An energy efficient and adaptive clustering for wireless sensor network (CH-leach) using leach protocol

Abstract

Researchers in wireless sensor networks (WSN), have given deep attention to large scale integration and energy-efficiency (energy consumption). Energy-efficient solutions can conserve valuable sensor-node energy. This is one of the main critical challenges that WSNs face, which plays a fundamental part in determining the lifetime of the network. Although, there are many WSN protocols, clustering based hierarchal routing protocols are given more consideration because of their improve scalability. In particular, sensors are battery-powered, often limiting available energy, which is not changeable in most of the situations. One of the most common energy-efficiency sensor networks protocols is Low Energy Adaptive Clustering Hierarchy (LEACH) as source. In this paper, we propose CH-leach. We present architectures, schemes and evaluate. Its performance using analytical study and simulations. The evaluation was based on the most critical metrics in WSNs, such as: energy-efficiency (energy consumption), and network lifetime. The evaluation and comparison with existing solutions show that our proposed CH-leach exhibits a reduction in energy consumption over LEACH and DEEC. While the overall network lifetime of CH-leach is improved 91% and 43% more than LEACH and DEEC protocols respectively.

Full text

An Energy Efficient and Adaptive Clustering for
Wireless Sensor Network (CH-leach) using Leach
Protocol
Walid Abushiba
Department of
Electronics and Electrical
Engineering
Liverpool John Moores
University
Liverpool L3 3AF, UK
w.m.abushiba@2012.ljm
u.ac.uk
Princy Johnson
Department of
Electronics and Electrical
Engineering
Liverpool John Moores
University
Liverpool L3 3AF, UK
P.Johnson@ljmu.ac.uk
Saad Alharthi
Department of
Electronics and Electrical
Engineering
Liverpool John Moores
University
Liverpool L3 3AF, UK
S.A.Alharthi@2014.ljmu
.ac.uk
Colin Wright
Department of
Electronics and Electrical
Engineering
Liverpool John Moores
University
Liverpool L3 3AF, UK
C.Wright@ljmu.ac.uk
Abstract— Researchers in wireless sensor networks (WSN),
have given deep attention to large scale integration and energy-
efficiency (energy consumption). Energy-efficient solutions can
conserve valuable sensor-node energy. This is one of the main
critical challenges that WSNs face, which plays a fundamental
part in determining the lifetime of the network. Although, there
are many WSN protocols, clustering based hierarchal routing
protocols are given more consideration because of their improve
scalability. In particular, sensors are battery-powered, often
limiting available energy, which is not changeable in most of the
situations. One of the most common energy-efficiency sensor
networks protocols is Low Energy Adaptive Clustering
Hierarchy (LEACH) as source. In this paper, we propose CH-
leach. We present architectures, schemes and evaluate. Its
performance using analytical study and simulations. The
evaluation was based on the most critical metrics in WSNs, such
as: energy-efficiency (energy consumption), and network lifetime.
The evaluation and comparison with existing solutions show that
our proposed CH-leach exhibits a reduction in energy
consumption over LEACH and DEEC. While the overall network
lifetime of CH-leach is improved 91% and 43% more than
LEACH and DEEC protocols respectively.
Keywords— Wireless sensor network; Cluster-head selection;
energy-efficiency; base station (BS); Network lifetime; sensor-
node; LEACH and DEEC.
I. INTRODUCTION
A typical wireless sensor network (WSN) can be considered as
an uncommon breed of wireless ad hoc networks with
decreased or no mobility. These networks combine wireless
communication and negligible on board computation facilities
with detecting and monitoring of physical and environmental
phenomena. Sensing is a technique used to gather information
about a physical object, process, environmental phenomenon
or the occurrence of events (e.g. changes in the state such as
rise or drop in temperature). These little size, low-cost sensor
gadgets have inserted on board radio transceiver, micro-
controller, memory, power supply and the real sensors. All
these components together in a single gadget shapes a so-
called remote Sensor-Node or basically a Sensor [1, 5, 9-
11].There are many fundamental factors in wireless network
which have impact of design a good protocols [7]. The
lifetime of the network is important, the network should
function for as long as possible but, it is too critical to
understand the parameters that suit sensor applications.
Moreover, the sensors node around the base station (BS) will
act as communicator for the sensors which are far from BS
[8], so by using the clustering algorithms associate with K-
means method can help the sensor nodes to extended and
maintain their live time which extends the live time of the
network overall.
There are different ways to design cluster-based WSNs. Since
all neighboring sensor nodes normally have the same data of
the same event and each node transmit to BS individually, this
cause energy consumption and the nodes will last very short
time. Cluster-Head architectures reduce the energy
consumption [7]. All the Cluster-Head transmit the data
directly to (BS) but the other nodes will only transmit the
collected data to CH. Hence the cluster-head selection will
determining the lifetime of the network.
In this work novel strategies for both topology and a routing
algorithm are proposed to maximization of the network life
time. CH-Leach Protocol, an approach of algorithm proposed,
this proposed research used number of connection in cluster,
and for each cluster head (CH) communicate with base station,
however the selection of the cluster head based on the number
of cluster on the network gird area, this method allow the
network to adopted the best scenario to extend life time of the
network, different ways of cluster are formed, in order to
avoid the condition that one cluster will contain large of
connection nodes and the rest is not, the maximum number of
the cluster head is chosen in different scenario to test the
network coverage.
The layout of the paper is as follows. Section II gives a
reviews of the Related Work, Section IV and V Simulation
Results and Conclusions of CH-leach are presented
respectively.

II. RELATED WORK
Leach was developed in order to provide a unique
necessities of wireless sensor networks. Most of the
application protocols architecture in a sensors network [6]
have main function which to forward the data gathered by
sensors to BS. In order to avoid energy consumption many
approaches have been proposed to achieve low energy
consumption. Hence base station is not near all sensor nodes,
therefore the node will use excessive energy to deliver data.
Cluster-Head algorithms aimed to solve this problem.
A. Leach (low-Energy Adaptive Clustering Hierachy)
The main task for clustering-based routing protocols over
the static clustering algorithms to minimize global energy
consume by nodes. The nodes are required to distribute the
load not in same time but over period of time [7]. Since the
sensor nodes will connect the appropriate cluster heads
depending on the signal strength, this methodology require the
nodes which have the highest energy within the cluster to
volunteer to be the cluster head and transmit the aggregate data
to Base Station.
B. Improved Leach Protocls
ED-LEACH [14] studied Euclidean distance between nodes
to improve location of cluster heads in a region, due to random
deployment of the nodes which become placed closed or
sometimes far away from each other. In [6] have proposed a
new cluster-head selection method for Leach. Same works
proposed in [14] takes in consideration of the remaining energy
of nodes and the protocol has two level of operation alike to
that on Leach, introducing a random delay before sending
ADV messages by cluster-head nodes made it better for cluster
to join the process which results in 17% of the reduction in
cluster-head but no unreachability nodes are mention.
LEACH-CE (Leach-Centralized Efficient) [13], in this
protocol although the improvement is made to Leach, but still
the nodes with highest energy in region will become a cluster-
head, due to nodes that chosen with less energy in some round
will die sooner. Leach-CH in setup phase choose the higher
nodes as cluster-head in each round, this will eliminates the
average life time of the network. ME-LEACH [3], also based
on LEACH, and more an energy efficient compare to original
LEACH, by reducing the communication distances between
sensor nodes, but this achievement comes by powerful radio
which will not work efficiently on large scale networks.
Leach has been improved in various areas, both in setup
and steady phase, for instance cluster-head selection, cluster
formation algorithms and energy reduction. Therefor this work
aims to apply a concept conserving energy overall and using it
to enhance Leach protocol in terms of novel cluster-head
selection.
C. Deec (Distributed Energy Efficient Clustering)
DEEC [12], this protocol is designed for heterogeneous
wireless sensor network, cluster-head selection is determined
by probability according to the remaining energy of each nodes
and average energy of the network. Nodes which have more
remaining energy have more probability to become cluster-
head than nodes have less remaining energy.
III. PROPSED APPROACH
This section present the concept of proposed network
model and characteristics, sensors network with a number of
nodes N, assumed that all nodes are randomly spread over the
100*100 m field with all the nodes are homogeneous which
means for every node having the same identical sensing,
communication capabilities and same initial energy. Regarding
the sink (Base station) is fixed, we consider two scenario for
(BS) location at edge of the network bounding and in the center
of the field, fig.1 describes the approach.
Fig. 1. The Network Topology.
A. Energy Efficient Cluster-Head Selection k-Means Approch
Leach selects a certain number of nodes as cluster head and
in order to maintain the energy dissipation, ration is taking
place, to centralize the data collection and transmission to Base
Station over number of rounds. Each round is divided by setup
phase and the steady state phase. Moreover based on the
probability calculated in advance whether the node will
become cluster-head or not, for the current round, considering
the number of times the node has been selected as cluster-head
before. The node will make the decision by choosing a random
number from 0 to 1 periodically. By setting the threshold T(n),
the node will becomes a CH for the current round if the number
is less than that threshold.
Whereas p, r, G percentage of the cluster- head, the current
round and the set of nodes which not been selected as cluster-
head in last 1/p round, respectively represented.
In Leach the choice of the cluster-head is randomly made.
This leads to unstable energy level among the nodes in
network. If the selected cluster head is far away from Base
Station, will spend more energy to transmit the data to Base
Station to those nearby the Base Station.

In the proposed protocol will consider k-Means approach to
ensure balanced energy distribution over the whole area of the
network, additional parameters will add to optimize the process
of setup phase (cluster-head selection) [4].
The aim is to maximum the life time of the network by
dividing the field of the network into cluster area size
equivalent, every cluster area has cluster head, the threshold
T(n) given in (1), represent the remaining energy of the node’s,
according to this threshold every node will decide whether to
become cluster head or not in each round.
Centralized k-Means in each area, although the number of
nodes not equally in each area, after diving the field of the
network to number of cluster area equal in size, using k-Means
algorithm (2), allow the process to be unsupervised, the
following steps describe the processes:
 Generate k pointes cluster center, the number of the
cluster required k.
 Set the data points. In this case the data points will
assign to location of the nodes, in every cluster area
required
 The cluster center (centroid) in each area will assign to
nearest node by calculating the mean value of the nodes
location (data points) in each cluster.
 This node will become the cluster-head if it has enough
energy, means above the threshold set.
 Repeat the steps when assigning node that is near to the
cluster center not able to act as cluster head.
CH-Leach selects a certain number of nodes as these nodes
are nearest to centroid in their cluster region, cluster head will
be responsible for transmit the data to Base Station. The
cluster-heads are randomly selected and in every round each of
the nodes that will become cluster-head is assigned to nearest
centroid. As the nodes that already chosen before have energy
under the threshold, then the cluster-head will updated by
chosen another node, these steps are depending on location of
the nodes which are randomly deployed. The centralized k-
Means [4], finding the mean of the locations of the nodes and
then decided on the centroid of the layout cluster area.
To ensure Centralized k-Means algorithm [4], works
unsupervised, the following inputs should consider:
 The set of data points which the nodes location.
 The number of the cluster area, in every cluster area
will have one centroid and the cluster-head selection
based on this determination.
 In proposed protocol, the nodes will be static so the
repetition of a process is limited in Centralized k-Means
algorithm.
The CH-Leach protocol is based on Leach to ensure energy
balancing and extended the network life time.
B. Setup Phase
The process of algorithms begins with the setup phase
where the cluster area is defined by certain numbers. The nodes
which are nearest to centroid will act as cluster-head, sensing
the data from the follower nodes. By doing so, the nodes will
not communicate with each other, but will communicate to
cluster-head nodes assign which will communicate with Base
Station, this cluster will act as cluster-head as long as it has
enough energy above than threshold. Rotation of the cluster-
head is based on energy load. Moreover when the cluster-head
become a normal node, it can communicate to active cluster-
head. Implanting Centralized k-Means algorithm added high
stability in setup phase which gives longer network life time.
C. Proposed Network Model
To evaluate performance of CH-Leach protocol, the
following performance metrics were used:
 The Network Life Time: the time period of sensor
nodes are still active which means transmitting and
receiving data, assuming after 90% of nodes have not
enough energy to do so, the network will consider
expired.
 Energy consumption: during the operational sensors
network, the amount of energy the nodes will dissipate
in transmitting and receiving data between normal
nodes and cluster-head nodes.
As showed in Fig. 1, the boundary of the network is defined
as 100*100 meter squire where a Base Station is placed firstly
in middle of the area and by the edge of the Network area in
the other layout of the network, the number of 100 sensors
nodes randomly deployed and all are active at the beginning of
the simulation with initial energy equal to 0.5 j as well as
homogeneous, All the nodes will transmit the sensing data to
Base Station via cluster-head in their region. In terms of the
number of the cluster area will divide the area in both scenario
and also considering the area of the Network 100*100, the
maximum cluster area will be not more 30, The Base Station is
stationary and resource high-energy for fair comparison to
previously published protocols.
TABLE I. PARAMETERS VALUE USED IN SIMULATION
Parameter
Value
Network size (M*M)
100m by 100m
Location of BS (Base Station)
50*50
By Edge of the network
Number of Nodes
100
Cluster head probability
0.1
Initial Energy
0.5 J
ETX and Erx (Eelec)
50 n J/bit
Number of the Cluster (K-
means)
5,10,15,20,25 and 30
The Data Paket Size (bits)
4000

IV. SIMULATION RESULTS
In this section, we provided two scenarios to illustrate the
proposed protocol algorithm capabilities using Matlab R2015a
simulation and compare its performance with Leach and Deec.
The network topology as describe on Fig.1 and specific
parameters as given in Table 1.
Firstly, the Base station is placed in the center of the
network 100 * 100. Here observation of the behavior of the
algorithm by chosen number of cluster area 5, 10, 15, 20, 25
and 30 were implanted, and in other scenario where the Base
station were place just by edge of network boundary, In Fig 2
and Fig 3 the CH-Leach improvement of Network Life Time
the over Leach and Deec are shown.
The improvement in CH-Leach is very clear as the number
of cluster areas was increased. Although the curve was a
straight line in most chosen experiments, comparing to Deec,
the curve is decreasing after the point were 30 cluster
comparing to Leach when the BS is placed on the edge.
Fig. 2. Number of cluster area, Improvement of CH- Leach (BS 50/50).
Fig. 3. Number of cluster area, Improvement of CH- Leach (BS 0/99).
TABLE II. NETWORK LIFE TIME COMPARISOPN
Number
of
Cluster
Area
Network Life Time
Leach
Deec
CH-Leach
BS
50/50
BS
00/99
BS
50/50
BS
00/99
BS
50/50
BS
00/99
5
911
972
1228
1177
1330
1310
10
888
900
1227
1198
1412
1397
15
872
859
1224
1204
1543
1484
20
904
878
1229
1198
1553
1539
25
940
826
1249
1261
1695
1641
30
919
932
1228
1216
1756
1641
Fig. 4. Network Life Time.
Fig. 5. Energy Dissipstion by Nodes over Number of Rounds.

In order to evaluate the reliability of CH-Leach protocols,
Fig. 6 and Fig. 7, show the Life Time comparing to Leach and
Deec protocols. It is clear from the figure that CH-Leach
protocol has performed well. When the Base Station is placed
on edge the improvement is quite significant due to efficient
cluster-head selection mechanism.
Fig. 6. The Life Time of the Network on Differint Cluster area, Base Station
Located on (50/50).
Fig. 7. The Life Time of the Network on Differint Cluster area, Base Station
Located on (edge).
V. CONCLUSIONS
In this paper, a series of experiments on different scenarios
were implemented and tested. The life time of the network in
CH-Leach shows major extension compared to Leach and
Deec protocols. The main aim of this work were to design and
implement a protocol which enhance exiting protocols in order
extend the Life Time of Network.
REFERENCES
[1] Mohammed Abo Zahhad, Sabah M. Ahmed, N. Sabor, and
Shigenobu Sa saki, “Mobile Sink-Based Adaptive Immune Energy
Efficient Clustering Protocol for Improving the Lifetime and
Stability Period” IEEE Sensors Journal,Vol. 15, No. 8, pp 4576-
4586, Aug 2015.
[2] Chaima Bejaoui, Alexandre Guitton and Abdennaceur Kachouri,
"Improved election of cluster heads in LEACH", IEEE/ACS 12th
(AICCSA), Marrakech, pp. 1-4, 2015.
[3] J. Chen H. Shen "MELEACH an energy-efficient routing protocol for
WSNs", vol. 9 no. 4 pp. 035 Aug. 2007.
[4] Tapas Kanungo, David M. Mount, Nathan S. Netanyahu, Christine D.
Piatko, Ruth Silverman, and Angela Y. Wu, "An efficient k-means
clustering algorithm: Analysis and implementation". IEEE
TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE
INTELLIGENCE, Vol. 24, No. 7, pp. 881-892, JULY 2002.
[5] Zahoor A. Khan and Srinivas Sampalli Khan, “AZR-LEACH: An
Energy Efficient Routing Protocol for Wireless Sensor Networks”, Int.
Journal Communications, Network and System Sciences, Vol 5, p 785-
795, Nov 2012.
[6] Chung-Shuo FAN, “Rich: Region-based intelligent cluster-head
selection and node deployment strategy in concentric-based WSNs”,
Advances in Electrical and Computer Engineering
Vol 13, No 4, 2013.
[7] M. Handy, M. Haase and D. Timmermann. “Low energy adaptive
clustering hierarchy with deterministic cluster-head selection”, 4th
International Workshop on Mobile and Wireless Communications
Network, pp 368–372, Sweden, Sep 2002.
[8] Wendi B. Heinzelman, Anantha P. Chandrakasan, and Hari
Balakrishnan, “Energy-Efficient Communication Protocol for
Wireless Microsensor Networks”, Hawaii International Conference
on System Sciences, 2000.
[9] Wendi B. Heinzelman, Anantha P. Chandrakasan, and Hari
Balakrishnan, “An a pplication-specific protocol architecture for wireless
microsensor networks,” IEEE TRANSACTIONS ON WIRELESS
COMM.,Vol.1,no. 4, Oct 2002.
[10] Hou, Rui, Weizheng Ren, and Yaodong Zhang. " A wireless sensor
network clustering algorithm based on energy and distance." Second
International Workshop on Computer Science and Engineering, 2009.
[11] A. Ihsan, K. Saghar T. Fatima. Analysis of leach protocol(s) using
formal verification. IEEE, International Bhurban Conference on Applied
Sciences & Technology (IBCAST ‘15), pp. 254-262, Jan 2015.
[12] Jiahu Qin, Weimiing Fu, Huijun Gao and Wei Xing Zheng, “Distributed
k-means algorithm and fuzzy c-means algorithm for sensor networks
based on multiagent consensus theory", IEEE Transactions on
Cybernetics, Vol. 99, pp. 1–12, 2016.
[13] Li Qing, Qingxin Zhu and Mingwen Wang, “Design of a distributed
energy-efficient clustering algorithm for heterogeneous wireless sensor
networks”, Computer Communications, Vol 29, Issue 12, Pp 2230-2237,
4 Aug 2006.
[14] Meenakshi Tripathi, R. B. Battula, M. S. Gaur and V. Laxmi, “Energy
Efficient Clustered Routing for Wireless Sensor Network”, 9th
International Conference on Mobile Ad-hoc and Sensor Networks, Dec
2013.