вход по аккаунту



код для вставкиСкачать
A Load Balanced Routing Protocol for Mobile
Sensor Network
Sukanya Chatterjee
Dept. of C.S.E
UIT Burdwan
Avijit Chakraborty
Lead QA Analyst
Abzooba India Infotech Private Limited
West Bengal, India
[email protected]
Shilpa Pareek
Dept. of C.S.E
UEM Jaipur
Rajasthan, India
[email protected]
Mohit Dokania
Dept. of E.E.E
IEM Kolkata
[email protected]
Rajasthan, India
[email protected]
[email protected]
Somen Nayak
Dept. of C.S.E
UEM Jaipur
Ashutosh Gautam
Dept. of C.S.E
UEM Jaipur
Rajasthan, India
[email protected]
Ratul Dey
Dept. of C.S.E
UEM Jaipur
Rajasthan, India
[email protected]
Hriday Banerjee
Dept. of C.S.E
UEM Jaipur
Rajasthan, India
[email protected]
Himadri Nath Saha
Dept. of C.S.E
IEM Kolkata
West Bengal, India
[email protected]
Mudassir Neyaz
Dept. of E.E.E
IEM Kolkata
[email protected]
management cost [12]. Information gathering is a fast growing
and challenging field in today’s world of computing. Sensors
provide a cheap and easy solution to these applications
especially in the inhospitable and low-maintenance areas
where conventional approaches prove to be very costly.
Sensors are tiny devices that are capable of gathering physical
information like heat, light or motion of an object or
environment. Sensors are deployed in an ad-hoc manner in the
area of interest to monitor events and gather data about the
environment. Networking of these unattended sensors is
expected to have significant impact on the efficiency of many
military and civil applications, such as combat field
surveillance, security and disaster management. Clustering
sensor nodes is an effective method to increase network
scalability and lifetime; it facilitates the distribution of control
over the network, and also saves energy and reduces network
contention by enabling locality of communication: a plain
node communicates its data over shorter distances to its
cluster head (CH). Then the CH aggregates these data into a
smaller set of meaningful information. Not all nodes, but only
Keywords— Wireless Communication, Clustering, Mobile Nodes, the CHs need to communicate to the base station (BS).
Therefore, the CHs are in charge of aggregating data and
Routing, Scalability, Network
transmitting the aggregated data to the BS [10]. Load
balancing of cluster heads in cluster based approaches lessens
A wireless sensor network (WSN) is typically composed of the load of CH leading to higher network life time. Traditional
large numbers of nodes called sensors. These sensor nodes are load balancing methods concentrates on shifting of load to
severely constrained in terms of cost, computational power, additional load to the additional gateway nodes commonly
battery power, and communication bandwidth. Besides these known as additional cluster head (ACH). This course of load
limitations, these sensors leverage the idea of creating a sensor balancing leads to consumption of higher amount of energy
network composed of large number of nodes to be deployed in for selection of ACH and selection of routing path to ACH.
difficult to access areas. A WSN can be deployed in two This paper aims to curtail the energy consumption by
manners: either structured or unstructured. In unstructured formation of nested cluster with the additional number of
way nodes may be deployed in an ad-hoc manner into the nodes responsible for augmentation of load threshold. The
field and remain unattended till the end of network lifetime. In rest of the paper is organized as follows. Section 2 includes a
this type of deployment number of nodes is large and because detailed survey of the related literatures. Section 3 gives a
of that maintenance is a challenge. In a structured WSN nodes brief overview of minimum system requirements. Section 4
are deployed on a pre-planned resolution. This type of gives brief description about the proposed method. Result and
deployment leads to lower network maintenance and
Abstract— Wireless Sensor Networks (WSN) have gained
worldwide attention in recent years, particularly with which has
facilitated the development of smart sensors. WSN consists of
small nodes with sensing, computation and wireless
communication capabilities. In WSN, clustering is used as an
effective technique to achieve scalability, self-organization, power
saving, channel access, routing etc. Lifetime of sensor nodes
determines the lifetime of the network and is crucial for the
sensing capability. Clustering is the key technique used to extend
the lifetime of a sensor network. In this proposed method, the
node next to the present cluster head is selected as the new
cluster head if its energy is greater than the threshold energy.
Moreover all the nodes used in this method are mobile nodes.
Mobile nodes increases the scope of gathering updated
information. Mobile nodes also have a chance of entering remote
area where cluster head cannot reach them. In this case the
nodes die. In the proposed method this problem is solved,
whenever the nodes tends to move out of the cluster, the nodes
generates a random angle to which it is redirected. This
approach will increase the network lifetime and will provide high
978-1-5386-2215-5/17/$31.00 ©2017 IEEE
its simulation discussions Are discussed in section 5.Section 6
concludes the paper and in section 7 the references are given.
The WSN literature is based on different Routing
Protocol for Static Nodes and Mobile Nodes. Low
Energy Adaptive Clustering Hierarchy (LEACH) [1]
first implemented the dynamic clustering method in
Wireless Sensor Network. LEACH protocol has
two phases set up phase and steady phase. In the set
up phase the cluster heads are chosen. In the steady
phase data transmission between the nodes are done
maintain the cluster head. Communication between
the cluster head is done by TDMA. LEACH
protocol doesn’t predict the location of the nodes
since it doesn’t support the mobility of the sensors.
.Leader Election with load Balancing Energy
(LELE)[2] Protocol compares the amount of energy
and distance of node with the neighbouring sensors
to elect it as leader. A WSN contains a great
number of sensors which are connected through
radio frequencies (RF). The architecture of sensor
networks is such that allows the sensors to
randomly distributed in an area and recognize, and
finally link them to a station called sink. In this
manner the load is balanced in this protocol.
Load Balancing bases on nodes distribution in
mobile nodes [3] protocol of load balancing, a
coordinator is static cluster head that collects data
from the associated mobile sensor nodes. From the
simulation results, it is seen that load balancing
scheme successfully balances the distribution of
mobile sensor nodes that attached to each
coordinators. This therefore balances the energy
consume among the coordinator and indirectly
delay in data reception is reduced there are large
mobile nodes in the system. The load balancing
scheme is achieved by applying it to a cluster
topology, therefore Mobile nodes in the
neighborhood do not need to broadcast there data to
all neighbor, but only to its coordinator. It is based
on Zig-Bee cluster tree. Dynamic Load Balancing
Protocol(DLBP) [4] has succeeded to build a load
balanced tree, eliminate the need for control
messages during data routing, keep the load of
WSN balanced during data routing, send messages
to next hops without route discovery delay, quickly
maintain and fix network errors and failures. DLBP
provides a technique the keep the load balancing of
WSN even sensors start collecting data from the
environment and sending them to sink.
. The coverage based inter – cluster
communication (CBIC) [5] approach comprises of
three distinct phases. The initialization phase in
which when the network starts for the first time
each node broadcast its location, range and the area
they cover. Each node then builds a table of
neighbors followed by the set up phase and steady
phases which is similar to LEACH [1]. In the
initialization phase, the proposed algorithm selects
those nodes in the network those sensing area
coverage is covered by its neighbors. If the sensing
area of 1 node is fully embraced by the union set of
its neighbors, that is neighbors can cover the current
node’s sensing area, this node becomes a temporary
cluster head without reducing the system overall
sensing coverage and later use these temporary
cluster heads in multi hop inter cluster
communication. All the nodes in the network
follow the same procedure and hence terminate in
O(n) times this follows the random selection cluster
heads which one selected broadcast advertisement
messages. Depending on the message strength each
nodes makes the decision for joining a particular
cluster. This phase use the CSMA MAC protocol
and during this period all the nodes are listening.
During each cycle the cluster head selection is
random and is depends on the amount of energy left
in the node and its probability of not being a cluster
head during the n rounds. Once the cluster heads
are selected and clusters are formed, the algorithm
then forms two layer of the network: the top layer
which will comprise of the temporary cluster heads
and the bottom layer which will comprise of
member nodes of the cluster after the set up phase
the steady phase (transmission phase) starts in
which all the nodes transmit data using TDMA
based scheduling. When all the nodes in the cluster
complete transmission the cluster head performs
computation on it and sends it to the base station
via multi hop using the temporary cluster heads.
The role of the cluster head is rotated randomly
after some pre – defined times.
In Performance Evaluation of Load Balanced
Clustering [6] network set up is performed in two
stages “Bootstrapping“ and “clustering”. In the
Bootstrapping, gateways discover the nodes that are
located within their communication range.
Gateways broadcast a message indication the start
of clustering. We assume that the receivers of
sensors are open throughout the clustering process.
Each Gateway starts the clustering at a different
instant of time in order to avoid collisions. In reply
the sensor also broad cast a message with their
maximum transmission power indication their
location and energy reserved in this message. Each
node discovered in this phase is included in a per –
Gateway range set. In the clustering phase,
Gateways calculate the cost of communication with
each node in the range set. This information is then
exchanged between all the gateways. After
receiving the data from all other gateways each
gateways start clustering the nodes based on the
communication cost and the current load on its
cluster. When the clustering is over, all the sensors
are informed about the id of the cluster they belong
to. Since gateways share the common information
during clustering, each sensor is picked by only one
For inter – Cluster communication all traffic is
routed through the gateways. PRMN: Predictive
Location Based Routing for Mobile Nodes[7]
supports mobility of sensors by implementing the
capability to predict their future location and it’s
also receive predictive location of neighbors within
its range. After the sensor nodes are deployed in the
target area the base station sends a frame to all the
sensor nodes which includes the location of the
base station. Sensor performs following operation
to predict data transmission path first a sensor node
predicts its future location and transmits it to the
neighbors within its range. It also transmits its
future distance with the base station. After
receiving the future location from neighbor, it
calculates the distance between the future location
of itself and the neighbors. It selects the neighbor
having minimum future distance and minimum
distance from base station.
a) All the sensor nodes are GPS enabled i.e. the nodes are
position aware.
After cluster formation, cluster head selection is held.
There are r rounds of data transmission. Each sensor node has
unique identification number, i.e. ni where i= 1 to n. Initially
all the nodes have similar energy of 2.0 nJoule and threshold
energy of each node is 0.1 nJoule.
The load balanced routing protocol for mobile nodes
works based on following assumptions.
b) All the nodes are mobile with velocity v and moves
with a random angle [0-2ʌ].
c) All the sensor nodes have same energy at the time of
d) Base Station (BS) is unique and stationary with endless
The mobile nodes in WSN have the capacity to predict
their future location and it also receives predictive location of
neighbors within its range. After the sensor nodes are
deployed in the target area the base station sends a frame to all
the sensor nodes which includes the location of the Base
A position aware sensor node transmits data by performing
following operations [7] a) Predicting its future location, a sensor node sends it to
the neighbor nodes within its range as well as the base station.
b) Gathering all the information about future location of
neighbor nodes, it calculates the distance between its own and
neighbor’s future location.
After deployment the sensors form multiple clusters in the
sensing zone. The number of clusters is dependent on –
1) Number of sensor nodes,
2) The area of the sensing zone
Illustration of cluster creation method
The co-ordinates of target area are xmax, xmin (according to
x axis) and ymax, ymin (according to y axis). The area is
calculated as
Total Area= {(xmax-xmin)*(ymax-ymin)} square unit.
The number of clusters is c which are formed by one of the
cluster formation method.
x co-ordinate of cluster ci is Xci and y co-ordinate is Yci.
The number of mobile sensor nodes deployed is n. The coordinates of each node according to x and y axis are –
For 1st node, x co-ordinate is Xn1 and y co-ordinate is Yn1.
For 2nd node, x co-ordinate is Xn2 and y co-ordinate is Yn2.
For nth node, x co-ordinate is Xnn and y co-ordinate is Ynn.
Now, after deployment of sensor nodes corresponding
clusters are formed according to their co-ordinates.
Illustration of cluster head selection method
In ci cluster, the number of nodes are n. The cluster head is
ni for ith round.
For first round, cluster head is n1.
After a certain amount of time interval, the cluster head is
n2, if it remains in the cluster zone. As the nodes are mobile,
they change their position with the time. If the node, which is
assigned to be next cluster head, goes out of its cluster, then it
floods a message to other nodes of corresponding cluster and
next available alive node having highest energy is declared as
cluster head. In this way, each of the nodes within a cluster
will get similar chance ( i.e. each node has probability of 1 ) to
be cluster node until and unless they are dead, i.e. they have
energy less than threshold value and the node does not remain
within same cluster.
position. Each node moves with constant velocity v and
random generated angle. This phase is based on
(PRMN) [7]. The algorithm is described below –
The proposed algorithm is divided into two phases.
1) Load balancing phase,
2) Routing phase
A. Load Balancing Phase
It is said earlier that the nodes are mobile. So
there remains possibility of each node to get out of
its cluster and moving on to another. In such cases,
the difference in number of outgoing nodes and
incoming nodes is calculated. Let, in a cluster,
number of incoming nodes is m and number of
outgoing nodes is k. If m is greater than k, then the
burden of the cluster head increases which leads to
the dissipation of cluster head’s energy in that
cluster. To solve this, the load balancing algorithm
is produced. In this case, (m-k) numbers of nodes
are reclustered to form new cluster with a new
cluster head.
Algorithm for Load Balancing
Input: incoming node list (X[max]), outgoing node list
(Y[max]), load balancing factor (Lfac);
Every incoming node (X[i]) sends a JOIN_MSG to the
cluster head (CH). Cluster head counts the no. of incoming
nodes from the no. of JOIN_MSGs. Every outgoing node
[Y[i]] sends an REMOVE_MSG to the CH.
1. If X[max] > Y[max] then,
2. CH calculates the Load Balancing Factor Lfac =
X[max] – Y[max].
3. A new cluster with Lfac no. of nodes will be formed
inside the present cluster.
B. Routing Phase
It is mentioned that the sensor nodes are mobile and position
aware. After a time interval the mobile nodes change their
Figure 1: No of alive nodes per round
The protocol has two phases:
a) Prediction phase
b) Transmission phase
Prediction phase:
A position aware sensor node is deployed in the location at
S(x,y). As it has a fixed velocity v and angle Ԧ [0-2ʌ], with
which it can move, according to Figure 1, after time t, S
node’s predictive location will be S’ with the coordinates:
Xt = X + |v| cosԦ * t
Yt = Y + |v| sinԦ * t
Using the above formula each node calculates its future
location. It also calculates the future distance from Base
Station (BS). After that it sends these information to other
neighbor nodes within its range and receives that information
of other neighbor nodes.
Transmission phase:
After sending and receiving the future locations of each
node as well as its future distance from BS in prediction
phase, each node calculates the distance from each neighbor
node by using Euclidian formula.
Dis = ¥(xt-xti)2+(yt-yti)2
Where (xt, yt) is the future location of a node and (xti, yti) is
the future location of a neighbor node (ni). Dis is the Euclidian
distance between these two nodes. Now, each node selects the
neighbor node which has minimum distance from the former
node and from the BS too. In each round these two phases i.e.
prediction and transmission phase occur using the algorithms
that are proposed in PRMN based protocol.
In order to evaluate the performance of the proposed protocol,
we simulated the algorithm using java with the parameters
shown in the table.
Network Size
Number of sensor nodes
Sensor node deployment
Random deployment
Percentage of cluster heads
Percentages of mobile censor 5-100%
Data size
2000 bits
Sensing range
1219 meters(CH),
400meters(member sensor)
Threshold value
0.1 joule
Initial energy
2.0 joule
Figure 3: Average energy per round
Average energy with respect to rounds is shown in Figure 3,
which depicts that energy decreament rate of the proposed
protocol is less than the LBCS(WRITE THE NAME), and
LELE(WRITE THE NAME). The proposed protocol achieves
the average energy of 1.999511265 joule after all the rounds.
Figure 4: Average energy of nodes after round 1
Figure 2: No of alive nodes per round
Figure 2 shows the numbers of alive nodes i.e. nodes have
their energy above the thresold monitored by the algorithm
and the state of the art algorithm. It is shown that, in the
proposed algorithm number of alive nodes per round is steady.
Up to round no 22 the number of alive nodes remains the
same and is equal to total number of nodes deployed . After
round no 22 no. of alive nodes decreases with a rate of
3.31%.The result shows that the no of alive nodes in the
proposed scheme is better than other cluster based load
balancing protocol . the reason behind the better consequence
is formation of nested cluster inside the target cluster province
which prohibits the cluster head from taking addition burden
of communication with the member nodes.
Figure 4 pronounces the the average energy of 1.99983478
joule of all nodes after round 1 is complete. In LELE Ch
election method is nearly same as what was followed in
LEACH; the only difference is consideration of residual
energy of the node for selection as a CH. The proposed
protocol exploits the mobility of the sensors by formation of
nester cluster with the nodes that involves extra burden to the
Figure 5: Average energy of nodes after round 2
M.J.Handy,M.Haas, D.Timmermaan “LEACH : Low
Energy Adaptive Clustering Hierarchy with deterministic
cluster head selection,” , IEEE , 2002.
[2] Mohammad
Faez,Mostafa Chhhardoli, “LELE : Leader Election with
Balancing Energy in Wireless Sensor Network,”
International Conference on Communication and Mobile
Computing IEEE2009 , 2009
[3] Karitnah
NazimJhamli,”Load balancing Based on nodes
Distribution in mobile sensor network”, 2011 7th
International Conference on IT, IEEE2011.
[4] Hamzed Aljawawdeh and Iman Almomani, “Dynamic
Load Balancing Protocol,” in Wireless Sensor Network,
IEEE Jodan Conference on Applied Electrical
Engineering and Computer Technologies , IEEE 2013.
[5] Nauman Israr and Irfan Awan,“Coverage Based
Intercluster Communication for Load Balancing,”,in
Wireless Sensor Network,21st International Conference
on advanced Information Networking and Application
[6] R. Nicole, “Performance Evaulation of Load Balanced
Clustering,”, inWireless Sensor Network,IEEE,2003.
[7] Rupam Some, and Indrajit Bannerjee, “PRMN: Predictive
Location Based Routing for Mobile Nodes,”,in Wireless
Sensor Network, International Conference on computing
and communication, IEEE,August 2014.
[8] Wendi
Chandrasekharan,and Hari Balakrishnan,”Energy Efficient Communication Protocol for wireless sensor
network”,Proceedings of 33rd Hawaii International
Conference on System Science ,2000.
[9] Fan Xianging and Song Yulin ,”Improvement on LEACH
International Conference on Sensor Technologies and
Applications ,IEEE ,20 October 2007.
[10] Younis, O.,&Fahmy,”HEED: A Hybrid Energy Efficient
,Distributed clustering approach for ad hoc sensor
nodes”,IEEE transaction on Mobile Computing ,2004.
[11] LBCS: A Load Balanced Clustering Scheme in Wireless
Sensor Networks- Shujuan Jin . 2009 Third
International Conference on Multimedia and Ubiquitous
Engineering. pp. 221. – 225.
[12] A. Kundu, R. Misra, A. Kar, S. Debchoudhury, S.
Pareek, S. Nayak, R.
Dey “On Demand Secure Routing
Protocol Using Convex-Hull & K-Mean Approach In Manet”
in proc. of 7th International Conference and Workshop on
Computing and Communication (UEMCON -2016), New
York City, USA.,IEEE Xplore Digital Library, October 2016,
pp. 1-5.
[13] R. Dey, H. N. Saha, ” Different Routing Threats and its
Mitigations Schemes for Mobile ad-hoc Networks (MANETs)
–A Review”, IPASJ International Journal of Electronics &
Communication (IIJEC) Vol.4 No.3 pp.27-34, March 2016.
[14] R. Dey, H. N. Saha, “Secure Routing Protocols For
Mobile Ad-Hoc Network (Manets) –A Review” International
Journal of Emerging Trends & Technology in Computer
Science (IJETTCS) Vol.5, No. 1, pp74-79, February 2016.
Figure 6: Average energy of nodes after round 3
Figure 5 and 6 depicts the average energy of all nodes after
round 3 and round 4. Average energy of all nodes after round
3 and round 4 is 1.9995470856 joule and 1.9993415537 joule
respectively which is less than LELE and LBCS. Most of the
traditional load protocols follow the approach of assigning the
additional load to a specialized node called Additional cluster
head (ACH) which leads to greater amount of energy
consumption. The process of using ACH is not only time
consuming but also not energy efficient due to the extensive
process of ACH selection and transmission of data to ACH.
The proposed method creates a cluster inside the existing one
reducing the burden of ACH selection and transmission to
Wireless sensor Network has opened a new large application
in our life and many researches emerges that need different
solution. In this paper, we have implemented the technique of
load balancing and routing path prediction for the mobile
nodes for the wireless sensor network. Load balancing feature
is considered one of the most important features for WSN
application and unique in this proposed protocol as compared
to most exiting load balancing protocol of WSN Stimulating
results shows that this protocol is more efficient is more
efficient than in terms of energy consumption, network life
and data transmission than those of exiting LEACH protocol.
The proposed algorithm for load distribution and routing
for mobile nodes in wireless sensor network is an approach for
balancing load of the cluster head (CH) of a cluster based
wireless sensor network. The algorithm balances the load of
CH by formation of a nested cluster with the additional nodes
due to which a cluster becomes unstable.
The proposed method of load distribution and routing can
be used in applications like surveillance of battle field,
surveillance of open cast mines.
We aim to consider variable speed of the mobile sensor
nodes in future. In addition to that, consideration for inclusion
of a routing access point in the every cluster and development
of accurate model for energy consumption is also our future
course of action
Без категории
Размер файла
508 Кб
2017, iemecon, 8079576
Пожаловаться на содержимое документа