layout, ultimately improving the network’s durability and
Ersid (si )
dependability. Q(si ) = (11)
d(si , BS)
This formula guarantees that nodes with greater energy reserves and
3.4. Data transmission phase
those situated closer to the BS have a higher chance of becoming CLs.
This, in turn, influences the MST construction to prioritize shorter
3.4.1. TDMA scheduling
transmission distances, ultimately leading to reduced energy consump­
In the context of this algorithm, the Base Station (BS) assigns time
tion throughout the network.
slots to each active node within the network to prevent data collision
Through these mathematical formulations, the proposed algorithm
and ensure an orderly transmission sequence. This is realized through a
seeks to optimize the network’s energy efficiency by structurally orga­
Time Division Multiple Access (TDMA) schedule. The scheduling can be
nizing data transmission paths and meticulously scheduling communi­
represented mathematically by assigning each node i a unique time slot
cation events. Fig. 5 illustrates the flow chart diagram of proposed
Ti within a cycle C, ensuring that Ti ∕
= Tj for any i ∕
= j, where i, j are active
algorithm.
nodes in the network.
4. Simulations and analysis of results
3.4.2. Data aggregation and transmission
Each Cluster Leader (CL) aggregates data from its child nodes and To evaluate the effectiveness of the proposed routing strategy, a
then transmits this aggregated data, along with its own data, to the BS. If comprehensive series of simulations was conducted utilizing MATLAB as
we denote the set of child nodes for CLi as Si , the total data Di trans­ the computational platform. These simulations were meticulously
mitted by i to the BS can be represented as: designed to enable a direct comparative analysis with the PEGASIS
∑
D i = di + dj (10) protocol, particularly emphasizing the strategic placement of the Base
j∈Si Station (BS) at the centroid of the network to parallel the chain forma­
tion approach inherent to PEGASIS. The simulation parameters, as
where di is the data from CL i itself, and dj is the data from each child outlined in Table 1, provide a structured framework for this analysis.
node j in Si . The routing approach presented here stands out for its creative uti­
lization of paired nodes during each communication cycle, resulting in a
3.4.3. Construction of the minimum spanning tree (MST) using Prim’s noticeable decrease in energy consumption. The primary objective is to
algorithm prolong the sensor network’s lifespan by optimizing the allocation of
The MST is constructed to connect all active nodes in the network Cluster Leader (CL) roles in every round. This allocation is determined
while minimizing the total edge weight, which in this context could by evaluating the remaining energy and proximity of nodes to the BS,
represent the energy cost of transmission or the distance between nodes. with a focus on energy efficiency and network longevity. The simula­
Prim’s algorithm starts with a single node (in this case, the CL) and tions were carried out systematically using network configurations of
incrementally adds edges and nodes to the tree to maintain the MST 100, 160, and 200 nodes distributed over a 100 × 100 m2 simulation
property. area. These simulations encompassed 5000 communication rounds,
Let G = (V, E) be the graph representing the network, where V is the producing a comprehensive dataset for performance evaluation. This
set of nodes, and E is the set of edges between nodes. Each edge (u, v) ∈ E detailed assessment framework allows for a thorough investigation of
has an associated weight w(u, v), representing the cost of transmission the routing strategy’s impact on operational efficiency, particularly
between nodes u and v. concerning energy consumption and network longevity. The outcomes
of these simulations suggest the potential for significant improvements
1 Initialization: Start with T = (VT ,ET ), where VT initially contains the in network performance, challenging existing protocols and introducing
CL node, and ET is empty.

7
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

Fig. 6. (a) (b) (c)illustrates the trajectory of node mortality against communication rounds within a network of 100 nodes.

a novel approach to managing energy-efficient sensor networks. PA suite, especially its hybrid variations, which demonstrate better
longevity. This suggests that the EEPEG-PA protocols are more effective
4.1. Network lifetime and load balancing at maintaining node functionality over multiple communication cycles.
The comparison underscores the importance of using well-designed
In this study, it is assumed that every node has data to send to the protocols to improve network resilience and stability.
Base Station (BS). This assumption leads to the idea that the network’s Fig. 6(c) illustrates the interrelation of node mortality with the suc­
performance starts to decrease when nodes stop functioning. Therefore, cessive communication rounds within an expanded network framework
the “First Node Die (FND)" metric is used to measure the network’s of 200 nodes. The graph illustrates the performance comparison be­
durability. The amount of energy left after the FND is used to assess how tween the PEGASIS protocol and its advanced versions, EEPEG-PA and
well the network balances its load. If there is less energy remaining after EEPEG-PA-H, in managing network operations. The findings indicate
the FND, it indicates that the network is distributing the load evenly that PEGASIS shows a faster decline in node viability than the EEPEG-
among its nodes, as shown by the extended operational time of all nodes. PA_H suite, particularly its hybrid forms, which exhibit superior
Given an initial energy allocation of 0.5 J per node, this examination is longevity. This implies that the EEPEG-PA-H protocols are more adept at
extended across varying network dimensions, as illustrated in Fig. 6(a) sustaining node functionality across numerous communication cycles.
(b)(c). The analysis highlights the significance of employing well-crafted pro­
Fig. 6(a) presents A comparison was made on the node mortality tocols to enhance network resilience and stability.
trends in a 100-node network, analyzing the performance of different
communication protocols as communication rounds increased. The 4.2. Alive nodes
findings revealed that the PEGASIS protocol showed a sharper increase
in node failures, indicating a higher failure rate. In contrast, the EEPEG- Nodes that still have ample energy for transmission and have not
PA protocol and its hybrids, EEPEG-PA-H, displayed improved resilience completely exhausted their power reserves are classified as being
with lower mortality rates. As the mortality rates of the EEPEG-PA “alive”. The results of each communication cycle for different coverage
variations converged up to a specific round count, it suggested an effi­ regions are shown in Fig. 7(a)(b)(c) as the number of surviving nodes in
ciency threshold where the hybrid protocols began to show varying the network.
levels of endurance. This highlights the potential for optimizing pro­ Fig. 7(a) presents a decay curve for a network of 100 nodes, detailing
tocols to extend the lifespan of nodes in communication networks. the survivability of nodes over communication rounds for four distinct
Fig. 6(b) delineates the trajectory of node mortality in relation to the routing protocols: PEGASIS, EEPEG-PA and EEPEG-PA-H. Initially, all
progression of communication rounds within a network consisting of protocols commence with the full complement of nodes. The curve for
160 nodes. The graph shows how the PEGASIS protocol compares to its PEGASIS exhibits sustained node life up to a certain threshold of rounds,
more advanced versions, EEPEG-PA and EEPEG-PA-H, in terms of their after which a rapid decline in alive nodes is observed, indicating a
ability to handle network operations. The data reveals that PEGASIS precipitous energy depletion. In contrast, EEPEG-PA shows a more
experiences a quicker decrease in node viability compared to the EEPEG- gradual descent, thereby implying enhanced energy efficiency. Further

Fig. 7. (a) (b) (c) Decay curve of Alive nodes vs communication rounds for 100,160 and 200 nodes.

8
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

Fig. 8. (a) (b) (c) Decay curve of Residual Energy vs communication rounds for 100,160 and 200 nodes.

refinement is observed with EEPEG-PA-H, which marginally out­ 4.3. Residual energy
performs its predecessor in prolonging node vitality. This graph un­
derscores the efficacy of the EEPEG-PA-H method in optimizing energy The overall energy remaining in the network can be inferred from the
distribution across the network, significantly increasing network lon­ total residual energy of nodes, which is calculated by summing up the
gevitya paramount consideration in scenarios demanding persistent energies of all functioning nodes at the end of each communication
network operation. round. The results indicate that the EEPEG-PA-H protocol consumes less
Fig. 7(b) elucidates the longevity of nodes in a network scenario energy than the PEGASIS protocol and so has a greater “residual
involving 160 nodes, comparing the persistence of operational nodes energy”.
across varying communication rounds among four routing protocols: Fig. 8(a) expounds upon the decay curve of residual energy as a
PEGASIS, EEPEG-PA and EEPEG-PA-H. The commencement of the graph function of communication rounds within a network composed of 100
displays an initial plateau where all nodes are alive, indicative of the nodes. The graphical comparison shows how the PEGASIS protocol’s
beginning stages of network operation. The PEGASIS protocol maintains energy retention differs from EEPEG-PA and EEPEG-PA-H. Over multi­
node longevity initially, followed by a sharp decline, suggesting a uni­ ple communication rounds, PEGASIS shows a quicker drop in remaining
form energy usage until a critical depletion point. Conversely, EEPEG- energy, indicating faster depletion compared to the other protocols. On
PA indicates an improved endurance with a less steep decline, reflect­ the other hand, EEPEG-PA protocols demonstrate more consistent en­
ing better energy management. This trend is incrementally enhanced ergy usage, with the hybrid EEPEG-PA-H version showing gradual en­
with EEPEG-PA-H, which evidences a further deceleration in node hancements in energy conservation. This analysis highlights the
mortality rate. Significantly. This graphical representation is imperative significance of optimizing protocols to improve energy efficiency and
in evaluating the effectiveness of the EEPEG-PA-H approach in extend­ prolong the operational lifespan of wireless sensor networks.
ing the functional duration of networks, especially in applications where Fig. 8(b) delineates the decay curve of residual energy relative to
the sustained operation of sensor nodes is critical. communication rounds within a network framework encapsulating 160
Fig. 7(c) presents a decay curve delineating the survival of nodes nodes. This illustration provides a critical comparative analysis of the
within a 200-node network configuration, set against the backdrop of energy sustainability under the PEGASIS protocol vis-à-vis the EEPEG-
accumulated communication rounds, for the purpose of contrasting the PA suite, inclusive of EEPEG-PA-H. It reveals that the PEGASIS proto­
durability of nodes under the governance of four routing protocols: col exhibits a more precipitous energy decline, whereas the EEPEG-PA
PEGASIS, EEPEG-PA, and EEPEG-PA-H. The depiction showcases an strategies, particularly the hybridized versions, manifest a more con­
equal starting point for all nodes, with subsequent divergence reflecting servative energy depletion rate.
the resilience of the nodes over time. The PEGASIS protocol exhibits Fig. 8(c) articulates the decay curve of residual energy in corre­
initial node survival stability which then sharply tapers off, indicative of spondence with the number of communication rounds for a network
a collective energy depletion threshold. In comparison, EEPEG-PA constituting 200 nodes. The graph compares the energy efficiency of the
demonstrates a more graduated decline, indicative of a superior en­ PEGASIS protocol with the EEPEG-PA protocol and its hybrids, EEPEG-
ergy conservation approach. Advancements in the routing technique, as PA-H. It illustrates a quicker depletion of energy reserves with PEGASIS,
seen in EEPEG-PA-H, yield a further moderated decline in node sur­ indicating a less efficient protocol compared to the EEPEG-PA family.
vivability. This graphical analysis serves as a crucial comparative tool The EEPEG-PA protocols and their hybrid versions exhibit a more
for ascertaining the efficacy of energy utilization strategies, particularly gradual decline in energy levels, showcasing their superior energy
in scenarios where extended network longevity is of paramount management abilities. This visual representation emphasizes the
importance. importance of well-designed protocols in promoting sustainable energy
use and prolonging the operational lifespan of extensive network
systems.
The First Node Die (FND) metric is pivotal in assessing the longevity

Table 2
FND and Residual energy in network for various network sizes.
Routing Method FND for 100 nodes FND for 160 nodes FND for 200 nodes

Round number Residual energy (%) Round number Residual energy (%) Round number Residual energy (%)

PEGASIS 1135 51 % 1169 42 % 1204 41 %

EEPEG-PA 2138 42 % 2191 41 % 2329 40 %
EEPEG-PA-H 2189 40 % 2231 39 % 2411 38 %

9
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

Table 3
First 10, 50 and 100 % of nodes die in network for various network sizes.
Routing Method/% of nodes 100 nodes network 160 nodes network 200 nodes network

10 % 50 % 100 % 10 % 50 % 100 % 10 % 50 % 100 %

PEGASIS 1981 2253 2742 2143 2312 2832 2162 2342 2892
EEPEG-PA 2218 3143 4152 2389 3201 4267 2415 3230 4313
EEPEG-PA-H 2257 3359 4456 2476 3416 4517 2502 3462 4565

of a network, gauging the precise round when the initial node depletion 4.5. Future directions
occurs critical juncture indicative of the network’s overall load distri­
bution and energy equilibrium. For the refined EEPEG-PA-H routing Future research should address the scalability of the proposed
protocol, the FND benchmark was attained at round 2357 within a 100- method, possibly through the development of distributed clustering and
node network, with a commendable 39 % of energy reserves remaining, pairing mechanisms to reduce the load on the BS. Investigations into
signaling adept load balancing. In the expanded network scenarios of alternative localization techniques that could complement or substitute
160 and 200 nodes, the FND was reached at rounds 2543 and 2682, with RSSI measurements for greater accuracy in node positioning are also
residual energies of 38 % and 33 % respectively, reflecting a slightly recommended. To ensure the robustness of the proposed algorithm, it
diminished but still substantial energy conservation. would be beneficial to test its performance in a variety of real-world
Nodes are considered “dead” when they have run out of power and settings and against physical factors that can affect sensor networks.
can no longer send data. Table 2 collates the FND alongside the Additionally, expanding the simulations to encompass a wider range of
concomitant residual energy percentages across different network sizes network sizes and densities will provide a more comprehensive under­
and routing methods. The PEGASIS protocol exhibits an FND occurrence standing of the algorithm’s performance. Exploring the integration of
at round 1135 with 51 % residual energy in a 100-node network, at 1169 machine learning techniques for predictive energy management and
with 42 % for 160 nodes, and at 1204 with 41 % for 200 nodes. EEPEG- adaptive clustering could also yield significant enhancements in
PA and its hybrid, EEPEG-PA-H, manifest FND at later rounds (2138 and network lifetime and efficiency.
2189 respectively for 100 nodes) with residual energies slightly above
40 %. The node ‘death’ in this context refers to the state where nodes are 5. Conclusions
completely depleted of power, rendering them incapable of transmitting
data, which marks a critical threshold for evaluating routing protocol A novel power-awaretree-based routing protocol has been intro­
efficiency. duced in our research aimed at prolonging the lifespan of wireless sensor
Table 3 presents a systematic enumeration of the rounds at which the network nodes. The routing strategy suggested in our study involves
initial 10 %, 50 %, and the complete 100 % node mortality thresholds turning off unnecessary nodes that detect the same data to conserve
are reached across various network configurations utilizing different energy. Additionally, by preventing paired nodes from switching be­
routing methods. The tabulated data delineates the performance of tween active and sleep modes within a single communication cycle, the
PEGASIS, EEPEG-PA and EEPEG-PA-H within networks of 100, 160, and sensor network’s longevity is increased by reducing energy consumption
200 nodes. For a network of 100 nodes, the PEGASIS protocol exhibits linked to frequent mode changes. The selection of the appropriate CL has
node mortality at 1981 rounds (10 %), 2253 rounds (50 %), and 2742 optimized energy efficiency and lifespan within the network. When
rounds (100 %). Conversely, the EEPEG-PA suite demonstrates height­ utilizing the minimal spanning tree for routing, energy consumption is
ened resilience, with EEPEG-PA achieving the respective mortality reduced as nodes only transmit data to their nearest parent node.
thresholds at later rounds (2218 for 10 %, 3143 for 50 %, 4152 for 100 Through MATLAB simulations, it has been confirmed that our EEPEG-
%), EEPEG-PA-H showing further improvement (2257, 3359, 4456), A PA-H method significantly enhances the network’s lifespan compared
similar pattern is observed in larger networks; for instance, in a 200- to PEGASIS and its variations.
node setup, PEGASIS reaches these mortality events at 2162, 2342,
and 2892 rounds, whereas EEPEG-PA-H extends these to 2502, 3462,
Declaration of competing interest
and 4565 rounds, thereby underscoring the advanced protocols’ supe­
rior efficacy in preserving network integrity over extended periods.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
4.4. Limitations of the study
the work reported in this paper.

The study’s innovative approach has some limitations that need

Data availability
further investigation. The centralized setup and clustering phases,
similar to LEACH-C, could raise scalability concerns in larger networks
No data was used for the research described in the article.
due to the computational and communication overhead at the base
station (BS). Additionally, relying on the Received Signal Strength In­
References
dicator (RSSI) for node localization, while common, may not always
offer the most accurate positioning due to environmental factors and [1] Z. Fei, B. Li, S. Yang, C. Xing, H. Chen, L. Hanzo, A survey on multi-objective
signal multipath effects. Although the proposed method outperforms optimisation in wireless sensor networks: metrics, algorithms and open problems,
PEGASIS in simulations, real-world deployment could introduce un­ IEEE Commun. Surv. Tutorials (2016) 1–38.
[2] W. Dargie, Dynamic power management in wireless sensor networks: state-of-the-
foreseen variables affecting its performance. The simulations are art, IEEE Sensor. J. 12 (5) (May 2012) 1518–1528.
restricted to specific network sizes and do not cover the full range of [3] A. Jain, A.K. Goel, Energy efficient algorithm for wireless sensor network using
possible sensor node deployments, potentially yielding different out­ fuzzy C-means clustering, Int. J. Adv. Comput. Sci. Appl. 9 (4) (2018) 474–481.
[4] X. Liu, A typical hierarchical routing protocols for wireless sensor networks: a
comes in denser networks or those with more irregular node review, IEEE Sensor. J. 15 (10) (2015) 5372–5383.
distributions. [5] L. Chan, K. Gomez Chavez, H. Rudolph, et al., Hierarchical routing protocols for
wireless sensor network: a compressive survey, Wireless Network 26 (2020)
3291–3314, https://doi.org/10.1007/s11276-020-02260-z.
[6] M. Chenait, B. Zebbane, C. Benzaid, N. Badache, Energy-efficient coverage protocol
based on stable and predictive scheduling in wireless sensor networks, Comput.

10
V.Rama Krishna et al. Measurement: Sensors 36 (2024) 101300

Network. 127 (Nov. 2017) 1–12, https://doi.org/10.1016/J. [19] P.S. Prakash, M. Janardhan, K. Sreenivasulu, S.I. Saheb, S. Neeha, M. Bhavsingh,
COMNET.2017.07.015. Mixed linear programming for charging vehicle scheduling in large-scale
[7] A. Abed, A. Alkhatib, G.S. Baicher, Wireless sensor network architecture, Int. J. rechargeable WSNs, J. Sens. 2022 (2022).
Comput. Network. Commun. Syst. 35 (2012) 11–15. [20] D.N. Wategaonkar, R. Phalnikar, Energy-efficient reliable sector-based clustering
[8] C. Gulzar, &AmeenaYasmeen, Maximum network lifetime with load balance and scheme to improve the lifetime of WSN, J. Inf. Sci. Eng. 39 (3) (2023).
connectivity by clustering process for wireless sensor networks, Int. J. Comput. [21] Abdullah Alhasanat, Bayan Sharif, C. Tsemendis, Efficient RSS-based collaborative
Eng. Res. Trends 3 (7) (2016) 375–383. localisation in wireless sensor networks, Int. J. Sens. Netw. 22 (1) (January 2016)
[9] A. Adamayanthi, Mohammad Riyaz Belgaum, A study of heterogeneity 27–36, https://doi.org/10.1504/IJSNET.2016.079335.
characteristics over wireless sensor networks, Int. J. Comput. Eng. Res. Trends 9 [22] C. Li, A.D. Mount, Euclidean Minimum Spanning Trees Based on Well Separated
(12) (2022) 258–262. Pair Decompositions, 2014.
[10] V. Rama Krishna, S. Sameen Fatima, POWER-AWARE sleep-scheduled tree-based [23] K.H. Rosen, K. Krithivasan, Discrete Mathematics and its Applications: with
routing protocol, Int. J. Emerg. Technol. Innovat. Res. 8 (10) (October-2021) Combinatorics and Graph Theory, Tata McGraw-Hill Education, 2012.
e283–e290, https://doi.org/10.1729/Journal.28432. www.jetir.org. ISSN:2349- [24] R. Srinivas, T. Rama Reddy, R.V.S. Lalitha, S. Vahida, A novel approach to find
5162. minimum cost spanning tree (MST) of a graph, in: Computer Communication,
[11] N.D. Tan, N.D. Viet, SSTBC: sleep scheduled and tree-based clustering routing Networking and IoT, Springer, Singapore, 2021, pp. 85–91.
protocol for energy-efficient in wireless sensor networks, in: Proceedings - 2015 [25] B. Biswas, K. Basuli, S. Naskar, S. Chakraborti, S.S. Sarma, A Combinatorial
IEEE RIVF International Conference on Computing and Communication Algorithm to Generate all Spanning Trees of a Weighted Graph in Order of
Technologies: Research, Innovation, and Vision for Future, IEEE RIVF 2015, Feb. Increasing Cost, arXiv preprint arXiv:1209.4206, 2012.
2015, pp. 180–185, https://doi.org/10.1109/RIVF.2015.7049896. [26] Ramdas Vankdothu, DrMohd Abdul Hameed, Husnah Fatima, A brain tumor
[12] M.J. Pasha, M. Pingili, K. Sreenivasulu, M. Bhavsingh, S.I. Saheb, A. Saleh, Bug2 identification and classification using deep learning based on CNN-lstm method,
algorithm-based data fusion using mobile element for IoT-enabled wireless sensor Comput. Electr. Eng. 101 (2022) 107960.
networks, Meas. Sensor. 24 (2022) 100548. [27] Ramdas Vankdothu, Mohd Abdul Hameed “Adaptive features selection and EDNN
[13] S. Lindsey, C.S. Raghavendra, PEGASIS: power-efficient GAthering in sensor based brain image recognition on the internet of medical things”, Comput. Electr.
information systems, in: Aerospace Conference Proceedings, vol. 3, 2002, Eng. 103 (2022) 108338.
pp. 1125–1130. [28] Ramdas Vankdothu, Mohd Abdul Hameed, Ayesha Ameen, Unnisa Raheem, Brain
[14] W.R. Heinzelman, A. Chandrakasan, H. Balakrishnan, Energy efficient image identification and classification on Internet of Medical Things in healthcare
communication protocol for wireless microsensor networks, in: Proceeding of 33rd system using support value based deep neural network, Comput. Electr. Eng. 102
Annual Hawaii Int. Conf. Syst. Sci., vol. 2, Jan. 2000, p. 10. (2022) 108196.
[15] A.P. Chandrakasan, H. Balakrishnan, W.B. Heinzelman, An application-specific [29] Ramdas Vankdothu, Mohd Abdul Hameed, Brain tumor segmentation of MR
protocol architecture for wireless microsensor networks, IEEE Trans. Wireless images using SVM and fuzzy classifier in machine learning, Meas. Sensor. J. 24
Commun. 1 (4) (October 2002). (2022) 100440.
[16] A. Somauroo, V. Bassoo, Energy-efficient genetic algorithm variants of PEGASIS for [30] Ramdas Vankdothu, Mohd Abdul Hameed” Brain tumor MRI images identification
3D wireless sensor networks, Appl. Comput. Inform. (2020), https://doi.org/ and classification based on the recurrent convolutional neural network, Meas.
10.1016/j.aci.2019.07.002 ahead-of-print No. ahead-of-print. Sensor. J. 24 (2022) 100412.
[17] M. Ali, F. Tariq, Dynamic chain reconfiguration in PEGASIS for enhanced network [31] Bhukya Madhu, M. Venu Gopala Chari, Ramdas Vankdothu, Arun Kumar Silivery,
lifetime, IEEE Commun. Lett. (2022). Veerender Aerranagula, Intrusion detection models for IOT networks via deep
[18] M. Abo-Zahhad, M. Farrag, A. Ali, O. Amin, An energy consumption model for learning approaches, Meas. Sensor. J. 25 (2022) 100641.
wireless sensor networks, in: 5th International Conference on Energy Aware [32] Mohd Thousif Ahemad, Mohd Abdul Hameed, Ramdas Vankdothu, COVID-19
Computing Systems & Applications, Cairo, Egypt, 2015, pp. 1–4, https://doi.org/ detection and classification for machine learning methods using human genomic
10.1109/ICEAC.2015.7352200. data, Meas. Sensor. J. 24 (2022) 100537.

11