Received: 4 May 2017 Revised: 13 July 2017 Accepted: 7 August 2017

DOI: 10.1002/dac.3407

RESEARCH ARTICLE

A new algorithm for cluster head selection in LEACH

protocol for wireless sensor networks

Ahmed Al‐Baz | Ayman El‐Sayed

Department of Computer Science and

Summary
Engineering, Faculty of Electronic
Engineering, Menoufia University, In wireless sensor network, a large number of sensor nodes are distributed to
Menouf 32952, Egypt cover a certain area. Sensor node is little in size with restricted processing
Correspondence
power, memory, and limited battery life. Because of restricted battery power,
A. El‐Sayed, Department of Computer wireless sensor network needs to broaden the system lifetime by reducing the
Science and Engineering, Faculty of energy consumption. A clustering‐based protocols adapt the use of energy by
Electronic Engineering, Menoufia
University, Menouf, 32952, Egypt. giving a balance to all nodes to become a cluster head. In this paper, we concen-
Email: [Link]@el‐[Link]. trate on a recent hierarchical routing protocols, which are depending on
[Link]
LEACH protocol to enhance its performance and increase the lifetime of wire-
less sensor network. So our enhanced protocol called Node Ranked–LEACH is
proposed. Our proposed protocol improves the total network lifetime based on
node rank algorithm. Node rank algorithm depends on both path cost and
number of links between nodes to select the cluster head of each cluster. This
enhancement reflects the real weight of specific node to success and can be rep-
resented as a cluster head. The proposed algorithm overcomes the random pro-
cess selection, which leads to unexpected fail for some cluster heads in other
LEACH versions, and it gives a good performance in the network lifetime
and energy consumption comparing with previous version of LEACH protocols.

KEYWORDS
cluster head selection, clustering‐based protocols, LEACH protocols, wireless sensor networks

1 | INTRODUCTION

Wireless sensor networks (WSNs) consider a base station (BS) that communicates with countless of wireless sensor nodes
by means of a radio channel. Data transmission and reception process are the major operations that consume power in
WSN. Elimination of redundant data is important issue in WSN, because multiple nodes may sense the same overlapped
area. This action will reduce the overall power consumption and use the bandwidth usage.
Routing protocols in WSNs can be divided according to network structure into 3 categories: location based, data
centric, and hierarchical protocols.1,2 Location‐based protocol: location information for all nodes is a major cornerstone
during data communication. Energy consumption during data transmission depends on the target position distance.
Data are sent not only to the target area but also to whole sensors network, which is known as flooding. Residual energy
is not considered for each node, so many nodes will run out of energy, which leads at the end to an accumulative data
transmission failure.3-5 Data‐centric protocol6: data transmission from sensor node to the BS through intermediate
nodes. The intermediate node may not have interest in the data it receives—it is known as relay node in this case. This
process reduces the number of packet sent from source to the sink, which leads to reduce the power consumption.

Int J Commun Syst. 2017;e3407. [Link]/journal/dac Copyright © 2017 John Wiley & Sons, Ltd. 1
[Link]
2 AL‐BAZ AND EL‐SAYED

Redundant data can be eliminated during the negotiation process among nodes, and nodes activities can be monitored to
report current state energy level. Hierarchical routing protocol: firstly, it is used in wired networks to save power then
used in WSN to prolong network lifetime by using power consumption. Higher energy sensors are used for data trans-
mission, while rest nodes do them traditional work in sensing target area. So nodes can be divided into groups, some
of them are control nodes and the other are member's nodes.4,5,7
This paper focuses on the hierarchical routing protocol that is based on clustering algorithm especially low energy
adaptive clustering hierarchy protocol (LEACH) that expands total network lifetime, and it is the first hierarchal energy
adapted protocol. Unexpected node death can be avoided in LEACH protocol by cluster head rotation process, which
alleviates the power consumption load on a certain node all the time.
However, LEACH has a considerable defect that is “cluster head selection process relies on probability method which
leads to additional overhead and power consumption increase.” LEACH can do its work well in homogeneous sensor
networks, but a weakness is noticed in heterogeneous sensor networks because the low‐residual energy nodes will die
before high‐residual energy nodes.
Data sensed from a target area is compressed at the sensor node (SN) and is transmitted either specifically directly to
the BS or through intermediate sensor nodes. The BS plays an important role that data from the wireless sensor system
will be assembled and prepared and might interact with the wireless SN using internet or satellite. Power consumption
cost is minimized to acceptable value through data transfer, considering the efforts exerted in clustering algorithm
research area.
In LEACH, wireless sensor nodes are arranged into groups to organize data transmission, and each group is
controlled by its cluster head (CH). In each cluster, the role of CH is exchanged among cluster nodes to save path cost.
So CH may suffer from energy overload that can be out of service in unexpected time. Also, CH re‐election process is
required among all nodes to achieve load balance in the sensor network. Data aggregation process aims to combine
and summarize the data packets so that the total data transmitted is reduced. Data aggregation process also has been
used to accomplish energy productivity and data transfer optimization in several routing protocols.8,9
LEACH splits the sensor nodes into small clusters and chooses one of them as a CH. LEACH uses randomized turn of
CH role among sensor nodes to evenly distribute the energy load to abstain from depleting the battery of any one sensor
in the system.8 The BS occasionally changes both the cluster membership and CH to ration energy. In other hand, CH
selection preformed randomly, so in many situations, some nodes that have less energy can be chosen as a CH, which
causes cluster struggle. Many improvements are performed on LEACH protocol to handle selection of CH obstacles.
LEACH operation depends on forming sensor nodes clusters in perspective of got signal strength and use neighborhood
CHs as switches to the sink.8-10 So this paper proposed a new algorithm to select the CH of each cluster, with considering
the increase of the SNs' lifetime and the improvement of the WSN performance.
Our proposed protocol is called Node Ranked–LEACH (NR‐LEACH) that enhances the performance of the LEACH
protocol. This enhancement is based on distributing the energy load among the sensor nodes when electing CH. This is
by using node rank algorithm that improved the total network lifetime. Node rank algorithm (NRA) depends on path
cost between nodes and number of links between nodes to elect CH. This enhancement reflects the real weight of specific
node to success and can be represented as CH. The proposed algorithm overcomes the random process selection, which
leads to unexpected failure for some CHs in other LEACH versions.
This paper is organized as follows: Section 2 describes the related works; this section discusses the hierarchical
routing protocols, which concerns with cluster head selection algorithms. Section 3 depicts the LEACH protocol and
its operation, which discusses hierarchical routing protocols research point concerned with cluster‐head selection mech-
anism and explains LEACH‐E and I‐LEACH which our proposed algorithm compared with. Section 4 outlines our pro-
posed algorithm in view of node rank determination. Section 5 describes the evaluation performance of our proposed
protocol with comparing with the other LEACH conventions. Finally, the paper is concluded in Section 6.

2 | RELATED WORKS

Many researches considered LEACH is the best protocol that save power and prolong the WSN lifetime. So LEACH ver-
sions have produced to benefit from LEACH characteristics and reduce its drawbacks. There are some versions of
LEACH protocols such as follows: LEACH‐C, Centralized,11,12 in which each node location is a vital information for
BS to run a centralized clustering algorithm to organize the clusters for current round. Most suitable nodes will be
selected to represent a CH, and the non‐CH nodes will be distributed as cluster member nodes. The first dead node
AL‐BAZ AND EL‐SAYED 3

metric enhancement occurred in LEACH‐C compared with traditional LEACH, but LEACH‐C is not suitable for large
scale networks. Power‐efficient gathering in sensor information systems, PEGASIS13 chooses the brief approach to
choose the neighbor node as CH, using BS data. PEGASIS is not self‐affiliation. Each node in PEGASIS can handle wire-
less communication, data positioning, and fusion. Data are sent from each node to its neighbor in the chain until arriving
to BS via chain head. Energy consumption is reduced during data transmission process.
In energy‐efficient clustering algorithm for event driven, EECED,14 all nodes considered to be stationary. Residual
energy is an operating condition factor in CH selection. That component can keep up a balanced energy usage movement
among different nodes to prolong the network lifetime, but results in extra overhead on BS. For mobile nodes, the mobil-
ity is considered with residual energy when CH selection in LEACH‐M, Mobile.15,16 LEACH‐M reduces the packet loss
occurred in mobility model, but a number of packets are lost during the CH selection for the next round of CH selection.
LEACH‐GA17 protocol is a genetic algorithm grouping system. Every node sends its node identifier, area data, and
CH choice in view of ideal rate of CH to the BS. The BS applies genetic calculation operations on got data to discover
the ideal limit likelihood and communicate it for group arrangement. Extra overhead is produced to calculate a desired
percentage of CHs.
In TL‐LEACH (2‐levels),18 CH collects data from other members in its cluster as original LEACH, but not transfer the
data to the BS directly. This concedes to better use of energy in WSN. Extra overhead is use to choose secondary CHs. TL‐
LEACH demonstrates an extraordinary upgrade of the energy consumption and system lifetime contrasted compared
with traditional LEACH. Depending on Vice‐LEACH19,20 protocol, data aggregated by the CHs will constantly arrived
the BS. Thusly, there is no convincing motivation to select another CH each time. The network lifetime extended, but
an overhead based on CH selection arises.
Multi‐hop LEACH10,21 protocol recommended to handle the power consumption load. All nodes are immobile, and
BS is fixed. The process of CH selection is based on random selection. It gets advantage of multihop technique to
minimize the total power consumption. Critical issue is denoted when one or more intermediate CHs died or failed.
In MELEACH‐L22 protocol, all nodes have the same energy level, immobile, and randomly distributed. Selection of
CH is based on residual energy. Economic consideration increased, applied thanks to multiple channel usage while
channel assignment among cluster groups problem is solved. Moreover, the energy efficiency is further improved, and
the framework lifetime extended.
In Enhanced‐LEACH (E‐LEACH),23 cost capacity calculation in perspective of the remaining energy and their
separation from sink exists in the process of CH selection. The results demonstrate that E‐LEACH calculation is a great
deal and more proficient. E‐LEACH calculation expands sensor lifetime, throughput, and the sensors power utilization,
but extra overhead exists.

3 | LEACH PROTOCOLS

Heinzelmon introduced a hierarchical clustering algorithm for sensor networks. LEACH is a self‐arranging, clustering
convention and based on round‐based technique. LEACH expect that the BS is settled and arranged far from the sensors,
all sensors are homogenous and have confined energy source, sensors can sense the earth at a changed rate and can
grant among each other, and sensors can particularly compare with BS.
The considered LEACH is to deal with the nodes and divides them into groups to distribute the energy among the
sensor nodes in the network, and in each group, there is a control node called a CH as shown in Figure 1. In Wu and
Wang11 and Fu et al,24 LEACH procedure is indicated. Each round in LEACH involves 2 phases. The process of cluster
set‐up, advertisement, and schedule creation phase forms setup phase.
At the beginning of the setup phase, every node picks a random number some place around 0 and 1, and after that
figures an edge condition.25 If the picked subjective number is less than the threshold numberT(n), the node becomes a
fortunate CH for such round. Threshold number T(n) is shown in Equation 1.
8
> P
>
<
if n ϵ G
1
T ðnÞ ¼ 1−P× r mod ; (1)
>
> P
:
0 otherwise

where P is the CHs desired percentage, r is the current round, and G is the competing nodes that was not chosen as CHs
in the last 1/P rounds. Node gets to be CH for the current round if the number is not as much as limit T(n). When node is
4 AL‐BAZ AND EL‐SAYED

FIGURE 1 Illustration of LEACH protocol

chosen as a CH, then it cannot get to be CH again until every one of the nodes of the group has gotten to be CH once.
This is valuable for adjusting the energy utilization. Steady state is the second phase; non‐CHs get the CH requests and
after that send join demand to the CH advising that they are individuals from the group under that CH.7 During the
steady‐state phase, each sensor node aggregates and transmits data to its CH in perspective of the TDMA schedule.
TDMA/CDMA MAC is used in LEACH protocol to resist intercluster and intracluster collision.26 The CHs get each
one of the data and aggregate it before being sent to the BS. After a time, which is determined from the before, the frame-
work starts another round by withdrawing to the setup and persisting state arranges yet again.7,26
There are several modifications to solve the problem of CH selection such as LEACH‐E, and I‐LEACH is explained in
the following subsections.

3.1 | LEACH‐E
A progressive algorithm for the CH selection based on the remaining energy and the distances is proposed in
Nawar et al.27 Take residual energy and distance into CH election question of life and death for WSN lifespan.
Inharmonic distribution of nodes leads to dense and slight clusters in the same network at the same time, which
affirms that some CHs may be loaded more than other sensors. In the setup phase, residual energy and distance
are considerable factors in calculation the value of T(n). The cluster head will be rotated by computing both dis-
tance and residual energy of nodes and select the node with maximal parameter. Enhanced LEACH (LEACH‐E)
in,27 cluster head selection is based on distance and residual energy. LEACH‐E calculates decision variable (dij)
as shown in Equation 2.
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ΔEavg
dij ¼ ∝sij þ ðEi −Em Þ2 ; where ∝ ¼ ;
2
(2)
ΔSavg

where Ei is the initial energy for each node, Em is the remaining energy of cluster head j, Sij is the distance
between the node i and the cluster head j, and dij is the node i decision value for cluster head j. This is fair load
distribution among the nodes in one cluster.27 However, it is not satisfied for the good performance metrics, so the
decision variable has to be verified again.

3.2 | I‐LEACH
In Improved LEACH (I‐LEACH),28,29 authors considered both the distance and residual energy as well to calculate the
decision value by Equation 3, where α is the average of the first‐order differences of the energy consumed and β is the
average of the first‐order differences of the distance. By Equation 3, CH can be selected to improve the performance
of ELEACH.
 2
dij ¼ ∝ sij þ β ðE i −E m Þ2 ; where∝ ¼ ΔE avg ; β ¼ ΔSavg : (3)
AL‐BAZ AND EL‐SAYED 5

ILEACH is an optimization equation 3, which leads to more improvement on packet delivery ratio, average delay,
throughput, and number of nodes alive. At the beginning of each round, each node receives the residual energy from
all the neighbors in its cluster range. Cluster head is selected based on its weight calculated based on its residual energy
and the received signal strength for the signal broadcasted by the sink. The proposed system will improve the network
life and total communication over the network. The selection of cluster head node in LEACH has serious defect that
leads to unreasonable CH selection that the nodes have different energy while big clusters and small clusters may exist
in the network at the same time.

3.3 | Our proposed protocol

Our proposed protocol is called Node Ranked–LEACH (NR‐LEACH) that improves the performance of the LEACH
protocol. This improvement depends on how to select CH to make the energy load of sensor nodes are distributed in
a good manner among all sensor nodes, to increase the total network lifetime. This is by using node rank algorithm that
gives a rank value of all sensor nodes. Node rank algorithm (NRA) depends on path cost between nodes and number of
links between nodes to elect CH. This enhancement reflects the real weight of specific node to success and can be
selected as CH. The proposed algorithm overcomes the random process selection, which leads to unexpected failures
for some CHs as in the other LEACH versions.

4 | OUR P ROPOSED P ROTOCOL: NODE RANKED– LEACH

Our proposed protocol is called Node Ranked–LEACH (NR‐LEACH) that enhances the performance of the LEACH pro-
tocol. This enhancement is based on distributing the energy load among the sensor nodes by using node rank algorithm
(NRA) used to elect the CH. The operation of our proposed protocol is divided into rounds. Forming clusters is per-
formed in setup phase which is followed by steady‐state phase where data can be transmitted among cluster member
nodes, CH and BS. In the setup phase, the process of CHs election in the proposed algorithm is depending on node rank
calculation. CH election is based on evaluation of each node by calculating its weight. LEACH protocol has the efficiency
to work well in networks that have big density but limited ability to be used in large‐scale network. Weight of a node in
our proposed protocol is controlled by 3 factors: received signal strength, residual energy for each node, and link connec-
tions number with other nodes. Distance between sensor nodes and BS and distance among nodes itself influence the
energy consumed. These 3 factors reflect the importance of node that can serve large number of nodes successfully in
the network and deserved to elect to be one of CHs group.25,30
Unfair distribution on large number nodes leads to one CH that manages big cluster members; some clusters have
small number of cluster members, and the rest of the nodes will not be covered by any CH control. CH that serves large
number of sensor nodes will be loaded and die rapidly in comparison with CH that has limited nodes. Also, random
selection of CH which results from giving all sensors in the network the same chance to be elected as CH will cause that;
a node having less energy may be future CH and will be out of service soon. Recent surveys performed on LEACH ver-
sion highlighted that CH selection is based on residual energy or nodes position far from BS.25,28,31,32
These ways used to select CHs produced an extra overhead that resulted from additional data transmission, which
consumes more power. These shortcomings are successfully handled in our proposed algorithm. BS has an accurate
information about each node such as its position, current energy, and connections available to other nodes. In our
proposed algorithm, each node is ranked according to its position, residual energy, and its connection links.
The selection of CH is based on a rank function. A node that has a biggest rank will be selected as CH.

4.1 | Node rank algorithm (NRA)

Our proposed algorithm is divided into 2 phases: setup phase where clusters organized followed by steady‐state phase
where data can be transmitted to BS. In setup phase, each node understands if it will act as a CH or a cluster member
for present cycle. The election of CH is based on algorithm called node rank algorithm (NRA). Node rank score for node
i, NR(ni), is calculated from Equation 4, NH is the set of neighbors for the k, djiout is the distance of out link edges
from node j to node i, PO(ni) is the current energy of node i , and α is the damping factor. Recent studies have checked
many damping factors (α) confirming that the suitable factor is 0.85 as mentioned in El‐Fishawy et al33 for Arabic
summarization in Twitter social network.
6 AL‐BAZ AND EL‐SAYED

  1= d jiout
NRðni Þ ¼ POðni Þ α ∑0 NR nj
j
þ ð1−αÞ: (4)
∑k∈NH d jkout

A node is considered important if it is pointed (linked) to by other important nodes. An initial node rank is assigned
for each node =1. Repeat the calculation for each iteration according to an iterative algorithm equation 4 until arrive the
convergence score. The calculation of a node's node rank achieved without defining the final value for node rank (NR)
for other nodes. In every iteration, a closer weight of a final value is achieved.
Figure 2 illustrates the concept of link connections importance factor: mathematical node ranks for a simple
network, expressed as percentages. Node C has a higher node rank than Node E, even though there are fewer links to
C; the one link to C comes from an important node (B) affects the total weight of the node C and hence is of high value
according to other factors residual energy, distance which will be considered according to node rank calculation. In the
presence of damping, the node rank computations require several passes, called “iterations,” through the collection to
adjust approximate node rank values to reflect the theoretical true value more closely. NRA algorithm archived the result
value each time it calculated and recalculate many times until the values stop changing. Now, the node that has the
biggest NR value will be selected as CH for the current round. CH node informs its members by small advertisement
message according to CSMA MAC protocol. A node will reply with an acknowledgement message that informs the
CH its interests to be a CH member.
CH allocates a TDMA schedule to active members to manage the data transmission and save power consumption
according to active/sleep status. Data transmission occurred in steady‐state phase respect to time slot allocated to each
member node. All workless nodes will turn off its power except CH waits sensed data from its member nodes. After data
delivery from nodes, CH aggregates the data and then transit the aggregated data to BS.
After finishing the steady‐state phase, the algorithm will return to setup iterations phase to elect another CHs group
and followed by steady‐state phase, etc. Our proposed algorithm avoids the random selection problem of cluster head by
replacing the random selection with selection based on node rank, which prolongs the overall network lifetime.
WSN consists of N sensor nodes; cluster head election is calculated according to Node Rank algorithm, which is
discussed in algorithm NRA shown in Figure 3.

4.2 | Node rank convergence scores

In Table 1, it is noticed that after the iteration number 50, the node rank for the nodes gets stable and no change
occurred. Node rank of N1 is higher than node rank of N2 and other nodes. It is because node N1 has more incoming
connections and out coming connections than other nodes. Each node has its own rank prestige; this rank increased also
when the node connected with another node that having a high rank score and so on. Overall in the network, each node

FIGURE 2 Link connections importance factor

AL‐BAZ AND EL‐SAYED 7

FIGURE 3 Node ranked algorithm

TABLE 1 Node rank convergence scores

Iteration N1 N2 N3 … N100

1 1.657 1.3987 1.2243 … 0.934

2 1.343 1.223 1.1123 … 0.897
3 1.223 1.112 1.089 … 0.776
4 1.220 1.065 1.231 … 0.636
5 1.192 1.044 1.204 … 0.556

49 1.21 1.0125 0.9879 … 0.808

50 1.21 1.0125 0.9879 … 0.808

affected by all nodes rank connected with it directly or inherited from indirect connections. Moreover, from Table 1, the
convergence of node rank calculation is shown for wireless sensor network that contains 100 nodes, and also this
scenario will be repeated for networks with 10, 25, 50, 75, 150, 200, and 250 nodes.

5 | PERFORMANC E EVALUATION

5.1 | Description of the simulator

All following results in this paper are gotten using NS234 simulation program. NS2 is used to assess the execution of
original LEACH, LEACH‐E, I‐EACH, and our proposed protocol (NR‐LEACH). WSN range size in meter is
600 × 600; the underlying energy of every node in the WSN is accepted as 0.5j, and the amount of sensor nodes differs
from 10 to 250; 20% of sensor nodes for each round get to be CHs, and the following metrics are used to evaluate the
WSN lifetime and its performance.

5.2 | Evaluation metrics

In this study, one metric can use throughput, normal end‐to‐end delay, monitor the packet delivery ratio, normal energy
proportion, and number of alive nodes for protocol assessment.
8 AL‐BAZ AND EL‐SAYED

(1) Throughput (Thr): The proportion of packets received (Nrecv) by a receiver from a transmitter for a period (T)
needed for getting the last packet by receiver measures in Kilo bit per seconds (kbps), where Pktsize is the packet
size in bits. It can be communicated numerically as appeared in Equation 5.

N recv PKT size

Thr ¼ (5)
T

(2) Average end‐to‐end delay of data packets (Davg): is the average delay between the packet source and its receipt at the
corresponding received. Equation 6 portrays the normal end‐to‐end data packets delay in milliseconds.

N T recv ðiÞ−T send ðiÞ

Davg ¼ ∑i¼0 (6)
N recv

(3) Packet delivery ratio (PDR): The proportion of packets successfully received (Nrecv) by the sink to the total number of
data packets sent (Nsend). PDR (%) is shown in Equation 7.

N recv
PDR ¼ x100 (7)
N send

(4) Average energy ratio (AER): Represented as the average energy consumed (Econsumed) at each node divided by the
total Energy available (Eavailable) for simulation time. AER (%) is evaluated as shown in Equation 8.

E consumed
AER ¼ x100 (8)
E available

(5) Number of alive nodes (Nalive): Number of nodes still alive after scenario finishing or during the simulation time.

5.3 | Simulation results

(1) Throughput (Thr)

In NR‐LEACH, each cluster has its cluster head based on link connections number with other nodes, residual energy,
and distance of node from BS. As shown in Figure 4 that portrays the throughput versus number of sensor nodes. It is
expanded with number of nodes build because of expanding the measure of information sent to sink. Noted additionally

FIGURE 4 Throughput versus number

of nodes for different LEACH protocols
AL‐BAZ AND EL‐SAYED 9

that our proposed protocol has a decent throughput with contrasting and different sorts of LEACH conventions. This is
because there is a heap parity of the CH part dispersion among the nodes in one group by considering both the
energy devoured of every node and the separation among nodes in bunch. Also, the throughput is improved because
of decreasing the packet loss.

(2) Average end‐to‐end delay of data packets (Davg)

Our proposed protocol is successful for distribution of CHs and does not require a large communication overhead
and reducing transmission of redundant data to cluster heads. So it reflects the enhancement of average end‐to‐end delay
as depicted in Figure 5 that shows the (Davg) versus number of sensor nodes. This proposed protocol has a minimum
delay with comparing with both the LEACH protocol versions because of considering the distance and residual energy
when selecting the CH.

(3) Packet delivery ratio (PDR)

NR‐LEACH has the best packet delivery ratio comparing with LEACH protocols as noted in Figure 6 that delineates
the PDR versus number of sensor nodes. PDR decreases with increasing the number of nodes because of the fixed
available bandwidth in the network and the worked area. Our proposed protocol has the best packet delivery ratio
comparing with LEACH protocols because there is a better apportionment of the nodes and a fair CH selection for each
cluster, which makes flow of data from CH to BS better.

FIGURE 5 Average delay versus

number of nodes for different LEACH
protocols

FIGURE 6 PDR versus number of

nodes for different LEACH protocols
10 AL‐BAZ AND EL‐SAYED

(4) Average energy ratio (AER)

NR‐LEACH prevents the random selection and uneven distribution of cluster heads as random selection of cluster
heads result in cluster heads being spread in a centralized area, thus leading to long transmission distance and wastage
of energy. So NR‐LEACH decreases the average energy consumption of nodes. Figure 7 delineates the (AER) versus
number of nodes. Noticed that the AER marginally increments with expanding the quantity of nodes. Likewise noticed
that our proposed protocol has a most minimal energy. Since the part of CH is appropriated among nodes in great way.
Also, our proposed algorithm consumes low energy because big rank node which considers a distance as the main factor
in its calculation will handle the problem of less energy CH be implicated in data transmission process which consumes
a large amount of energy. In this way, it reduces the energy dissipation in communication and achieve the purpose of
energy saving of the sensor nodes.

(5) Number of alive nodes (Nalive)

Nodes remain alive for long time for NR‐LEACH because its CH selection algorithm considers link connections
number with other nodes, residual energy, and distance of node from BS. Figure 8 depicts the (Nalive) versus number
of nodes. It is noted that our proposed protocol increases the nodes lifetime comparing both the original LEACH,
LEACH‐E, and EB‐EACH protocol. Cluster‐based protocols over node rank algorithm increases the total network life-
time as shown in this figure. This enhancement results from the selection of an optimized CH for every cluster, which
represents the actual weight and more suitable node to be CH. In addition, the clustering overhead is reduced, which
assists nodes to be alive more than before. So extra duties of cluster heads that increase energy usage results in node
death will decreased.

FIGURE 7 AER versus number of

nodes for different LEACH protocols

FIGURE 8 Number of nodes alive

versus number of nodes for different
versions of LEACH protocols
AL‐BAZ AND EL‐SAYED 11

5.4 | Discussion
One of the previous solutions of CH selection is a distance‐aware algorithm and some of LEACH versions are consider-
ing both distance and residual energy of each node. For the later, the weights are adapting of both distance and residual
energy to make more efficient and long‐life time such as LEACH‐E and I‐EACH. To prolong lifetime of wireless sensor
network considering another parameter namely link connections number with other nodes. Because the available num-
ber of connection to other nodes increases, it makes more processing resulting more power consumption. In NR‐
LEACH, to prolong network lifetime of wireless sensor network, it is important to rank the nodes according to 3 factors:
distance factor, residual energy for each node, and link connections number with other nodes.
In the original LEACH protocol, the nearest neighbor is chosen by measuring the received signal strength (RSS); the
highest RSS is corresponding to the nearest neighbor. Also, in our algorithm (NR‐LEACH), RSS with the residual energy
for each node and link connections number with other nodes are considered to rank the nodes.
In NR‐LEACH, each cluster has its cluster head based on link connections number with other nodes, residual energy,
and distance of node from BS. This results in improved throughput by decreasing the packet loss. This algorithm is
successful for distribution of CHs and does not require a large communication overhead also, reducing transmission
of redundant data to cluster heads. It reflects the enhancement of average end‐to‐end delay. NR‐LEACH has the best
packet delivery ratio comparing with LEACH protocols. Because there is a better apportionment of the nodes and a fair
CH selection for each cluster, which makes flow of data from CH to BS well, NR‐LEACH prevents the random selection
and uneven distribution of cluster heads as random selection of cluster heads result in cluster heads being spread in a
centralized area, thus leading to long transmission distance and wastage of energy. So NR‐LEACH decreases the average
energy consumption of nodes. In this way, it reduces the energy dissipation in communication and achieve the purpose
of energy saving of the sensor nodes. Nodes remain alive for long time for NR‐LEACH because its CH selection
algorithm considers link connections number with other nodes, residual energy, and distance of node from BS. So extra
duties of cluster heads that increase energy usage results in node death will decreased.
It is noted that NR‐LEACH increases the nodes lifetime comparing with other LEACH versions. This enhancement
results from the selection of an optimized CH for every cluster, which represents the actual weight and more suitable
node to be a CH.

6 | CONCLUSION

Clustering is an effective technique that can participate to entire system lifetime, scalability, and energy efficiency in
WSNs. Low energy adaptive clustering hierarchy protocol is a very effective protocol that enhances the network lifetime.
A clustering‐based protocol balances the energy usage by giving equal chance to all nodes to become a cluster head. This
paper proposed a new version of LEACH protocol called Node Ranked–LEACH (NR‐LEACH) that is enhanced to
improve the total network lifetime based on node rank (NR) algorithm. Node rank algorithm depends on path cost
between nodes and number of links between nodes to elect cluster head. This enhancement reflects the real weight of
specific node to success and can be represented as cluster head. The proposed algorithm overcomes the random process
selection, which leads to unexpected failure for some cluster heads in other LEACH versions, and it gives a good
performance in the network lifetime and energy consumption. Our proposed protocol increases both sensor lifetime,
throughput, and the delivery packet ratio and decreases both the packet delay and the sensors power consumption.
The results show that our proposed algorithm is more efficient than previous algorithms.

ORCID
Ayman El‐Sayed [Link]

R EF E RE N C E
1. Goswami A, Kumar M. A review on energy harvesting in wireless sensor networks. Int J Innov Res Sci Eng Technol (IJIRSET).
April 2017;6(4):6863‐6867.
2. Babayo AA, Anisi MH, Ali I. A review on energy management schemes in energy harvesting wireless sensor networks. Renew Sustain
Energy Rev. September 2017;76:1176‐1184.
3. Huang C‐F, Tseng Y‐C. The coverage problem in a wireless sensor network. Mobile Network Appl. 2005;10(4):519‐528.
12 AL‐BAZ AND EL‐SAYED

4. Al‐Karaki JN, Kamal AE. Routing techniques in wireless sensor networks: a survey. IEEE Wireless Comm. 2004;11(6):6‐28.
5. Singh SK, Singh MP, Singh DK. A survey of energy‐efficient hierarchical cluster‐based routing in wireless sensor networks. Int J Adv Netw
Appl (IJANA). 2010;2(02):570‐580.
6. Zia Q. A survey of data‐centric protocols for wireless sensor networks. Comput Sci Syst Biol. 2015;8(3):127‐131.
7. Heinzelman WR, Chandrakasan A, Balakrishnan H. Energy‐efficient routing protocols for wireless microsensors networks.
In: Proceedings of Hawaii International Conference on System and Sciences; 2000.
8. Braman A, Umapathi GR. A comparative study on advances in LEACH routing protocol for wireless sensor networks: a survey. Int J Adv
Res Comput Commun Eng. 2014;3(2):5683‐5690.
9. Heinzelman WB, Chandrakasan AP, Balakrishnan H. An application‐specific protocol architecture for wireless microsensor networks.
IEEE Trans Wirel Commun. 2002;1(4):660‐670.
10. Aslam M, Javaid N, Rahim A, Nazir U, Bibi A, Khan ZA. Survey of extended LEACH‐based clustering routing protocols for wireless sensor
networks. In: High Performance Computing and Communication & 2012 IEEE 9th International Conference on Embedded Software and
Systems (HPCC‐ICESS), 2012 IEEE 14th International Conference on. IEEE; 2012: 1232‐1238.
11. Wu XH, Wang S. Performance comparison of LEACH and LEACH‐C protocols by NS2. In: Proc. 9th International Symposium on
Distributed Computing and Applications to Business, Engineering and Science. Hong Kong, China; 2010: 254‐258.
12. Gnanambigai J, Rengarajan N, Anbukkarasi K. Leach and its descendant protocols: a survey. Int J Commun Comput Tech. September
2012;1(3):15‐21.
13. Lindsey S, Raghavendra CS. PEGASIS: power‐efficient gathering in sensor information systems. In: Aerospace conference proceedings,
2002. IEEE, vol. 3, IEEE, 2002: 3‐3.
14. Buyanjargal O, Kwon Y. “An energy efficient clustering algorithm for event‐driven wireless sensor networks (EECED).” In: INC, IMS and
IDC, 2009. NCM'09. Fifth International Joint Conference on, IEEE, 2009: 1758‐1763.
15. Kim D‐S, Chung Y‐J. Self‐organization routing protocol supporting mobile nodes for wireless sensor network. In: Computer and
Computational Sciences, 2006. IMSCCS'06. First International Multi‐Symposiums on, vol. 2. IEEE; 2006: 622‐626.
16. Kumar GS, Vinu PMV, Jacob KP. Mobility metric based leach‐mobile protocol. In: Advanced Computing and Communications, 2008.
ADCOM 2008. 16th International Conference on, IEEE; 2008: 248‐253.
17. Liu J‐L, Ravishankar CV. LEACH‐GA: genetic algorithm‐based energy‐efficient adaptive clustering protocol for wireless sensor networks.
Int J Mach Learn Comput. 2011;1(1):79
18. Loscri V, Morabito G, Marano S. A two‐levels hierarchy for low‐energy adaptive clustering hierarchy (TL‐LEACH). In: IEEE vehicular
technology conference, vol. 62, no. 3, IEEE; 1999, 2005: 1809.
19. Sindhwani N, Vaid R. V‐Leach: an energy efficient communication protocol for WSN. J Mechanica Confab. Feb‐March 2013;2:79‐84.
20. Ahlawat A, Malik V. An extended vice‐cluster selection approach to improve V Leach protocol in WSN. In: 3rd International Conference
on Advanced Computing and Communication Technologies (ACCT), 6‐7 April 2013: 236‐240.
21. Liu Y, Xu K, Luo Z, Chen L. A reliable clustering algorithm base on LEACH protocol in wireless mobile sensor networks. In: Mechanical
and Electrical Technology (ICMET), 2010 2nd International Conference on, IEEE; 2010: 692‐696.
22. Chen J, Shen H. MELEACH‐L: more energy‐efficient LEACH for large‐scale WSNs. In: Wireless Communications, Networking and
Mobile Computing, 2008. WiCOM'08. 4th International Conference on, IEEE; 2008: 1‐4.
23. Xiangning F, Yulin S. Improvement on LEACH protocol of wireless sensor network. In: Sensor Technologies and Applications, 2007.
Sensor Comm 2007. International Conference on, IEEE, 2007: 260‐264.
24. Fu C, Jiang Z, Wei W, Wei A. An energy balanced algorithm of LEACH protocol in WSN. Int J Comput Sci. 2013;10(1):354‐359.
25. Ramesh K, Somasundaram K. A comparative study of clusterhead selection algorithms in wireless sensor networks. arXiv preprint
arXiv:1205.1673; 2012.
26. Mahalik NP. Sensor Networks and Configuration. Fundamentals, Standards, Platforms, and Applications. Berlin: Springer‐Verlag; 2007.
27. Nawar NM, Soliman SE, Kelash HM, El‐sayed HS. Enhancement of mobility model for cluster hierarchical routing protocol for wireless
sensor networks. Int J Comput Appl (IJCA). May 2014;94(9):12‐16.
28. Al‐baz A, El‐Sayed A. Energy‐aware enhancement of LEACH protocol in wireless sensor network. In: 25th International Conference on
Computer Theory and Applications (ICCTA 2015), Alexandria, Egypt, 24‐26 October 2015.
29. Al‐Baz A, El‐Sayed A. Cluster head selection enhancement of LEACH protocol in wireless sensor network. Minufiya J Electron Eng Res
(MJEER). January 2017;26(1):153‐169.
30. Deosarkar BP, Yadav NS, Yadav RP. Clusterhead selection in clustering algorithms for wireless sensor networks: a survey. In: Computing,
Communication and Networking, 2008. ICCCn 2008. International Conference on, IEEE; 2008: 1‐8.
31. Soro S, Heinzelman WB. Cluster head election techniques for coverage preservation in wireless sensor networks. Ad Hoc Netw.
2009;7(5):955‐972.
AL‐BAZ AND EL‐SAYED 13

32. Al‐Zou'bi A, Khamayseh Y, Mardini W, Yassein MB. Improvement on LEACH protocol of wireless sensor network (VLEACH). Int J Digit
Content Technol Appl (IJDCTA). Feb. 2009;3(2):132‐136.
33. El‐Fishawy N, Hamouda A, Attiya GM, Atef M. Arabic summarization in Twitter social network. J Ain Shams Eng. 2014;5:411‐420.
34. Zhang J, Li W, Cui D, Zhao X, Yin Z. The NS2‐based simulation and research on wireless sensor network route protocol. In: Wireless
Communications, Networking and Mobile Computing, 2009. WiCom'09. 5th International Conference on, IEEE; 2009: 1‐4.

How to cite this article: Al‐Baz A, El‐Sayed A. A new algorithm for cluster head selection in LEACH protocol
for wireless sensor networks. Int J Commun Syst. 2017;e3407. [Link]