Sunteți pe pagina 1din 5

Proceedings of 2015 Global Conference on Communication Technologies(GCCT 2015)

Simulative Study of Random Waypoint Mobility


Model for Mobile Ad hoc Networks
Aniket Pramanik1 , Biplav Choudhury2 , Tameem S. Choudhury3 , Wasim Arif4 , J. Mehedi5
Department of Electronics and Communication Engineering
National Institute of Technology, Silchar
Silchar, Assam-788010, India
aniketpramanik@yahoo.co.in1 , biplav93@gmail.com2 , salmantameem360@gmail.com3 , arif.ece.nits@gmail.com4 ,
j.mehedi@gmail.com5
AbstractConnectivity of Mobile Ad-hoc Networks (MANET)
is highly dependent on the mobility model it follows. From
previous studies, it has been found that Random Waypoint
Mobility Model (RWP) is widely used to simulate performance
of the MANETs. This mobility model resembles the mobility
of people using mobile phones in a given area. Connectivity
is an important issue in a MANET following RWP model and
hence researchers are trying out different ways to enhance
connectivity. So, this model is considered and the variation of
connectivity for different varying parameters is studied. The RWP
model is simulated in MATLAB, followed by the study of the
node connectivity by changing different parameters like number
of user nodes, simulation time and simulation area. Further,
it is shown that connectivity is enhanced in each case after
introducing agent nodes. This study gives an estimate of number
of nodes required for achieving desired connectivity of the network.
Keywords: Random Waypoint Mobility Model, Connectivity,
MANET, Agent nodes

I.

I NTRODUCTION

Mobile Ad hoc Networks (MANET) are wireless networks


which are infrastructure-less and are used mostly in battleelds
and railways. Unlike wired or antenna based networks, these
involve devices like mobile phone communicating to each
other directly without any intermediate entity in a deployed
area [1]. It is a dynamic network where each device is a node
and follows a random or denite mobility pattern.
Mobility determines the location of each node with respect to
the other nodes at different points of time. The connectivity
between the nodes depends on the relative position of each
node with respect to the other nodes, which in turn affects the
signal strength, data rate and the shortest path between them.
Therefore, mobility pattern is a vital factor which affects the
performance of the network. Some popular mobility patterns
that have been studied are: Gauss-Markov Mobility Model,
Random Walk, Random Waypoint Mobility Model and City
Selection Model in[2], [3], [4], [5], [6] and the most common
out of them is Random Waypoint Mobility Model.
According to this model, each node selects a random destination, and travels towards it in a straight line at a randomly
chosen uniform velocity. After reaching its destination, it halts
for a specic time known as pause time [4] and the process
repeats itself. Here, each node is implemented within a xed
area. Every node selects its destination, as well as, moves randomly within the area. Average connectivity is obtained after
each beacon interval. RWP model is widely used to simulate

978-1-4799-8553-1/15/$31.00 2015 IEEE

MANETs, since it emulates the condition where people use


mobile phones in a given area, while at the same time they
move randomly within the area. With random mobility of each
node, connectivity is an issue which is to be dealt with. Hence,
RWP model is used here.
This paper presents a study of variation in connectivity of the
nodes with variation in different RWP mobility parameters.
Also, it compares these variations in connectivity in the presence, as well as, absence of agent nodes. The parameters that
have been considered are: number of user nodes, simulation
time of the RWP model and the simulation area of the network.
It has been shown that connectivity is greater in presence of
agent nodes irrespective of other parameters.
The remaining sections of the paper are organized as follows:
Section-II contains a review of related work in this eld
followed by Section-III which consists of the objective of the
paper and denition of various parameters used. Section-IV
lists the simulation results in the form of plots of connectivity
versus different parameters and nally Section-V concludes
the paper.
II.

R ELATED WORK

Performance of a MANET always depends on the way


users move in the network area. Since, any user in real life
while communicating may move in any direction at a random
velocity, RWP has been the major mobility model studied.
Initially, the behaviour of RWP model was studied in terms
of spatial distribution of nodes where structure of node distribution was obtained as weighted sum of parameters like
static, pause time and mobility component for each node
[7]. William Navidi and Tracy Camp in [8], studied the
varying probability distribution of the location and speed of the
nodes and determined the stationary distribution for location,
speed and pause time of the RWP model. Later, stationary
distribution was further investigated in [9] in which it was
integrated to a normalized constant and then the ad-hoc network was analysed for connectivity and trafc. However, RWP
has not been considered as a perfect representation of human
mobility in the real world, so in [10], a survey of a person in
Melbourne has been done to validate some characteristics of
RWP. RWP was found to have some limitations like decreasing
average node velocity [11]. The authors in [12] therefore,
have investigated Time-Based Random Waypoint (TBRWP)
mobility model in which, motion time is chosen independent of
path length. Therefore, each node selects velocity proportional
to its trajectory length. On using NS-2 simulator, it was found



Proceedings of 2015 Global Conference on Communication Technologies(GCCT 2015)

that its movement generator gave border effect and on applying


quadrats count statistical testing in [2], the effect was reduced.
Xiang, Liu and Kuang have further analyzed and gave a
connectivity formula for a two-dimensional MANET which
eliminated the border effect [13].
In recent studies, for a non-uniform RWP, inter-arrival time has
been studied and estimated using two schemes namely Grid
and Monte Carlo estimation techniques [14]. As it was found
in [11], decay of average node velocity had lead to inaccurate
simulations so it has been modied in [15]. Here, Moraes
and Pontes have proposed a method by which node speeds
are chosen from BETA(2,2) distribution which stabilizes the
mean node speed. The lifetime of nodes based on the energy
left in them has also been investigated. Authors, Colletti and
Moraes in [16] have evaluated the residual link lifetimes of
each node in a network following RWP model and have found
that links remain active longer for mobility in steady-state
region. This idea can be used to develop new communicating
protocols for MANETs. Latest research has been done on the
performance evaluation of various routing protocols used in
MANETs following RWP model. In the latest conference paper
published in this eld [17], performance of routing protocols
like Ad hoc On-Demand Distance Vector (AODV), Dynamic
Source Routing (DSR) and Dynamic MANET On-demand
(DYMO) protocols have been evaluated and compared for
variable pause times in RWP mobility model.
Therefore, due to the need of study of connectivity in RWP
model, the study of connectivity between nodes and its
variation with total number of nodes, simulation time and
simulation area has been taken up.
III.

sion range. Connectivity of the network is calculated


as
C=[

A. Parameter denition

User node (UN) : A node which demands network


services.

Agent node (AU): A node which helps user nodes for


better network services. These nodes are intelligent
nodes. They rush according to the position of the user
nodes in the network to improve connectivity.

Communication Range: The maximum distance between two nodes to perform single hop communication.

Connectivity: It is the number of directly connected


mobile nodes to a specic mobile host in its transmis-

978-1-4799-8553-1/15/$31.00 2015 IEEE

(1)

where f(i, j, t) = 1 if there is a path between the ith


and jth users at time t, and f(i, j, t) = 0, otherwise[18],
[19].
The connectivity of a network calculated using the
given formula always ranges from 0 to 1, which means
C belongs to [0, 1].

Network Connectivity Time: It is the time interval


after which connectivity of the network is calculated.

Average Network Connectivity: It is the mean of


all the connectivity values of the network calculated
after each Network Connectivity Time throughout the
simulation.

B. Calculation of Connectivity
The connectivity of the network is calculated using the
above described formula, after each network connectivity time
or interval which is dened as 1 second. So, n number of
connectivity values are obtained for a simulation time of n
seconds.
After obtaining n connectivity values, average connectivity is
calculated by nding the mean of these values given as,
C(avg) =

O BJECTIVE AND PARAMETER DEFINITION

MANETs are expected to operate in a highly dynamic


environment. Therefore, network performance depends on the
extent to which routing protocols adapt to the topology dynamics. It is inadequate to use only one model. While evaluating a
protocol, it is run on various models to see how its performance
changes with different models. Thus the performance of a
specic protocol varies if underlying mobility models are
different. Thus, routing protocols are inuenced by different
mobility models in different ways. The objectives of this paper
are: i) to simulate the Random Waypoint Mobility Model ii) to
study the node connectivity by changing different parameters
(number of user nodes, simulation time, simulation area) and
iii) to enhance the connectivity by introducing special kind of
node called agent node which is described in the next section

UN 
UN

2
f (i, j, t)]
][
U N (U N 1) i=1 j=i+1

IV.

Sum of n connectivity values


n

(2)

S IMULATION RESULTS

The simulator is designed in MATLAB in WINDOWS


environment. Simulation is carried out to study the connectivity
of the network varying i) number of nodes ii) simulation time
iii) length of the simulation area and iv) width of the simulation
area.
A. Study I
In Fig. 3, variation in the connectivity of the network with
increase in the total number of user nodes has been plotted.
There are two plots for two different cases.
In the rst case, along with the user nodes, agent nodes have
also been introduced in the network. Agent nodes are meant
for connectivity between two user nodes which are far away
from each other and hence require an intermediate node for
better data transfer services. Fig. 1 shows the positions of the
user and agent nodes after a simulation time interval of ten
seconds in a simulation area of 10 metres x 5 metres. There
are four user nodes and two agent nodes in the simulation area.
The red circles are user nodes while the blue circles are agent
nodes. It can be seen that the agent nodes are near the user
node which is far away from the remaining three user nodes.
So, both of these nodes help the user node to connect to the
remaining three user nodes.
In the second case, only user nodes have been introduced in
the network. Fig. 2 shows the positions of the user nodes after
a simulation time interval of ten seconds in a simulation area
of 10 metres x 5 metres. There are four user nodes. All the
parameters considered for the network connectivity calculation



Proceedings of 2015 Global Conference on Communication Technologies(GCCT 2015)

have been entered in Table 1.


So, in both the cases, average connectivity has been calculated.
It has been calculated ten times in each case for ten different
number of user nodes varying from 5 to 50 and plotted in the
graph. The blue graph represents the rst case, in which the
agent nodes are present and the red one represents the other
condition where there are only user nodes and no agent nodes.
The conclusions that can be drawn are:
TABLE I.

I NPUT PARAMETERS

Simulation Parameters
Number of Agent nodes
Simulation time
User node speed interval
Agent node speed interval
Pause interval
Walk interval
Direction interval
Simulation area dimensions
Mobility model
Network connectivity time

Values
2
10 seconds
0.2 to 0.9 metres/second
0 to 0.9 metres/second
0 to 1 second
1 to 3 seconds
-180 to +180 degrees
10 metres x 5 metres
Random waypoint
1 second

Fig. 3.

Connectivity vs Number of user nodes

Overall, the connectivity in presence of agent nodes


is better.

It is seen that connectivity, mostly increases with


increase in user nodes as expected but decreases at few
instances which proves randomness of the mobility
model.

The connectivity, in each case, is found to be constant


with the number of user nodes beyond forty ve. It
gives the approximate amount of user nodes, which is
forty ve, required in a simulation area of 10 metres
x 5 metres to maintain best possible connectivity.

B. Study II

Fig. 1.

Snap shot of the network connectivity using agent nodes

In Fig. 4, variation in the connectivity of the network with


increase in the simulation time has been plotted. There are
two plots for two different cases as described earlier, that is,
simulation in presence and absence of agent nodes. All the
parameters considered for the network connectivity calculation
have been entered in Table 2.
So, in both the cases, average connectivity has been calculated.
It has been calculated nine times in each case for nine different
simulation times varying from 10 seconds to 50 seconds and
plotted. The blue graph represents the rst case, in which the
agent nodes are present and the red one represents the other
condition where there are only user nodes and no agent nodes.
The conclusions that can be drawn are:
TABLE II.

I NPUT PARAMETERS

Simulation Parameters
Number of Agent nodes
Number of User nodes
User node speed interval
Agent node speed interval
Pause interval
Walk interval
Direction interval
Simulation area dimensions
Mobility model
Network connectivity time

Fig. 2.

Snap shot of the network connectivity without using agent nodes

978-1-4799-8553-1/15/$31.00 2015 IEEE

Values
2
20
0.2 to 0.9 metres/second
0 to 0.9 metres/second
0 to 1 second
1 to 3 seconds
-180 to +180 degrees
10 metres x 5 metres
Random waypoint
1 second

Overall, the connectivity in presence of agent nodes


is better.

Connectivity, in each case neither strictly increases nor


strictly decreases, rather it follows a random pattern



Proceedings of 2015 Global Conference on Communication Technologies(GCCT 2015)

Fig. 5.
Fig. 4.

Connectivity vs Simulation area length

Connectivity vs Simulation time

as seen from the graph. This is due to the randomness


in the motion of the user nodes.
C. Study III
In Fig. 5, variation in the connectivity of the network with
increase in the length of the simulation area, keeping its width
constant, has been plotted. The width of the area has been kept
10 metres. There are again two plots for two different cases as
described earlier. All the parameters considered for the network
connectivity calculation have been entered in Table 3.
So, in both the cases, average connectivity has been calculated.
It has been calculated nine times in each case for nine different
lengths of the simulation area varying from 2 metres to 18
metres and plotted. The blue graph represents the rst case, in
which the agent nodes are present and the red one represents
the other condition where there are only user nodes and no
agent nodes. The conclusions that can be drawn are:
TABLE III.

D. Study IV
In Fig. 6, variation in the connectivity of the network with
increase in the width of the simulation area, keeping its length
constant, has been plotted. The length of the area has been
kept 10 metres. There are again two plots for two different
cases as described earlier. All the parameters considered for
the network connectivity calculation have been entered in
Table 3.
So, in both the cases, average connectivity has been calculated.
It has been calculated nine times in each case for nine different
widths of the simulation area varying from 2 metres to 18
metres and plotted. The blue graph represents the rst case, in
which the agent nodes are present and the red one represents
the other condition where there are only user nodes and no
agent nodes.
The conclusions that can be drawn are:
TABLE IV.

I NPUT PARAMETERS

Simulation Parameters
Number of Agent nodes
Number of User nodes
Simulation time
User node speed interval
Agent node speed interval
Pause interval
Walk interval
Direction interval
Simulation area breadth
Mobility model
Network connectivity time

Values
2
20
10 seconds
0.2 to 0.9 metres/second
0 to 0.9 metres/second
0 to 1 second
1 to 3 seconds
-180 to +180 degrees
10 metres
Random waypoint
1 second

Again, the connectivity in presence of agent nodes is


better.

Connectivity, in each case decreases with increase in


length of the simulation area. This is mainly because
increase in simulation area length leads to more space
availability for the movement of the nodes. Hence, the
nodes get more scattered across the simulation area
leading to reduced connectivity.

978-1-4799-8553-1/15/$31.00 2015 IEEE

I NPUT PARAMETERS

Simulation Parameters
Number of Agent nodes
Number of User nodes
Simulation time
User node speed interval
Agent node speed interval
Pause interval
Walk interval
Direction interval
Simulation area length
Mobility model
Network connectivity time

Values
2
20
10 seconds
0.2 to 0.9 metres/second
0 to 0.9 metres/second
0 to 1 second
1 to 3 seconds
-180 to +180 degrees
10 metres
Random waypoint
1 second

Again, the connectivity in presence of agent nodes is


better.

Connectivity, in each case, in general, decreases with


increase in width of the simulation area. This is mainly
because increase in simulation area width leads to
more space availability for the movement of the nodes.
Hence, the nodes get more scattered across the simulation area leading to reduced connectivity. However,
in some cases, connectivity has also increased which
is due to the randomness in the motion of the nodes.



Proceedings of 2015 Global Conference on Communication Technologies(GCCT 2015)

[9]

[10]

[11]

[12]

[13]

[14]
Fig. 6.

Connectivity vs Simulation area width


[15]

V.

C ONCLUSION

This paper studied the connectivity of the network of


MANETs considering Random Waypoint Mobility model by
changing number of nodes, simulation time, length and width
of the simulation area. It also studied the connectivity if agent
nodes are used. It gives an approximate idea of minimum
number of nodes required to be deployed to achieve network
connectivity of particular area if Random Waypoint Mobility
model is considered. This paper also shows the usefulness
of agent nodes to improve connectivity if Random Waypoint
Mobility pattern of the nodes are considered. Authors wish to
extend this work to nd a suitable algorithm for movement
control of the nodes in distributed manner in emergencies to
establish better connectivity.

[16]

[17]

[18]

[19]

E. Hyytia, P. Lassila, and J. Virtamo, Spatial node distribution of the


random waypoint mobility model with applications, Mobile Computing, IEEE Transactions on, vol. 5, no. 6, pp. 680694, 2006.
A. Rojas, P. Branch, and G. Armitage, Validation of the random
waypoint mobility model through a real world mobility trace, in
TENCON 2005 2005 IEEE Region 10. IEEE, 2005, pp. 16.
J. Yoon, M. Liu, and B. Noble, Random waypoint considered harmful,
in INFOCOM 2003. twenty-second annual joint conference of the IEEE
computer and communications. IEEE societies, vol. 2. IEEE, 2003,
pp. 13121321.
A. Nayebi, M. Rahimi, and H. S. Azad, Analysis of time-based random
waypoint mobility model for wireless mobile networks, in Information Technology, 2007. ITNG07. Fourth International Conference on.
IEEE, 2007, pp. 4247.
H. Xiang, J. Kuang et al., Minimum node degree and connectivity
of two-dimensional manets under random waypoint mobility model,
in Computer and Information Technology (CIT), 2010 IEEE 10th
International Conference on. IEEE, 2010, pp. 28002805.
T. Wang and C. P. Low, Determine the inter-arrival time in the nonuniform random waypoint mobility model, in Wireless Days (WD),
2010 IFIP. IEEE, 2010, pp. 15.
R. M. de Moraes, F. P. de Araujo, and A. S. Pontes, A proposal
to stabilize the random waypoint mobility model for ad hoc network
simulation, in Wireless Communications and Networking Conference
(WCNC), 2010 IEEE. IEEE, 2010, pp. 16.
R. R. Colletti and R. M. de Moraes, Evaluation of link lifetime for
the random waypoint mobility model, in Networks (ICON), 2013 19th
IEEE International Conference on. IEEE, 2013, pp. 15.
A. K. Maurya, D. Singh, A. Kumar, and R. Maurya, Random waypoint
mobility model based performance estimation of on-demand routing
protocols in manet for cbr applications, in Computing for Sustainable
Global Development (INDIACom), 2014 International Conference on.
IEEE, 2014, pp. 835839.
A. Konak, O. Dengiz, and A. E. Smith, Improving network connectivity in ad hoc networks using particle swarm optimization and agents,
in Wireless Network Design. Springer, 2011, pp. 247267.
A. Konak, G. E. Buchert, and J. Juro, A ocking-based approach to
maintain connectivity in mobile wireless ad hoc networks, Applied Soft
Computing, vol. 13, no. 2, pp. 12841291, 2013.

R EFERENCES
[1]

[2]

[3]
[4]
[5]

[6]

[7]

[8]

Z. Ismail and R. Hassan, A performance study of various mobility


speed on aodv routing protocol in homogeneous and heterogeneous
manet, in Communications (APCC), 2011 17th Asia-Pacic Conference
on. IEEE, 2011, pp. 637642.
R. Alghamdi, J. DeDourek, and P. Pochec, Avoiding border effect in
mobile network simulation, in ICN 2013, The Twelfth International
Conference on Networks, 2013, pp. 184189.
V. A. Davies et al., Evaluating mobility models within an ad hoc
network, Masters thesis, Citeseer, 2000.
D. B. Johnson and D. A. Maltz, Dynamic source routing in ad hoc
wireless networks, in Mobile computing. Springer, 1996, pp. 153181.
J. Broch, D. A. Maltz, D. B. Johnson, Y.-C. Hu, and J. Jetcheva, A
performance comparison of multi-hop wireless ad hoc network routing
protocols, in Proceedings of the 4th annual ACM/IEEE international
conference on Mobile computing and networking. ACM, 1998, pp.
8597.
B. Liang and Z. J. Haas, Predictive distance-based mobility management for pcs networks, in INFOCOM99. Eighteenth Annual Joint
Conference of the IEEE Computer and Communications Societies.
Proceedings. IEEE, vol. 3. IEEE, 1999, pp. 13771384.
C. Bettstetter, G. Resta, and P. Santi, The node distribution of the
random waypoint mobility model for wireless ad hoc networks, Mobile
Computing, IEEE Transactions on, vol. 2, no. 3, pp. 257269, 2003.
W. Navidi and T. Camp, Stationary distributions for the random
waypoint mobility model, Mobile Computing, IEEE Transactions on,
vol. 3, no. 1, pp. 99108, 2004.

978-1-4799-8553-1/15/$31.00 2015 IEEE



S-ar putea să vă placă și