You are on page 1of 7

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/332824518

Multi-hop PSO based routing protocol for Wireless Sensor Networks with
Energy Harvesting

Conference Paper · March 2019


DOI: 10.1109/ICTAS.2019.8703638

CITATIONS READS

11 224

3 authors, including:

Marcel Ohanga Odhiambo


Vaal University of Technology
56 PUBLICATIONS   124 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

1. Integrated Image Capture and Motion Detection Home Security Systems. View project

Call for Chapters: Handbook of Research on Mitigating the Challenges of the Internet of Things View project

All content following this page was uploaded by Marcel Ohanga Odhiambo on 16 May 2019.

The user has requested enhancement of the downloaded file.


2019 Conference on Information Communications Technology and Society (ICTAS)

Multi-hop PSO based routing protocol for Wireless Sensor


Networks with Energy Harvesting
Tlholiso W. Tukisi, Tebello N.D. Mathaba, Marcel Ohanga Odhiambo
Department of Process Control and Computer Systems,
Faculty of Engineering,
Vaal University of Technology
1tlholisot@vut.ac.za
2tebellom1@vut.ac.za
3marcelo@vut.ac.za

Abstract— In this paper, three energy efficient routing carry when collecting, aggregating and sending data to the
strategies for a design of a Particle Swarm Optimization based Sink [2], [3]. Therefore, load balancing is very important when
Wireless Sensor Network (WSN) routing algorithm are designing a WSN routing protocol.
presented. The first strategy maximizes the energy of the node Regardless of how good the routing protocol can be, the
with the lowest energy (the worst performing node) in the WSN,
the second strategy maximizes the overall WSN energy and the
battery charge will eventually deplete, and sensor nodes will
third strategy maximizes the energy of the worst performing node stop functioning. One viable option is to incorporating energy
while improving network latency. The best of these strategies is harvesting in WSN [4], [5], [6]. In this case, sensor nodes can
further tested in a WSN with energy harvesting nodes. The use rechargeable cells and be recharged from different sources
results are compared with a benchmark variant of the Low of renewable energy such as solar energy, wind energy or heat
energy adaptive clustering hierarchy (LEACH) protocol, called from nature [6]. However, if the routing algorithm is not
the LEACH-Centralized Sleeping (CS). The results show that the energy harvesting aware, the energy harvested might not be
strategy that maximizes the overall network energy can improve optimally utilized in the WSN.
the network lifetime better. The results further show that the In this paper, we compare three new energy efficient routing
energy harvesting aware routing protocol can extend the lifetime
of a WSN more than a protocol that is not energy harvesting
strategies based on Particle Swarm Optimization (PSO)
aware. algorithm together with a benchmark LEACH-Centralized
Sleeping (CS) algorithm. These strategies take into
Index Terms—Wireless Sensor Networks, Energy harvesting, consideration the residual energy of alive nodes in the
PSO, Two-tier Clustering, LEACH-CS. network, the transmission distance to the Sink as well as
latency for every round of data transmission. In this way,
I. INTRODUCTION energy balancing across all sensor nodes can be achieved.

I n the recent past years, wireless sensor networks have


gained more attention due to enormous advantages they
The three routing algorithms are executed in two stages.
During the first stage, the WSN is organized into a two tier
architecture by adopting a dynamic clustering algorithm from
bring to various application domains such as medical, LEACH-CS [7]. The second stage routes data by solving an
agriculture, military and industry. However, energy optimization problem using PSO by implementing each
consumption in Wireless Sensor Networks (WSNs) remains a strategy’s energy cost function as the optimization’s objective
key challenge during their operation. Traditionally, sensor function. When the results are compared with the LEACH-CS
nodes in WSNs are powered by non-chargeable batteries, as a protocol, it is clear that the use of relaying nodes at the upper
result, when that battery charge depletes, the node stops tier hierarchy of two-tiered WSN can extend the lifetime of the
functioning. network. The results also show that the strategy that maximize
Energy in WSN nodes is consumed mainly during the overall network energy improves the network lifetime
transmission and reception of data. Most of the research better. Lastly, we assess the impact of incorporating energy
literature, like Low energy adaptive clustering hierarchy harvesting in the most energy efficient strategy. The results
(LEACH) in [1], adopt an energy model where the obtained show that the energy harvesting aware routing
transmission energy (𝐸) is mainly dominated by the overall protocol can extend lifetime of WSN than the protocol that is
distance between the transmitter and receiver (d), i.e.𝐸 ∝ 𝑑 𝜆 , not.
where λ is the path loss exponent and 2 ≤ λ ≤ 4 [2]. This means The outline of the paper is organized as follows. Section II
that the minimization of transmission distance between nodes presents the related literature on routing protocols, energy
involved in the routing path can reduce the energy harvesting WSNs and PSO overview. Section III presents the
consumption. network system model. The analysis and discussion of the
Although LEACH and other single-hop routing protocols results obtained via MATLAB simulation are provided in
still use cluster-based architecture, also known as hierarchical Section IV. Finally, we conclude our finding in Section V.
or two-tier network, there is no inter-cluster communication in
the network because cluster heads (CHs) directly II. RELATED WORK
communicate with the Sink [3]. As a result, this process In WSN, clustering and routing mechanisms/techniques have
requires high range of transmission power in the network been studied extensively to improve the performance and
which lead to a significant amount of energy consumption. prolong lifetime of the network. The core of this mechanisms
Moreover, in many WSNs, CHs are selected from normal is based on energy optimization during the routing processes.
sensor nodes which die quickly from the extra workload they In this section, we present a brief overview of energy efficient
ISBN 978-1-5386-7365-2/19/$31.00 ©2019 IEEE
routing protocols, energy harvesting and optimization tools present the Time-slot based Information-updating (TI)
with an example of PSO. Mechanism to update the routing table periodically without
extra overhead.
A. Energy efficient routing approaches
The authors in [5], propose a solution for energy limitation
In their survey on energy efficiency in wireless sensor
of WSN by introducing an Energy Neutral Clustering (ENC)
network, Rault et al, [8], highlighted that energy efficient
protocol. This protocol group the sensor nodes in the network
routing involve clustering architectures, energy as routing
into several clusters with the goal of providing a perpetual
metric approaches, multi-path routing and multi-hop routing,
network operation. ENC routing employs a novel CH Group
relay node placement and sink mobility. In a multi-hop
mechanism that allows a cluster to use multiple CHs to share
routing, it is important to emphasize on the energy as a metric
heavy load traffic load and reducing the frequency of cluster
as illustrated by [2], [9], [10], [11], because it can greatly
re-formation. ENC routing protocol also prevents cluster
improve energy balancing in the network. Furthermore,
failures by ensuring a network wide energy neutral operation.
placement of relay nodes also play a major role as it minimizes
In this paper, we design a routing protocols using PSO
energy consumption and improves link quality by minimizing
algorithm to balance energy usage between EH-nodes and O-
the distance between the source node and the next hop [2].
nodes in a two-tied WSN. Our analysis deals with an ordinary
B. Heuristics based Routing protocols WSN and then a later section of the paper deals with analysis
Low energy adaptive clustering hierarchy (LEACH) [1] is of a homogeneous WSN with energy harvesting nodes, similar
one of the most popular distributed cluster-based routing to those in [5] and [6].
protocol in WSN that has been proven to be an effective
approach to prolong the network lifetime. More routing III. OVERVIEW OF PSO AND SYSTEM MODEL
protocols were developed to improve on the performance of
A. Overview of PSO
LEACH in the following research; [7], [12], [13], [14] and
[15]. In LEACH-CS [7], authors implement a sleeping mode PSO is a swarm intelligence based optimization method. It was
for a percentage of the nodes in the network depending on the inspired by social behavior of bird flocking or fish schooling
data sent at a certain time. They also select CHs based on the and on how they can explore and exploit the multi-dimensional
residual energy where nodes with high remaining energy have search space for food and shelter [18], [21]. PSO has
high chance of becoming CHs. This approach corrects one of demonstrated success in solving complex non-linear
the major limitation of LEACH and improves energy optimization problems. It generally optimizes a cost function
balancing in the network. However, most LEACH based using a candidate particle population in the search-space while
protocol have a common drawback of encouraging single hop conforming to formulae for velocity and position of the
communication which can be expensive if the distance from particle [22]. It accomplishes best results using a series of
the cluster head (CH) and the sink is large [16]. iterations attempting to enhance a candidate solution’s cost
function value [23]. In each iteration, velocity of each particle
C. Metaheuristics based multi-hop routing protocols is updated using the current velocity of the particle and the
Due to the nature of multi-hop routing process in WSNs, previous local best and global best position [22]. Finally, the
which are characterized by determining a set of intermediate swarm will meet the optimal positions.
nodes to be selected to form a data forwarding path between In this paper, we adopt PSO’s intuitive formulation presented
the source and the destination [17], metaheuristics in [24]:
optimization approaches are well recommended because of
their ability to test every possible solution towards finding the 𝑣𝑙𝑚+1 = 𝜑 𝑚 𝑣𝑙𝑚 + 𝛼1 [𝛾1,𝑙 (𝑃𝑙 − 𝑥𝑙𝑚 )] + 𝛼2 [𝛾2,𝑙 (𝐺𝑙 − 𝑥𝑙𝑚 )] (1)
optional path from the source node to the Sink. Among many 𝑥𝑙 (𝑚 + 1) = 𝑣𝑙 (𝑚 + 1) + 𝑥𝑙 (𝑚) (2)
metaheuristics algorithms, PSO has attracted more attention to
researchers focused on WSN lifetime maximization through The vectors 𝑥𝑙𝑚 and 𝑣𝑙𝑚 are the current position and velocity
enhanced routing protocols, e.g. [9], [11], [2], [18], [19]. of the l-th particle in the m-th generation. The swarm consists
According to [20] and [21], PSO has many advantages over of N particles, i.e., l = {1,...,N}. 𝑃𝑙 is the personal best position
other alternative optimization techniques like Genetic of each individual and 𝐺𝑙 is the global best position observed
Algorithms (GA) which include ease of implementation on the among all particles up to the current generation. The
hardware or software, high quality solutions because of its
ability to escape from local optima and quick convenience. Algorithm 1. PSO algorithm
D. Energy Harvesting in WSNs 1: for each particle do
2: initialize particle
While sensor nodes in LEACH have a high chance of
3: end for
becoming CHs because the network is homogeneous, in EH-
4: while target fitness or maximum epoch is not attained do
WSN proposed by [4], network homogeneity is invalid.
5: for each particle do
Therefore, the authors in [4] introduced an Energy Potential 6: calculate fitness
Function (EP-Function) to measure a nodes’ capability of 7: if current fitness value better than (pbest) then
energy harvesting. EP-Function extend the traditional protocol 8: pbest=current fitness
LEACH to Energy Potential LEACH which is suitable for 9: end for
energy harvesting WSNs. 10: set gbest to the best one among all pbest
In [6], a routing protocol named energy harvesting routing 11: for each particle do
which takes energy harvesting into account is designed to 12: update velocity using (1)
improve energy efficiency. In their design, they introduce 13: update position using (2)
hybrid routing metric which combines residual energy and 14: end for
energy harvesting rate. To efficiently select the next hop, they 15: end while
parameters 𝛾1,2 ∈ [0,1] are uniformly distributed random path loss model can be approximated as a two ray multi-path
4
values and 𝛼1,2 are acceleration constants. The parameter 𝜑 is propagation model with 𝑑𝑖.𝑗 power attenuation.
the particle inertia which gives rise to a certain momentum of Due to different task requirements, the Cluster Member
the particles. The original PSO algorithm is shown Algorithm (CM) and the Cluster Head (CH) consume energy in different
1. Our algorithm uses priority coding to encode routes like in ways. For sensors that are selected to be the CM, energy will
[21]. The routing tree is built from the encoded particle in a be spent on information sensing and data packet transmission.
branch growth process. Each branch is a route from the cluster Thus, based on (5), the total amount of energy consumed by a
head to the Sink. CM to sense and transmit bits in the message packet is
estimated by;
𝐸𝐶𝑀(𝑖) = 𝜀𝑠𝑥 + 𝐸𝑇𝑥(𝑖,𝑗) (5)
where 𝜀𝑠𝑥 is the amount of energy consumed to sense
information and 𝐸𝑇𝑥(𝑖,𝑗) is given by (4).
A CH will consume energy to receive data packets sent
from the CMs. It also consumes energy to send the combined
data packets to the Sink. Thus, for a cluster k, if it has 𝑁𝑘 CMs,
the total amount of energy consumed by the CH to handle
(sense, receive and then transmit) information sent by the CMs
Figure 1: A sensor node energy consumption components.
is estimated as follows;
B. System Model 𝐸𝐶𝐻(𝑖) = 𝜀𝑠𝑋 + 𝐸𝑅𝑥(𝑖) ∙ 𝑁𝑘 + 𝐸𝑇𝑥(𝑖,𝑗) ∙ (𝑁𝑘 + 1) (6)
1) WSN model
The energy consumed for relaying traffic messages is given
For our model, a two-tier WSN is considered. The network by:
consists of N sensor nodes, k cluster heads and one Sink. A
stationary deployment of sensor nodes and the Sink is 𝐸𝑅(𝑖) = 𝐸𝑅𝑥(𝑖) ∙ 𝑁𝑅 + 𝐸𝑇𝑥(𝑖,𝑗) ∙ 𝑁𝑅 ) (7)
assumed. A sensor node’s energy consuming components where 𝑁𝑅 is the number of nodes whose traffic is being
include sensing module, processing module and transceiver relayed. For each node in the network, the energy used per
module as shown in Figure 1. Each sensor node is identified messaging round is calculated using (8). When the node is not
by a unique ID and the Sink is given the ID of 0. The CH
a relay node, 𝑁𝑅 = 0 and hence 𝐸𝑅𝑥(𝑖) = 0 as well.
connect nodes closer to it and each sensor node connect to only
one CH. In the case of an energy harvesting network, we also 𝐸𝐶𝑀(𝑖) + 𝐸𝑅(𝑖) ; 𝑖 ≡ 𝐶𝑀
assume that all nodes have energy harvesting features to 𝐸𝑢𝑠𝑒𝑑𝑖 = { (8)
𝐸𝐶𝐻(𝑖) + 𝐸𝑅(𝑖) ; 𝑖 ≡ 𝐶𝐻
provide random amounts of additional energy to the nodes.
The lower tier is made up of cluster members (CM) while the It is also important to compute the remaining energy after a
upper tier is made up of cluster heads (CH) and relay nodes messaging round by using the formula below:
(RN). CMs can also be relay nodes. Energy is consumed by 𝐸𝑁𝑒𝑥𝑡𝑖 = 𝐸𝑖𝑛𝑖𝑡𝑖 − 𝐸𝑢𝑠𝑒𝑑𝑖 (9)
both CH & CM for sensing, receiving and transmitting data.
Where 𝐸𝑁𝑒𝑥𝑡𝑖 is the remaining energy after a round, 𝐸𝑖𝑛𝑖𝑡𝑖 is
2) Energy consumption model the previous round energy of a node in the network, 𝐸𝑢𝑠𝑒𝑑𝑖 is
Multiple models for energy consumption and link quality exist the energy consumed by the i-th node.
– i.e. Euclidian Distance Factor (EDF) energy model used in C. Energy Efficient Protocols
LEACH [1], Receiver Signal Strength Index (RSSI) energy
and link quality model [21] and power transmission control In this paper, we compare three energy efficient routing
(PTC) energy model [25], [26]. strategies based on (PSO) algorithm. The three strategies are
In this paper we adopt a simplified EDF energy model similar individually implemented as fitness functions of the PSO
to that applied in [5], for receiving and transmission energy as algorithm in Algorithm 2. The optimal route is therefore
shown in (3) and (4), respectively. obtained from the global best particle, 𝐺𝑙 . The simulation

𝐸𝑅𝑥(𝑖) = 𝜀𝑟𝑥 ∙ 𝑏𝑡𝑖 (3) Algorithm 2. Simulation Algorithm


2
1: Load WSN and energy parameters
𝜀𝑡𝑥 ∙ 𝑏𝑡𝑖 + 𝑏𝑡𝑖 ∙ 𝜀ℎ𝑜𝑝 ∙ 𝑑𝑖.𝑗 𝑖𝑓 𝑑 < 𝜆 2: while the stopping condition has not been met do
𝐸𝑇𝑥(𝑖,𝑗) = { 4 (4)
𝜀𝑡𝑥 ∙ 𝑏𝑡𝑖 + 𝑏𝑡𝑖 ∙ 𝜀ℎ𝑜𝑝 ∙ 𝑑𝑖.𝑗 𝑖𝑓 𝑑 ≥ 𝜆 3: Establish network clusters
4: Find the optimal route using Algorithm 1
where constant 𝜀𝑟𝑥 and 𝜀𝑡𝑥 are the amounts of energy 5: Decode the routing particle into a routing tree
consumed by the sensor node electronics to receive and 6: Run messaging round using the new routing tree
transmit message packets, respectively. 𝜀ℎ𝑜𝑝 is the 7: if minimum network energy is less than zero then
transmitting amplifier coefficient, 𝑏𝑡𝑖 is the number of bits in 8: The stopping condition has been reached
the message packet and 𝑑𝑖,𝑗 is the Euclidian distance between 9: end if
10: if maximum number of rounds is reached then
node i and j while λ is the cross over distance. When d is
11: The stopping condition has been reached
smaller than λ, the path loss can be treated as Fris free space
2 12: end if
loss with 𝑑𝑖.𝑗 power attenuation. When d is larger than λ, the 13: Display results
algorithm for these energy efficient routing strategies is residual energy. Furthermore, CHs are allocated an equal
presented below: number of member nodes per messaging round. Member
i. Maximizes the minimum energy in the network –this strategy nodes are subsequently associated to CHs closest to them.
discourages the use of the weakest node in the network to save
B. Network lifetime simulations
its energy and therefore extend the network’s lifetime. This
leaves the nodes with low energy unused until their energy is In this section, the performance of the proposed routing
higher relative to all other nodes in the network while protocols is investigated. The simulation was carried out in
sacrificing nodes with the highest energy. As a result, energy MATLAB platform. The routing Algorithm 2 incorporates the
consumption is equally balanced across the network which PSO to execute the cost functions discussed in Section III.C.
extend network lifetime under a definition of first node dead The PSO parameters used are as follows: φ = 0.1, 𝛼1 = 0.25,
(FND). The strategy fitness function is given by (10). 𝛼2 = 3, N = 60 and number of generation is 60.
A random WSN was created with a Sink, CMs and CHs. We
𝐸𝑐𝑜𝑠𝑡1 = max {min(𝐸𝑁𝑒𝑥𝑡 )} (10) assume that all the sensors have the same hardware capabilities
𝑙 𝑖 𝑖
and initial energy 𝐸𝑖𝑛𝑖𝑡𝑖𝑎𝑙 of 20 J. The energy consumption
Where 𝑙 is the particle index. model uses the following parameters adapted from [5] for
ii. Maximize the overall average network energy – this strategy
simulation settings; 𝜀𝑆𝑥 = 100 nJ / bit, 𝜀𝑅𝑥 = 100 nJ / bit, 𝜀𝑇𝑥 =
computes different routes and select the routes that result with
50 nJ / bit, 𝜀ℎ𝑜𝑝 = 100 pJ / bit / m2. The network is executed
the lowest average energy consumption. In this way, the
strategy attempts to balance the energy used by all nodes in the until the last node dies with sensor nodes generating messages
network: that have a payload of 6400 bits per messaging round. In this
way, we obtain results allowing us to assess the performance
1 of these protocols in terms of energy efficiency and lifetime.
𝐸𝑐𝑜𝑠𝑡2 = max {𝑁 ∑𝑁
𝑖=1 𝐸𝑁𝑒𝑥𝑡𝑖 } (11)
𝑙 For energy efficiency, we consider average node energy to
Where N is the number of nodes in the network. determine how effective the routing strategy is. For network
iii. Maximize the minimum energy while improving network lifetime, we consider three common definitions of network
latency– Besides being energy-efficient, this strategy also lifetimes as defined by [27]. That is, first node dies (FND), half
attempts to improve network latency by reducing the number nodes die (HND) and last node dies (LND)
of hops from the CH to the sink. In this way, data from the CM C. Results and analysis
takes the shortest route to the sink. This strategy is represented
as follows: Figure 2 shows the results of average node energy per
messaging round when comparing the three strategies and
𝐸𝑐𝑜𝑠𝑡3 = max {𝑤 ∙ min(𝐸𝑁𝑒𝑥𝑡 )} + (1 − 𝑤) ∙ 𝑁(𝑄), 0 ≤ 𝑤 ≤ 1 (12) LEACH-CS in terms of energy efficiency. 𝐸𝑐𝑜𝑠𝑡2 out performs
𝑙 𝑖 𝑖

where 𝑤 is the multi-objective weighting parameter and 𝐸𝑐𝑜𝑠𝑡1 , 𝐸𝑐𝑜𝑠𝑡3 and LEACH-CS when balancing average node
𝑁(𝑄) the total number of nodes used by the resulting routing energy across the network. This is because 𝐸𝑐𝑜𝑠𝑡2 maintains
path. The value of w determines the importance of a one the energy of nodes with higher energy high to keep the
objective over the other. If the multi-objective weighting network average energy high as it selects routing paths with
parameter (w), is equal to 1, then network latency is ignored. low average energy. 𝐸𝑐𝑜𝑠𝑡1 , on the other hand, cannot maintain
However, if w is 0, the new route tree is solely based on average node energy since its objective is to balance average
maximizing the network lifetime. If w is less than 1, the network energy by keeping the minimum node energy as high
protocol is partially focused on reducing network latency as as possible throughout the network lifetime. Furthermore, the
well reducing average network energy. This cost function is network of 𝐸𝑐𝑜𝑠𝑡1 suffer sudden death as shown in Figure 2
similar to the one implemented in [21] for energy balancing in because all the node always has more or less the same energy
the WSN. after every round, hence they all die simultaneously.

D. Energy Harvesting
From the above energy efficient routing strategies, we choose
the one that gives the highest network lifetime and incorporate
energy harvesting into it to further improve the network
lifetime. To make the newly constructed protocol energy
harvesting aware, the formula for 𝐸𝑁𝑒𝑥𝑡𝑖 is modified as
indicated in (13).
𝐸𝑁𝑒𝑥𝑡𝑖 = 𝐸𝑖𝑛𝑖𝑡𝑖 − 𝐸𝑢𝑠𝑒𝑑𝑖 + 𝐸𝑅𝑒𝑐ℎ𝑖 (13)

IV. ANALYSIS AND DISCUSSION


A. Benchmark protocol
In this paper we use a variant of the original LEACH protocol
that implements centralized sleeping (CS), called LEACH-CS,
as a benchmark. We also adopted its clustering algorithm in
our proposed PSO routing algorithm. LEACH-CS, introduced
in [7], uses a clustering algorithm that dynamically selects a Figure 2: Comparison of average node energy per messaging round
predefined percentage of nodes as CHs based the highest
Figure 3: Comparison of alive nodes per messaging round

Figure 5: Comparison of network lifetime with Energy Harvesting (Alive


nodes per round)

respectively. Since nodes are getting random amounts of


energy during the harvesting process for every messaging
round, 𝐸𝑐𝑜𝑠𝑡2 +EH still maintain the average node energy well
until the average node energy reaches 20% of initial energy
after 800 messaging rounds as shown in Figure 4. However,
the network still survives until 1800 messaging rounds. It is
important to indicate that 𝐸𝑐𝑜𝑠𝑡2 +EH extended network
lifetime until 50% of nodes die and eventually suffer a sudden
death as shown in Figure 5. In general, the results show that
energy harvesting can improve network lifetime by more than
50% when compared to non-energy harvesting network using
𝐸𝑐𝑜𝑠𝑡2 .

V. CONCLUSION
Figure 4: Comparison of network lifetime (𝐸𝑐𝑜𝑠𝑡2 +Energy harvesting In conclusion, the strategy that maximizes the average network
Average node energy) energy can extend the lifetime of the WSN considerably as
The results further show that average node energy is opposed to maximizing minimum node energy when network
proportional to network lifetime as shown in Figure 2 and 3. lifetime is defined by the last node dead. However, this is
𝐸𝑐𝑜𝑠𝑡1 strategy result in the WSN taking periods of 385 achieved by sacrificing nodes with the least amount of energy.
messaging rounds before the first nodes dies while 𝐸𝑐𝑜𝑠𝑡2 takes Furthermore, this proposed protocol also adapts easily to
337 messaging rounds before FND as shown in Figure 3. It can energy harvesting function since it considers nodes with the
also be noted that 𝐸𝑐𝑜𝑠𝑡2 performs better than other strategies high energy level for relaying massages traffic. Results show
in terms of HND as it takes up to 1025 messaging rounds that energy harvesting can double the network lifetime. More
before 50% of the nodes become dead. 𝐸𝑐𝑜𝑠𝑡2 further gives the
work needs to be done to consider other factors such as link
best performance of almost 1178 messaging rounds before the
last node is dead. This performance is superior to that of the quality, network coverage, traffic aggregation by CH, the route
benchmark LEACH-CS protocol that gives only 988 cost and network convergence rate.
messaging rounds before the last node is dead.
𝐸𝑐𝑜𝑠𝑡3 executes a similar routing process as 𝐸𝑐𝑜𝑠𝑡1 , but also REFERENCES
reduces network latency. For example, when w=1,
𝐸𝑐𝑜𝑠𝑡3 =𝐸𝑐𝑜𝑠𝑡1 , and network latency is ignored. In this study, w [1] W. B. Heinzelman, A. P. Chandrakasan and H.
was set to 0.5, hence why the network last longer than 𝐸𝑐𝑜𝑠𝑡1 Balakrishnan, "An application-specific protocol
in terms of FND, HND and LND. architecture for wireless microsensor networks," IEEE
Since 𝐸𝑐𝑜𝑠𝑡2 proves to be the most energy efficient strategy, Transactions on Wireless Communications, vol. 1, no. 4,
it is enhanced to make it EH-aware by substituting (7) with pp. 660-670, 2002.
(11) into a new algorithm called 𝐸𝑐𝑜𝑠𝑡2 +EH. A performance [2] P. Kuila and P. K. Jana, "Energy efficient clustering and
comparison of 𝐸𝑐𝑜𝑠𝑡2 +EH with LEACH-CS in an energy routing algorithms for wireless sensor networks: Particle
harvesting WSN is shown in Figure 4 and 5 in terms of average Swarm Optimization approach," Engineering
node energy and alive nodes per messaging round
Applications of Artificial Intelligence, vol. 33, pp. 127- Networks," Procedia Computer Science, vol. 57, pp.
140, 2014. 1005-1014, 2015.
[3] A. Braman and G. R. Umapathi, "A Comparative Study [16] J. Roselin, p. Latha and S. Benitta, "Maximizing the
on Advances in LEACH Routing Protocol for Wireless wireless sensor networks lifetime through energy
Sensor Networks:A survey," International Journal of efficient connected coverage," Ad Hoc Networks, vol.
Advanced Research in Computer and Communication 62, pp. 1-10, 2017.
Engineering, vol. 3, no. 2, pp. 5683-5690, 2014. [17] M. A. Zungeru, L.-M. Ang and K. P. Seng, "Classical
[4] M. Xiao, X. Zhang and Y. Dong, "An Effective Routing and swarm intelligence based routing protocols for
Protocol for Energy Harvesting Wireless Sensor wireless sensor network: A survey and comparison,"
Networks," in 2013 IEEE Wireless Communications and Journal of Network and Computer Applications, vol. 35,
Networking Conference (WCNC): NETWORKS, 2013. pp. 1508-1536, 2012.
[5] S. Peng, T. Wang and C. P. Low, "Energy neutral [18] M. Azharuddin and J. K. Prasanta, "Particle swarm
Clastering for energy harvesting wireless senoror optimization for maximizing lifetime of wireless sensor
networks," Ad Hock Networks, vol. 28, pp. 1-16, 2015. networks," Computer & Electrical Engineering, vol. 51,
[6] Y. Cao, X.-Y. Liu, L. Kong, M.-Y. Wu and M. K. Khan, pp. 26-42, 2016.
"EHR: Routing Protocol for Energy Harvesting Wireless [19] R. Sudarmani and K. R. Shanker Kumar, "Particle
Sensor Networks," in 2016 IEEE 22nd International Swarm Optimization-Based Routing Protocol for
Conference on Parallel and Distributed Systems, 2016. Clustered Heterogeneous Sensor Networks with Mobile
[7] A. A. Hady, S. M. Abd El-kader and H. S. Eissa, Sink," American Journal of Applied Sciences, vol. 10,
"Intelligent Sleeping Mechanism for wireless sensor no. 3, pp. 259-269, 2013.
networks," Egyptian Informatics Journal, vol. 14, no. 2, [20] W. Guo and W. Zhang, "A survey on intelligent routing
pp. 109-115, 2013. protocols in wireless sensor networks," Journal of
[8] T. Rault, A. Bouabdallah and Y. Challal, "Energy Network and Computer Applications, vol. 38, pp. 185-
Efficiency in wireless sensor networks: A top-down 201, 2014.
surver," Computer Network, vol. 64, pp. 104-122, 2015. [21] R. S. Elhabyan and M. C. Yagoab, "Two-tier particle
[9] R. S. Elhabyan and M. C. Yagoub, "PSO-HC: Particle swarm optimization protocol for clustering and routing
Swarm Optimization Protocol for Hierarchical in wireless sensor network," Journal of Networks and
Clustering in Wireless Sensor Networks," in 10th IEEE Computer Application, vol. 52, pp. 116-128, 2015.
International Conference on Collaborative Computing: [22] P. C. Srinivasa Rao, J. K. Prasanta and H. Banka, "A
Networking, Applications and Worksharing particle swarm optimization based energy efficient
(CollaborateCom 2014), Miami, FL, USA, 2014. cluster head selection algorithm for wireless sensor
[10] A. Adamou Abba Ari, B. Omer Yenke, N. Labrauoi and networks," The Journal of Mobile Communication,
I. Damakoa, "A power efficient cluster-based routing Computation and Information, vol. 22, no. 3, pp. 1-16.
algorithm for wireless sensor networks: Haneybees [23] R. RejinaParvin and C. Vasanthanayaki, "Particle
swarm intelligent based aproach," Journal of Networks Swarm Optimization-Based Clustering by Preventing
and Comuter Applications, vol. 69, pp. 77-97, 2016. Residual Nodes in Wireless Sensor Networks," IEEE
[11] S. Kaur and R. Mahajan, "Hybrid meta-heuristic Sensors Journal, vol. 15, no. 8, pp. 4264-4274, 2015.
optimization based energy efficient protocol for wireless [24] S. Ebbesen, P. Kiwitz and L. Guzzella, "A generic
sensor networks," Egyptian Informatics Journal, 2018. Particle Swarm Optimization Matlab Function," in 2012
[12] M. J. Usman, Z. Xing, H. Chiroma and T. Herawan, American Control conference, 2012.
"Modified Low Energy Adaptive Clustering Hierarchy [25] S. Lin, F. Miao, J. Zhang, G. Zhou, L. Gu, T. He, J. A.
Protocol for Efficient Energy Consumption in Wireless Stankovic, S. Son and G. J. Pappas, "ATPC: Adaptive
Sensor Networks," 2014. transmission power control for wireless sensor
[13] M. O. Farooq, A. B. Dogar and G. A. Shah, "MR- networks," ACM Trans. Sen., vol. 12, no. 1, 2016.
LEACH: Multi-hop Routing with Low Energy Adaptive [26] S. Kurt, H. U. Yildiz, M. Yigit, B. Tavli and V. C.
Clustering Hierarchy," in 2010 Fourth International Gungor, "Packet Size Optimization in Wireless Sensor
Conference on Sensor Technologies and Applications Networks for Smart Grid Applications," IEEE
(SENSORCOMM), Venice, Italy, 2010. TRANSACTIONS ON INDUSTRIAL ELECTRONICS,,
[14] F. Fanian, M. K. Rafsanjani and V. K. Bardsiri, "A vol. 64, no. 3, pp. 2392-2401, 2017.
Survey of Advanced LEACH-based Protocols," [27] N. H. Mak and W. K. Seah, "How long is the lifetime of
International Journal of Energy, Information and a wireless sensor network?," in 2009 International
Communications, vol. 7, no. 1, pp. 1-16, 2016. Conference on Advanced Information Networking and
[15] R. P. Mahapatra and R. K. Yadav, "Descendant of Applications, Bradford, UK, 2009.
LEACH Based Routing Protocols in Wireless Sensor

View publication stats

You might also like