880114 1569846515
LEACH Enhancements for Wireless Sensor Networks Based on Energy Model ' 3 4 Mohammad Shurman , Noor Awad 2, Mamoun F. AI-Mistarihi , and Khalid A. Darabkh (I) Jordan University of Science and TechnologylNetwork Engineering and Security Department, Irbid, Jordan (2) Jordan University of Science and Technology/Computer Engineering Department, Irbid, Jordan (3) Jordan University of Science and Technology/ Electrical Engineering Department, Irbid, Jordan (4) The University of Jordan/Computer Engineering Department, Amman, Jordan
[email protected],
[email protected],
[email protected],
[email protected]
Abstract - The low-energy adaptive clustering hierarchy
task. Due to the large number of these collaborative nodes, it is
(LEACH) protocol is one of the most adaptive protocols used in
not feasible to assign a global identifier to each node. This
the wireless sensor networks (WSNs). LEACH is a hierarchical model that provides a powerful strategy for energy balancing, using head rotations. However, the original LEACH protocol suffers from many drawbacks and many researchers proposed new methods to mitigate them. In this paper, we propose two approaches based on an energy model to enhance cluster heads (CHs) selection method through not only minimizing the power
consideration has led to data centric routing, whereby the base station (BS) sends queries to certain regions and awaits data from sensors located in the selected regions. The low-energy adaptive clustering hierarchy (LEACH) protocol is one of the hierarchical clustering-based routing protocols in which time is divided into rounds for the sake of allowing all available
consumption of network nodes, but also minimizing the number
nodes to compete to be a cluster head (CH). In location-based
of CHs.
protocols, sensor nodes are addressed by means of their
head
The first proposed approach effectively selects a cluster that
has
the
lowest
power
consumption
when
communicating with other nodes. In addition to what is proposed in the first approach, the idea of incorporating the shortest
locations. The distance between neighboring nodes can be estimated on the basis of incoming signal strengths. In the last few years, many researchers have proposed new
distance has been considered in the second proposed approach. This
significantly
decreases
the
energy
consumption
and
increases the lifetime of associated nodes.
techniques and enhancements to improve the performance of the LEACH protocol, as it is considered the dominant protocol for WSNs. Extensive research has been done to introduce new
Keywords: LEACH protocol, Wireless sensor network, Power
methods to determine the cluster heads and their optimal numbers, as reported in [1-2]. Many researchers focused on
consumption, Energy model.
efficient energy and power consumption reduction for wireless I.
sensor network nodes [3]. One interesting comparison, based
INTRODUCTION
on
A considerable amount of research has been done on wireless sensor networks (WSNs). A WSN is a system which mainly consists of many distributed sensor nodes that collect and process their data in an efficient manner. There are several types of sensor nodes in the market out of which Berkeley's motes, Shockfish's tiny nodes, Moteiv's telos, UCLA's ibadge and
RFID.
The
WSNs
have
proven
to
be
a
powerful
differing
variants,
was
used
to
define
the
LEACH
shortcomings [4]. Authors in [5] introduced an improved version of LEACH for data gathering and aggregation. Other approaches concentrate on increasing the nodes' lifetime and power consumption minimization in LEACH protocol [6]. Other techniques have been merged with LEACH
[7-8]
protocol such as fuzzy logic that was used to elect a cluster head among available nodes [8].
mechanism in diverse applications of various fields such as
In this paper, we introduce new approaches that enhance
(but not limited to) military, health, commerce, environment
the lifetime of the LEACH protocol based on an efficient
and ecology, and forest fire detection. Many routing protocols
energy model. The new approaches change the cluster head
have been introduced for WSN such as flooding, gossiping,
election strategy. This paper is further presented as follows; Section II provides the related work. Section III provides a
flat-routing, hierarchical routing, and location based routing. Flooding is a traditional technique in which the data is transmitted
without
incorporating
any
algorithms
or
maintenance. Unfortunately, it suffers from data implosion, overlapping, and blindness of resources. Gossiping is an enhanced version of flooding that solves the issue of data implosion,
whereas
each
node forwards the
data
to
its
brief description of the LEACH protocol. Section IV provides the proposed schemes. Performance evaluation is presented in Section V. We concluded this paper in Section VI. II.
RELATED WORK
Some researchers introduced new methods based on the
neighbors, and introduces a higher delay in transmission. The
idea of stochastic cluster head selection algorithm [2, 5, 9-10].
nodes in multi-hop flat routing protocols play the same role
In [2], the authors proposed the idea of adding the remaining
and sensor nodes collaborate together to perfonn the sensing
energy level of a node to the
978-1-4799-3866-7/14/$31.00 ©2014
IEEE
T(n)
equation. Simulation results
showed that this modification increased the lifetime of a
less than the threshold
network by 30%. However, one of the drawbacks in this
then chosen as a cluster head, according to:
model
is
the
inefficiency
of
network
encountered
after
surpassing a certain number of rounds. In
[11],
a
grid-based
cluster
T(n) partition
method
was
presented. This new method divides the monitoring region into
where
grids with equal slots where nodes in each grid form a cluster
=
(
T(n) ,
whereby an associated node is
1- p[r * :Od(1/P)]' 0,
ifn
E G
(1)
otherwise
p is the desired cluster head percentage, r is the current
round, and G is the set of nodes which have not been cluster
and propose a numerical computing method to determine the
heads in
number of cluster heads. Another protocol, based on the energy clustering algorithm, was proposed in [1] which mainly
lip rounds.
The second phase, or the steady phase, is the one in which
determines the number of cluster heads according to the afore
the sensor nodes start sensing and transmitting data to the CH.
mentioned grid-based clustering methods.
The CH then aggregates the data from available nodes prior to sending their data to the base station.
One more interesting protocol which is based on the consideration of node residual energy during CH election was
B. Shortcomings of Original LEACH
proposed in [lO] whereas the concept of energy classification was introduced. During the setup phase, each node produces a
The original LEACH protocol has some drawbacks such as
timer in accordance to its energy, where the timer value has an
[8-9]:
inverse relationship with this specific node's energy. In this
•
case, the node chooses a random number between 0 and 1 and once the timer expires, energy levels are assessed to elect CH.
are within the range of the BS wireless transmission
The higher the energy a node has, the smaller number it randomly picks. If the number is less than the threshold
T,
In many WSN protocols, LEACH assumes that nodes
•
the
node reverts to a CH for the current round; otherwise, the node
even if this is not the case. The expansion of the network is not suitable in real networks since the BS is usually located at the
picks another random number, then CH sends ADV messages and other cluster members send join request messages as in
•
LEACH protocol. If the node receives cluster information before the timer expires, it will cancel the timer and send a
•
join request to the CH. Simulation results showed a better performance over LEACH protocol in terms of network
•
lifetime and loads equalization degree.
longest distance from sensor node region. The election of cluster heads is based on a probability model which does not consider node energy. The lifetime of the network is not as long as the randomly chosen cluster heads. There is no definitive location distribution of cluster heads; hence, some parts of the network may not be
Distributed optical fiber sensors have been intensively
covered by nodes.
studied and deployed for analyzing loss of external pressure,
•
or temperature distribution along a fiber link. Hybrid sensor
Time division multiple access (TDMA) schedule has some drawbacks: At each TDMA frame, the cluster
networks consisting of both optical and wireless nodes have
head will send data in the designated slot even if
also been studied. Authors in [12] investigated the hybrid
there is no new value.
sensor networks that were composed of a distributed fiber sensor (DFS) links and WSNs. They proposed a new optical LEACH (O-LEACH) protocol based on the LEACH protocol,
IV.
whereas a DFS link is located at the center of sensors field where it generally requires a constant and relatively high
The proposed schemes aim to enhance the election method
power supply. Hence, it is usually used as the "sink" or base
of cluster heads by minimizing power consumption of network
station for the network. In addition to standard parameters
nodes and minimizing the number of CHs. The schemes take
used in LEACEH, the coverage area of DFS can be varied to obtain
optimum
value.
Simulation
results
of
into consideration the drawbacks of the original LEACH, and
O-LEACH
accordingly try to minimize the energy consumption. The
showed about a 30% longer lifetime than that of LEACH, and notably,
the
performance
stays
relatively
constant
PROPOSED SCHEMES
main idea of our proposed schemes is split into two stages.
when
The first stage is achieved through the application of a basic
varying the DFS coverage percentage.
equation which determines the number of CHs while the second stage part I consider the energy factor in the cluster
III.
head election process. In the part II of second stage, the
PRELIMINARIES OF LEACH PROTOCOL
distance parameter has been also appended to the threshold equation. These stages are discussed as follows:
A. Original LEACH LEACH is one of the earliest and most popular protocols
A. Stage One
used in WSNs [13]. It is based on spreading the dissipating energy required for communication from the base station to all
Cluster head numbers are determined using a theoretical
available sensor nodes. It randomly selects a sensor node as a
model which is based on the node coverage monitoring
cluster head, using two phases. The first phase, or the set-up
regions and their communication radius of nodes versus the
phase, is the one in which each sensor node selects a random
previous described randomized method. The new equation
number between 0 and 1 and tests if this random number is
which determines CH numbers is:
2
B. Simulation Results
(2)
In order to achieve a fair comparison, a total of 20 runs are where
M
is the monitoring region area,
communication, and
r
used
is the radius of node
k is the number of cluster heads.
*
with
Table 2 shows a comparison of the three approaches based and worst total energy consumed. Results found in this table show that the original LEACH protocol is slightly better than approach
1
in
terms
of
all
parameters
except
standard
deviation. However, the best energy values for the original
model which incorporates the energy method, whereby energy
LEACH protocol and approach 1 are almost close to each
consumption of every node in the network is considered in
other. Considering the results of approach 2, it is certainly
order to determine the threshold factor that will be used for the
clear that this approach achieves superior results in terms of all
new enhanced CHs election.
parameters. These results confirm that the modified approach which is based on the energy model and shortest distance is
Second Stage (Part II):Modified Energy Based Election Strategy (Approach 2)
the best proposed model. TABLE 2 Comparison of the three LEACH models based on statistical results
In this subsection, another modification to the LEACH protocol is introduced. This modified approach appends the
Approach I
distance as a new parameter, based on the energy model discussed in subsection
B.
This approach keeps the same
strategy as the energy model, in terms of calculating the number of CHs and updates. The threshold equation becomes as follows:
T(n) where,
d is
£. * k =-! £tot+d
(4)
-
the distance between node
i
= .j[nodeX(O
- nodeY(j)F
33.50018
1.780965
31.4739
33.6612
1.69795
1.007296
0.774913
0.665018
29.8989
31.2879
0.4523
32.7623
34.5766
3.4527
and other nodes and is
calculated according to:
d
protocol
on the mean, median, standard deviation, as well as the best
It is worth mentioning that this equation presents a new
C.
LEACH
total energy consumed.
(3)
tot
original
number of packets delivered, number of alive nodes, as well as
heads in which it considers the energy factor given as: £. k i-
the
determine the best approach. The best and worst energy values
The second stage changes the basic election of cluster
=
comparing
are also provided. The figures presented below refer to the
B. Second Stage (Part J): Energy Based Election Strategy (Approach 1)
T(n)
for
proposed approaches 1 and 2. The average of energy is used to
Figs. 1, 2, and 3 show the total energy consumed, number +
of alive nodes, and number of packets delivered for the three
[nodeY(O - nodeX(j)F
approaches, respectively. It is undoubtedly shown in Fig.l that
(5)
approach 2 outperforms the LEACH protocol and proposed approach 1. On the other hand, the total energy consumed for
V.
the LEACH and approach 1 is on average close to each other.
PERFORMANCE EVALUATION
As far as the number of alive nodes is concerned and as depicted in Fig. 2, the positive results of LEACH protocol
A. Parameter Settings
over approach 1 indicate having inappropriate choice of cluster heads in approach 1 inside wireless sensor network
We used MATLAB to test the modified LEACH protocols.
(i.e., longer distances were selected). On the contrary, due to
The simulation environment is composed of 100 nodes with 20
an efficient use of node's energy for cluster head election, the
rounds. The simulation area covered up to 100xlOO. All others
better results become belong to approach 2. Packets being
parameter settings are presented in Table 1.
delivered for the LEACH
Parameter
Value
Packet Size
2000
Idle Energy
100*IOE-12
Transition Energy
500*IOE-12
Sensing Energy
100*IOE-12
Position of Base station
50,175
Percentage CH
0.05
protocol along with proposed
approaches are presented in Fig. 3. Among all approaches, it
TABLE 1 Parameter Settings
is seen that in almost all available rounds, the LEACH provides the lowest number of delivered packet. The obtained results of approach 1 indicate having minor improvement over the LEACH protocol. It is quite interesting to notice that the approach 2 gives the highest number of delivered packets which certainly can be justified to the efficient power use which definitely prolongs the network lifetime. VI.
CONCLUSION
An energy-based LEACH approaches are presented in this paper. The purpose of the proposed approaches is to improve
3
evaluation of selective and adaptive heads clustering algorithms over wireless sensor networks," Journal oj Network and Computer Applications, vo1.35, no.6, pp.2068-2080, November 2012.
the cluster head election process by modifying the calculation that determines the number of cluster heads. However, this paper introduced two approaches based on the energy model of LEACH protocol. The first approach is based on the pure
Total Energy Consumed
energy of a current node in comparison to all other nodes. The
2.5 ,------
second approach, which represents a modification of the first
" QI
approach, is based on two important parameters, namely, the energy
and
distance,
whereby
the
distance
is computed
:; 2
simulation
experiments
were
conducted
to
evaluate
"0 QI
the
E1.5
�
significance of our proposed approaches. Though the results show
minor
enhancements
of approach l
over
1.o-----/L....lo----I---II r-------+_+_------
g
between the current node and all working nodes. Many
+-----+-f-�--,/-+-_I'�_I_IJ__+_----I-A_fI'l '----
c o u > 1 +--�----�-----\-_i�-----
LEACH
protocol, but without a competing claimant, they confirm that
1:0 QI C
approach 2 is the one that should be employed.
w
.. 0 .5
�
REFERENCES
*-��........,:;i'..., ""*'........-c...: . �r_.��=k____;/�"*""-=-: � L EACH
[I] L. Wang, J. Liu, and W. Wang, "An Improvement and Simulation of LEACH Protocol for Wireless Sensor Network", in proceedings oj the
+Approach 1
10
......Appro.ch 2
12
16
14
20
18
Rouned Number
First International ConJerence on Pervasive Computing, Signal Processing and Applications, pp. 444-447, Harbin, September 2010.
Fig. 1 Comparison of total energy for LEACH versus proposed models
[2] N. Kumar and J. Kaur, "Improved LEACH Protocol for Wireless Sensor Networks", In proceedings oj 201i 7th International ConJerence on
Wireless Communications, Networking and Mobile Computing (WiCOM 2011), pp.I-5, Wuhan, September 201I.
Number of Alive Nodes
[3] N. Mittal, D. P. Singh, A. Panghal, and R.S. Chauhan, "Improved LEACH Communication Protocol for WSN", in proceedings oj National ConJerence on Computational Instrumentation (NCCI 2010), pp. 153-156, Chandigarh, March 2010. [4] M. Haneef and Z. Deng, "Comparative analysis of classical routing protocol LEACH and its updated variants that improved network life time by addressing shortcomings in Wireless Sensor Network", In proceedings
l00 �-----
95
II> QI "0 0 Z QI .�
90
�
85
QI .D
80
... 0
oj 2011 Seventh international ConJerence on Mobile Ad-hoc and Sensor Networks (MSN 2011), pp. 361 - 363, Beijing, December 2011. [5] Wei Bo, Hu Han-ying, and Fu Wen, "An Improved LEACH Protocol for Data Gathering and Aggregation in Wireless Sensor Networks", In proceedings oj 2008 International ConJerence on Computer and Electrical Engineering (lCCEE 2008), pp. 398 - 401, Phuket, December
�
E
" Z
75
2008. [6] V. Katiyar, N. Chand, G. C. Gautam, and A. Kumar, "Improvement in LEACH Protocol for Large-scale Wireless Sensor Networks", In
70 +--,--,--,---,--,--.--,--,��---, 10
proceedings oj 2011 International ConJerence on Emerging Trends in Electrical and Computer Technology (ICETECT 20i 1), pp. 1070 - 1075,
12
18
16
14
20
Rouned Number
Tamil Nadu, March 2011. [7] Y. Li, N. Yu, W. Zhang, W. Zhao, X. You, and M. Daneshmand, "Enhancing the Perfonnance of LEACH Protocol in Wireless Sensor Networks", in proceedings oj IEEE iNFOCOM Workshop on M2MCN 20II, Shanghai, April 20II. [8] G. Ran, H. Zhang, and S. Gong, "Improving on LEACH Protocol of Wireless Sensor Networks using Fuzzy Logic", Journal oj Itiformation & Computational Science, vol. 7, no. 3, pp. 767-775, 2010. [9] Fan Yiming and Yu Jianjun, 'The Communication Protocol for Wireless Sensor Network about LEACH", In proceedings oj International
Fig, 2 Comparison of alive nodes for LEACH versus proposed models Number of Packets Delivered 700 "0
�
• .� .. Cl
Cotiference on Computational intelligence and Security Workshops (CISW 2007), pp. 550 - 553, Harbin, December 2007.
.':J
• .l< U • Q. ... 0
[10] Xiaoyan Cui, "Research and Improvement of LEACH Protocol in Wireless Sensor Networks", In proceedings oj International Symposium
on Microwave, Antenna, Propagation and EMC Technologies Jor Wireless Communications, pp.251-254, Hangzhou, August 2007. [II] Liu Jun-feng and Gao Sui-xiang, "Grid-based cluster partition method in wireless sensor networks", Computer Engineering and DeSign, vol. 28, no.9, pp. 2023-2025, 2007. [12] L.-S. Van, W. Pan, B. Luo, J.-T. Liu, and M.-F. Xu, "Communication Protocol Based on Optical LowEnergy-Adaptive-Clustering-Hierarchy (O-LEACH) for Hybrid Optical Wireless Sensor Networks",
650 600 550 500 450
·
• .D
400
E
, Z
350 300 10
12
14
16
18
Rouned Number
Communications and Photonics Cotiference and Exhibition (AC? 2009), pp. 1-6, Shanghai, November 2009. [13] Khalid A. Darabkh, Shereen S. Ismail, Mohammad AI-Shurman, Iyad F. Jafar, Eman Alkhader, and Mamoun F. AI-Mistarihi, "Performance
Fig, 3 Comparison of number of packets delieverd for LEACH versus proposed models
4
20
