A study on multi-hop routing scheme for wireless sensor networks
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
A Study on Multi-Hop Routing Scheme for Wireless Sensor Networks
Thi Nguyen1, Thuy-Anh Nguyen2*, Trong-Minh Hoang3
1Communication Technology RDC, The Voice of Viet Nam, Hanoi, Vietnam
2 School of Electronics and Telecoms, Hanoi University of Science and Technology, Hanoi, Vietnam
3 Telecommunication Faculty, Posts and Telecommunications Institute of Technology, Hanoi, Vietnam
*Email: [email protected]
Abstract
Wireless sensor networks (WSN) play an important role in IoT (Internet of Things) as an interconnecting
infrastructure. Working with a limited energy source, the vital challenge for WSN is to prolong the network
lifetime as an important performance metric. Furthermore, the limitations of regular transmission technologies
create localized network areas of a multi-hop fashion form that adds more constraints to enhance the network
performance. Hence, the clustering strategies initially have solved these problems and received the attention
of many studies, an approach using unequal clustering strategy has yielded some positive results since
consumed energy gaps are avoided in regions near base stations. However, the routing strategy among
cluster heads in multi-hop wireless networks is still a big challenge because of its inefficiency in energy
consumption aspects. Therefore, in this paper, we propose a novel method that combining an unequal
clustering problem and a simple multi-hop routing to prolong network life. The numerical results show that the
proposed solution is more effective than other models in recent studies
Keywords: Wireless sensor network, fuzzy logic, clustering technique, routing algorithm, lifetime.
1. Introduction*
form have been developed according to various criteria
for selecting cluster heads such as probability, node
residual energy, density, etc. [6]. To integrate multiple
criteria with reasonable computational complexity,
some fuzzy logic-based clustering algorithms have
also been developed recently [7-9]. These protocols
operate in both single-hop and multi-hop
communication environments. For the multi-hop
fashion, the appearance of energy holes nearby the
base station has created new challenges for clustering
solutions. In addition, the efficiency of multi-hop
communication between cluster heads also poses a
need to develop new algorithms to improve wireless
sensor network performance.
The WSN act as a communication infrastructure
in IoT with many applications in multiple fields like
military, agriculture, smart control, etc. The WSN
usually consist of a large number of small sensors to
gather information or environment parameter based on
single- or multi-hop wireless communication. The
sensor network not only collects environmental data
but also is used to control the actuators according to
intelligent computational decisions. However,
wireless sensor nodes often use a limited energy source
like a battery (rechargeable and non-rechargeable) and
thus leading to the need to save energy and prolong the
life of the wireless sensor network. Therefore, one of
the challenging sensor network performance issues is
related to energy-saving operating mechanisms [1-4].
In this paper, a novel multi-hop routing scheme
is proposed. In which, the fuzzy logic-based clustering
solution with a greedy routing algorithm is used to
improve the lifetime of the wireless sensor network.
The proposal is validated through numerical
simulation in general scenarios. The results show the
effectiveness of the proposed model when it comes to
better load balancing and prolonging the life of the
network compared to the previously researched model.
The article includes the following sections: Session 2
presents the recent research on clustering and routing
techniques. The proposed clustering method based on
fuzzy logic and routing scheme between cluster head
nodes is demonstrated in session 3. Simulation and
comparison results with previous studies are given in
section 4, and finally, conclusions about the research
An effective approach to save energy
consumptions of sensors in a network is clustering.
Some geographically close sensors are organized into
clusters and select a node to be the head of the cluster
CH (Cluster Head). The information is sent from the
sensors to the cluster head and from there forward to
the BS (Base Station). Since the distance from sensors
to CHs is limited and data is gathered at the cluster
head before being transferred to the base station, the
overall network lifetime is increased. From a basic
energy-efficient clustering protocol called LEACH
(Low Energy Clustering Hierarchy) [5], a series of
variant protocols based on the principle of clustering
ISSN 2734-9373
Received: May 18, 2021; accepted: August 02, 2021
1
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
as well as future development directions are in the last
session.
algorithm is given based on the cluster radius of CH
nodes and the cost function of transmission energy
using an approximate optimization approach. This
proposed scheme can prolong the lifetime of a
multihop sensor network. However, it used three fuzzy
inference sets to support the clustering process that
makes increasing problem complexity. In fact, the
complexity of clustering and routing based on fuzzy
inference system (FIS) is dependent on input
parameters and several rules. Some proposed multi-
hop routing schemes based on FIS were focused on
routing algorithm, traffic balancing, energy hole, and
reliability [14-15].
2. Related Work
In previous studies, many clustering algorithms
based on the idea of LEACH protocol have been
developed to improve the performance of WSN.
Operating in a single-hop network, LEACH chooses
randomly in the sensor nodes to elect a CH cluster head
to get balance energy consumption. The detailed
process of the LEACH protocol is described below.
The data communication process in wireless
sensor networks based on LEACH protocol consists of
two phases in a periodic time frame. This process
consists of two stages: The first phase is the cluster
building phase. The cluster head node is elected during
an election process by voting. The cluster head node
then sends a message to the member nodes informing
them about the cluster head role. Member nodes send
request messages to the cluster head to confirm cluster
membership. The second phase is stable data
communication. The sensor nodes in a cluster send
data to their cluster head. The cluster head gathers data
and sends it directly to the Sink node. It is easy to
recognize that the CH node consumes more energy
than the member nodes because it is responsible for
sending all the data in its cluster. In other words, the
cluster head node’s energy is depleted too soon as
compared to other cluster member nodes. Therefore, a
random CH node selection mechanism is proposed to
prolong the lifetime of the network. The algorithm is
based on a simple principle, each node executes a
random number process in the range [0 1] and takes
that as its value to compare with the threshold T(n).
T(n) probability is a function related to the probability
of selecting the cluster head in the current round. The
node that was selected as a cluster head in the previous
round will have zero chance in the current round. The
details of the LEACH algorithm are presented in [5].
In [14], it is shown that the lifetime of the
network is strongly affected by the uneven distribution
of cluster heads. Therefore, the authors [14] proposed
a fuzzy clustering algorithm based on dynamic particle
cluster optimization (PSO MF). In which, clusters are
formed according to C-Means fuzzy algorithm to
create stability for clusters. From these clusters,
suitable cluster heads are selected through Mamdani
fuzzy inference system (MFIS) through some key
parameters. The input parameters of MFIS include
residual energy, node degree, and distance to the Sink
node. The novel method of the proposal is to use the
particle swarm optimization algorithm to optimize the
rules. Thus, the cluster head selection process will be
adjusted according to the change predefined rules in
Mamdani FIS. However, the introduction of the
principle of forming clusters and the optimization
algorithm also leads to an increase in computational
complexity.
Based on the operating principle of the LEACH
protocol, [15] highlights the limitation of the
probability-based computational model. Clusters that
form based on probability can lead to cluster heads
being close together and subsequently reduce the
energy efficiency of the network. From that
perspective, the authors proposed a new protocol
called LEACH-Fuzzy Clustering (LEACH-FC). This
proposed LEACH-FC protocol performs cluster head
selection based on fuzzy logic and cluster formation to
maximize network lifetime. A cluster deputy head is
proposed as a solution to increase the reliability of the
cluster head selection problem. However, this is a
centralized approach instead of distributed computing,
so it is not suitable for wide-area wireless sensor
networks.
Although LEACH has improved the lifetime of
the network, random selection can select CH with low
energy and lead to significant energy loss and energy
hole formation. Hence, variants of LEACH aim to
address this limitation by using more efficient
parameters to select CHs such as residual energy,
density, or distance to BS [10-12].
Instead of using a single parameter, some input
parameters are combined to make decisions in a fuzzy
method that gives positive results. These proposed
models have brought a good performance and a
reasonable complexity [13]. However, the above
studies are conducted for single-hop wireless sensor
networks, so they may not be suitable for large-scale
networks.
A remarkable approach in previous studies is to
use the entire fuzzy model for multi-hop
communication networks presented in [9, 16]. In this
study, the authors pointed out that the cause of the
network exhausted energy is that the CHs near the base
station is subjected to high transition loads; As a result,
their energy depletes much faster than other CHs. To
eliminate this phenomenon (hot spot), and unequal
clustering approach based on distributed fuzzy logic
and routing algorithm (DFCR) is proposed to solve this
In large-scale WSNs, multi-hop communication
requires not only an appropriate clustering process but
also an appropriate routing method. In [9], a routing
2
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
problem. After the clusters form, a fuzzy logic-based
goal for the sensor nodes. The main functions of the
FIS model are described below.
multi-hop routing algorithm is implemented for
relaying the aggregate data between the cluster heads
to the Sink node. Although hot spots are avoided, the
number of rules of the two fuzzy models causes an
increase in the problem complexity. Through these
previous studies, we recognize that besides
introducing an extra supported algorithm, the number
of rules can bring more complexity to improve
network lifetime.
-
Fuzzification: The fuzzification has the effect of
mapping each point of the input data to values in
the range [0, 1] or retaining the fuzzy values at the
input.
-
Mapping Rules: Represents the relationship
between input parameters and output parameters
based on certain rules.
Hence, to overcome these illustrated limitations,
a combined scheme of a lightweight clustering
problem and a simple greed routing algorithm is
proposed to prolong the lifetime of the wireless sensor
network. The numerical simulation results show a
clear advantage of the proposal compared to recent
studies.
-
-
Inference System: An inference system that maps
input and output according to fuzzy rules.
Defuzzification: The values obtained at the output
will be back-mapped for correct identification.
The basic fuzzy inference system is divided into
two types according to different principles: the one
based on the non-additive principle called the
Mamdani FIS model and the other based on the
additive model called the addition rule model Takagi-
Sugeno-Kang FIS. The main difference between the
two FIS models is the way sharp outputs are generated
from fuzzy inputs. The Mamdani FIS model uses the
output fuzzy technique that has fuzzy sets in the
consequent part. The TKS-FIS model uses a weighted
average to calculate the output through the linear
functions of the input variables [17].
3. The Proposed Model
3.1. Background
Fuzzy systems are an intelligent computational
approach and are used to model noisy and imprecise
environments. The biggest advantage of this approach
is that it is compatible with the real environment with
reasonable computational complexity because it is
based on imitating human reasoning. This is important
for many real-time applications and networks with
limited resources.
A rule in the Mamdani FIS type rule base has a
general form expressed by the formulation as
Fuzzy systems allow inputs with relative values
to give a smooth output function. A range of real
values described by fuzzy sets is called a domain, and
a membership function is defined. Each point in the
fuzzy set domain will be assigned a truth value. The
membership function can be one of a number of
triangular, trapezoidal, Gaussian functions with
domain values between 0 and 1.
(1)
If input x is A1 and input y is B1
Then output z is C1
A Sugeno FIS model type rule is in the form of
the statement expressed by
(2)
If input x is A1 and input y is B1
Then output z is z1= p1.x + q1.y + r1
where A1 and B1 are the values of the membership
functions associated with the input values x and y.
Due to the interpretable and intuitive nature of the
rule base, Mamdani FIS is widely used in particular for
decision support applications, artificial intelligence,
and so on. Hence, this study used a Mamdani FIS to
construct clusters based on several key input
parameters such as residual energy and distance to the
sink node. The input and output variables in the
Mamdani FIS model are characterized by triangular
membership functions.
Fig. 1. A typical fuzzy inference system model
In wireless sensor networks, the problem of
cluster formation and cluster head selection with input
variables such as residual energy, distance, or density
in a dynamic network are all uncertain factors.
Therefore, the fuzzy logic approach would be
advantageous to solve these problems. Proven through
previous studies [8-9], fuzzy logic is suitable to solve
the clustering for WSNs by calculating the choice in
many different dependent parameters. Fig. 1 shows the
general fuzzy inference system model. In the study,
property parameters of a wireless sensor network are
applied to function blocks according to the clustering
The network lifetime is dependent on the energy
consumption of whole sensor nodes. If the current
energy of the node is less, then the sensor node
becomes a dead node, resulting in broken links and
decreasing network lifetime. Therefore the energy of
the node should be a key factor of a network lifetime
problem as a part of input parameters in the study.
3
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
In this paper, we use the energy model [9], which
and ꢄꢇꢈ is the free space and multipath gain,
respectively, d0 is the calculated distance value
threshold:
is calculated based on the distance Dist(I,j) between
source and destination; free space channel fs, and
multipath channel mp. The energy consumed when
transferring data of l bits between node Si and Sj with
distance Dist(I,j) is determined by:
ε fs
(4)
d0 =
εmp
(Eelec +ε fs Dist(2i, j) )×l, Dist(i, j) < d0
ETX
=
The energy received during data reception of l
bits is calculated as:
(3)
(E +εmp Dist(4i, j) )×l, Dist(i, j) ≥ d0
elec
(5)
Er (Si , Sj ) = Eelec ×l
where, ꢀꢁꢂꢁꢃ is the 1-bit data transmission power, ꢄꢅꢆ
Fig. 2. Input parameters (a) Energy level; (b) Distance to the Sink node;
Fig. 3. Output probability of the Mamdani FIS
4
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
3.2. Clustering and Electing the Cluster Head
range of [10m 15m 20m 25m 30m]. In each hop, the
CH nodes will increase the cluster radius according to
the Fibonacci rule as follows:
To cluster and elect the cluster head, this study
uses two input parameters to reduce the complexity as
the same previous approaches [9, 16]. In [9], to
perform unequal clustering based on fuzzy logic, the
nodes that are likely to be candidates for cluster master
are selected through the fitness1 function. The
proposed FIS model used two input parameters such as
residual energy and distance to the sink node. The
input variables of fitness1 are shown below
f (1) =1, n =1
f (2) = 2, n = 2
(7)
f (n) = f (n −1) + f (n − 2),n∈[1,6]
The cluster radius of the CH node in each hop is
determined by the equation:
(8)
RCH = f (n)× β
The input parameters of the fuzzy set are as follows:
-
Energy level: The energy level of a Si node is the
ratio of the remaining energy to the initial energy.
Nodes with higher energy levels will be selected
as candidate nodes. Furthermore, as the energy
level decreases with time, the cluster radius is
adjusted to conserve its energy. The input energy
script of sensor node is presented in Fig. 2(a).
where β is a random variable in the interval [1, 1.5],
added to increase the accuracy of the cluster radius.
After calculating the burst radius, a CH will broadcast
a CH_ADVERTISE message within its cluster radius.
CH_ADVERTISE contains the energy level of the CH,
its location, and distance to the BS. If a node Sj receives
the message, it does not elect a CH and considers it a
non-CH node for that round. If a non-CH node receives
a CH_ADVERTISE message from multiple CHs, it
selects one of them based on cost. The cost function
CH_Cost(Sj, CHi) is defined as follows:
-
Distance: The path length from a Si node to the
BS root node is denoted DistBS(Si), this is to avoid
the problem of having too many CH nodes
located near the BS that have DistBS(Si) smaller
than the cluster radius. This input distance script
of sensor node to base station is presented in Fig.
2(b).
Dist(Sj ,CHi )× DistBS (CHi )
CH _ Cost(Sj ,CHi ) =
Energyres (CHi )
Table 1. If-then mapping rules for fuzzy fitness1
Clustering Algorithm
Rule Energy level
Distance
Near
Fuzzy fitness1
Medium
Weak
1.For each node Si do
2.Compute delay (Td(Si)) using Equation 3
3.End For
1
2
3
4
5
6
7
8
9
Small
Small
Moderate
Distant
Near
4.For each node Si do
Small
Very weak
Strong
5.If (Td(Si) == 0), ie, Delay time expires then
Average
Average
Average
High
6.Si becomes CH and calculates cluster radius using
Fibonacci algorithm and broadcasts
Moderate
Distant
Near
Medium
Very weak
Very strong
Strong
7. CH_ADVERTISE message within-cluster
radius
8. End If
High
Moderate
Distant
9. If Sj receives the CH_ADVERTISE message
from Si then
High
Weak
10. Sj update Si in Candidate(Sj) = [Si]
11. End If
After calculating through fitness1 as an outcome
of the FIS model illustrated in Fig. 3, each Si node will
give a value that is the delay time Td before self-select
the CH node, Td is determined by the following
formula:
12. End for
13. CH of Sj is the most residual energy in
Candidate(Sj)
(6)
Td (Si ) = α ×(1 − Fuzzyfitness1) ×Tc
3.3. Routing Scheme
where α is a random variable in the interval [0.9, 1],
added to minimize the delay of the two nodes and Tc is
the maximum waiting time when electing CH. When
the delay time expires, it selects itself as the CH node
and calculates the cluster radius. Then with the output
of fitness1, we will recalculate the cluster radius based
on the distributed rule of Fibonacci values. At the BS
node, we extend the distance to the sensor nodes in the
The routing process is based on the transmission
energy cost function and the CH node radius [9]. Each
CH node after being clustered will be assigned the
following level:
5
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
any node CHsi, when routing to the BS node, it will
DistBS (Si )
L(S ) =
(9)
open a scan angle with radius Rmax such that vector
[CHsi, BS] is the bisector of that angle. During the
scan, the CHsi node will find the CHsj nodes belonging
to that scan angle and determine the projection length
R(S) of that node down the line connecting the CHsi
and the BS node. After determining all the projection
lengths of the CH nodes in that angle, CHsi will choose
the CH node whose R(S) is the largest to transmit data.
Denote α is the scan angle, l is the scan length. Then,
the network continues to be routed to the BS node until
the distance from the intermediate CHsj node to the BS
i
Rmax
To determine the next CH nodes to the BS node,
each CH Si node will transmit a FIND message
containing information about its L(Si) level and its id
in the range k × Rmax (where k is initially 2). If a lower
level CH receives such a message, it returns an ACK
packet containing the id, energy, and distance
information to the BS. If no ACK is received within
the timeout from any of the lower CH, then Si
increments to k (k = 3, 4,…, L). If the lower-level CH
node does not exist, it will transmit directly to the BS,
the routing process will be complete.
is less than Rmax
.
(10)
R(Sj ) = max{cos(α)×l(Sj )}
The Greedy algorithm uses vector concepts to
construct routing strategies as illustrated in Fig. 4. At
Fig. 4. Greedy routing algorithm
Fig. 5. (a) BS at the edge; (b) BS at the central
6
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
Greedy routing Algorithm
Function Greedy(Rmax, angle, finish_id,start_id)
1. vector_to_finish_id = vector(start_id, finish_id)
2. Infr_path = []
3. While true
4. If length(vector_to_finish_id) ≤ Rmax
5. update finish_id to Infr_path
6. break
7. End If
8. For each Si in all nodes
9. vector = vector(CH, Si)
10. cosine = cos(vector,vector_to_finish_id)
11. If(length(vector) ≤ Rmax and cosine ≤ cos(angle))
12. R(Si) = max(cosine*length(Si))
13. End If
14. End For
15. update Si to Infr_path
16. Start_id = Si.id
17. End While
18. return Infr_path
End Function
Table 2. Parameters and values
Parameter
Symbol Value
100×100 m2
Network Size
BS location
A
BS
N
(50, 50); (50, 0)
100
Number of nodes
Initiation energy
Cluster radius
Energyinit 0.5 J
Rmax
30 m
Transmission range Rs
10 m
Packet Size
Dp
εfs
500
Free space coefficient
10 pJ/m2/bit
0.0013 pJ/bit/m4
Fig. 6. Comparison of network life cycle results:
Multipath delay
factor
εmp
(a) Comparison of FND; (b) Comparison of HND;
(c) Comparison of AND
Tx and Rx signal
Etx or Erx 50 nJ/m2/bit
4. Numerical Results
In this section, we conduct numerical simulations
performed on MATLAB with input parameters as
previously proposed [8, 9]. The BS base station is
located at two locations (center and edge) as illustrated
in Fig. 5, and the input parameters are presented in
Table 2.
To save energy, the routing algorithm considers
the energy required to transmit messages between CH
nodes. After determining the total energy consumed by
the CH node, the routing will go through the cost
function to determine the path to the BS node. The cost
function will increase when the remaining energy of
the node is low, so it must be routed to another node
with higher residual energy.
The criteria for evaluating network performance
through the lifetime of the network include: the first
node’s energy depleted (FND : First Node Die) when
7
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
Fig. 7. Comparison of the number of alive nodes: (a) The BS is at the edge; (b) The BS node is at the center
half of all nodes are exhausted (HND: Half Node Die)
and all nodes are exhausted (AND: All Node Die). The
AND case will likely not occur, but we include it only
to calculate the relative life of the network. A node is
considered dead (out of energy) when it has no energy
left to receive or transmit data. The main numerical
results are presented in Fig. 6 for comparing the life of
a sensor network in general cases. Fig. 7 presents two
general cases with different position (a) for the base
station located at the edge examples network and (b)
for the center of the simulation network.
proposed and proven through simulation. The
validated numerical results show the ability to operate
efficiently and respond well to the uncertain conditions
of the multi-hop wireless sensor network. From the
above initial results, cooperative calculations between
cluster heads will be our next research.
References
[1]. C. V. Mahamuni, A military surveillance system based
on wireless sensor networks with extended coverage
life, in Proc. ICGTSPICC2016, 2017, pp. 375–381.
5. Conclusion
[2]. D. K. Rathinam, D. Surendran, A. Shilpa, A. Santhiya
Grace, and J. Sherin, Modern agriculture using wireless
sensor network (WSN), in Proc. ICACCS 2019, 2019,
pp. 515–519.
The lifetime of a multi-hop wireless sensor
network has always been a practical criterion for
evaluating network performance. A simple and
efficient multi-hop routing scheme has been proposed
in this paper, the intelligent computational approaches
based on fuzzy logic provide effective solutions for the
clustering process and cluster head selection. To avoid
energy holes by efficient routing algorithms between
cluster heads, the proposed algorithm has been
[3]. M. G. R. Maldonado, Wireless sensor network for smart
home services using optimal communications, Proc. -
INCISCOS 2017, vol. 2017-November, pp. 27–32.
8
JST: Smart Systems and Devices
Volume 31, Issue 2, September 2021, 001-009
[11].R. Priyadarshi, L. Singh, Randheer, and A. Singh, A
[4]. B. S. Kim, K. Il Kim, B. Shah, F. Chow, and K. H. Kim,
Wireless sensor networks for big data systems, Sensors
(Switzerland), vol. 19, no. 7, pp. 1–18, 2019.
novel HEED protocol for wireless sensor networks, in
Proc. SPIN2018, Noida, India, 2018, pp. 296–300.
[12].I. Sharma, R. Singh, and M. Khurana, Performance
evaluation of PEGASIS protocol for WSN using NS2,
In Proc. ICACEA 2015, 2015, pp. 926–929.
[5]. A. O. Abu Salem and N. Shudifat, Enhanced LEACH
protocol for increasing a lifetime of WSNs, Personal
and Ubiquitous Computing, vol. 23, no.5–6, pp. 901–
907, 2019.
[13].[1] Z. Siqing, T. Yang, and Y. Feiyue, Fuzzy logic-
based clustering algorithm for multi-hop wireless
sensor networks, Procedia Comput. Sci., vol. 131, pp.
1095–1103, 2018.
[6]. A. R. Khan, N. Rakesh, A. Bansal, and D. K.
Chaudhary, Comparative study of WSN protocols, In
Proc. 2015 Third International Conference on Image
Information Processing (ICIIP), Waknaghat, India,
2015, pp. 422–427.
[14].T. Bhowmik and I. Banerjee, Dynamic PSO based
fuzzy clustering algorithm for WSNs, in Proc.
TENCON 2019, 2019, pp. 1992-1997.
[7]. E. Farahmand, S. Sheikhpour, A. Mahani, and N.
Taheri, Load balanced energy-Aware genetic algorithm
clustering in wireless sensor networks, in Proc. CSIEC
2016, 2016, pp. 119–124.
[15].S. Lata, S. Mehfuz, S. Urooj and F. Alrowais, Fuzzy
clustering algorithm for enhancing reliability and
network lifetime of wireless sensor networks, in IEEE
Access, vol. 8, pp. 66013-66024, 2020.
[8]. M. Ahmed, Energy efficient distributed clustering and
scheduling algorithm for wireless sensor networks with
non-uniform node distribution, Int. J. Adv. Res.
Comput. Sci., vol. 9, no. 2, pp. 527–532, 2018.
[16].Kiran, W.S., Smys, S. & Bindhu, V. Enhancement of
network lifetime using fuzzy clustering and
multidirectional routing for wireless sensor networks,
Soft Computing 24, pp. 11805–11818, 2020.
[9]. N. Mazumdar and H. Om, Distributed fuzzy approach
to unequal clustering and routing algorithm for wireless
sensor networks, Int. J. Commun. Syst., vol. 31, no. 12,
2018.
[17].Bagis, A. and Mehmet Konar. Comparison of Sugeno
and Mamdani fuzzy models optimized by artificial bee
colony algorithm for nonlinear system modelling.
Transactions of the Institute of Measurement and
Control 38, pp. 579 – 592, 2016.
[10].J. P. A. Pirishothm, Study of wireless sensor networks
using leach, teen and apteen routing protocols, Int.J.Sci.
Res., vol. 4, no. 5, pp. 1221–1224, 2015.
9
9 trang
16480
Free
Bạn đang xem tài liệu "A study on multi-hop routing scheme for wireless sensor networks", để tải tài liệu gốc về máy hãy click vào nút Download ở trên
File đính kèm:
- a_study_on_multi_hop_routing_scheme_for_wireless_sensor_netw.pdf
