LOCALIZATION ALGORITHM USING VARYING SPEED MOBILE SINK FOR WIRELESS SENSOR NETWORKS

International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.1, No.3, September 2010
DOI :10.5121/ijasuc.2010.1311 130
LOCALIZATION ALGORITHM USING VARYING
SPEED MOBILE SINK FOR WIRELESS SENSOR
NETWORKS
B.Amutha 1
, Dr. M. Ponnavaikko2
, N.Karthick3
and M.Saravanan3
1
Research Scholar, School of Computer Science and Engineering, SRM University
bamutha62@gmail.com
2
Provost, SRM University
ponnav@gmail.com
3
B.E[CSE], Valliammai Engineering College,Anna University
b.d.best.4ever@gmail.com
3
M.Tech[CSE], SRM University
Wellwish1989@gmail.com
ABSTRACT
Localization of sensor nodes is important in many aspects in wireless sensor networks. The known
location of sensor node helps in determining the event of interest. A mobile sink is introduced to track the
event driven sensor nodes in the path of the event, thus conserving energy and time. We present a novel
range based localization algorithm which helps the mobile sink to compute the location of the sensor
nodes efficiently. The data transfer from the mobile sink and the sensor nodes is used to estimate the
sensor location. The sensor nodes do not need to spend energy on neighbouring interaction for
localization. The localization mechanism has been implemented in TOSSIM. The simulation results show
that our scheme performed better than other range-based schemes.
KEYWORDS
Mobile Sink, Location sensing, Range based location scheme, Beacon Communication range, IDSQ
1. INTRODUCTION
Due to advances in wireless technologies, Wireless Sensor Networks have been used for many
applications such as military, home, health, industries etc. Wireless Sensor Network is a
dynamically changing localized network formed on-the-fly as mobile nodes move in and out of
each others' transmission ranges [1]. Actually, the mobile ad hoc networking model followed by
wireless sensor networks make no assumption that nodes know their own locations. However,
recent research on localization shows that location-awareness can be beneficial to fundamental
tasks such as routing and energy-conservation [2]-[16].Because of the deployment of a large
number of sensor nodes specific to a particular application, it is often not possible to hand place
these sensor nodes. Fundamental design objectives of sensor networks include reliability,
accuracy, flexibility, cost effectiveness and ease of deployment [3]. It is not possible to keep
Global Positioning System (GPS)for all the nodes. The capabilities of individual Sensor nodes
are extremely limited due to many factors and their collaboration is required with minimum
energy expenses. This paper proposes a mechanism that allows non-GPS equipped nodes in the
network to derive their exact locations from a limited number of GPS-equipped nodes. In this
method, a mobile sink is introduced, which has the prior knowledge about the sensor field. The
mobile sink is used for data collection whenever any event of interest occurs. It travels in the
sensor field in a randomized way with dynamic speed. During its travel, it broadcasts its
location information to all sensor nodes within the sensor field. A mobile sink can compute the

131
location of sensor nodes when it finds the entry and exit point over the communication circle of
the sensor node, if the sensors are homogeneous. This scheme only requires broadcasting of
beacon messages by the mobile sink and performs accurately in spite of irregular radio
transmission due to obstacles.
1.1 Related Work
Several schemes have been proposed to locate the events which are broadly classified into two
categories such as range-based and range-free. Range Based Localization introduced ways to
localize nodes based on range (i.e. distance or angle) information. Several mechanisms require
more hardware on sensor nodes in order to obtain higher location accuracy. Localization
protocols either make simplifying assumptions (e.g., line of sight with sensor nodes, high
density of anchor nodes, deployment knowledge) or require sophisticated hardware [2]-[4].
On the other hand, Range free localization approaches expect no knowledge about
communication range to be available; instead they include other metrics to calculate a node’s
position. The approaches typically need a large amount of stationary reference points for
achieving higher accuracy and extensive communication among neighbouring sensor nodes [5]-
[9].
Over the years, many systems have addressed the problem of automatic location sensing.
Triangulation, scene analysis, and proximity are the three principal techniques for automatic
location sensing. One of the most well known location-based systems is the Global Positioning
System (GPS), a satellite-based navigation system made up of a network of 24 satellites placed
into orbit. GPS is widely used to track moving objects located outdoors. However, GPS, as it is
satellite dependent, has an inherent problem of accurately determining the location of objects
located inside buildings. Different approaches have been proposed and tested for their
effectiveness and utilities in order to achieve the ability to locate objects within buildings. At
present, there are several types of location-sensing systems, each having their own strengths as
well as limitations. Infrared, 802.11, ultrasonic, and RFID are some examples of these systems.
1.1.1 Using Mobile Beacons
Many schemes have been introduced, using mobile sink as a beacon node, in estimating the
location of sensor nodes in a wireless sensor network. One of the scheme proposed by Sichitiu
and Ramadurai, uses a single mobile beacon transmitting its current location, to find the
location of unknown sensor node [12]. Any sensor node receiving the beacon packet will be
able to infer that it must be somewhere around the mobile beacon with a certain probability.
This information constrains the possible locations of a node. Combined with the RSSI
technique, possible locations of the sensor node can be estimated. The accuracy can be
improved when the sensor node receives more beacons. The precision of the localization is good
and uniform as long as the trajectory of the beacon covers the entire deployment area in such a
way that each sensor node receives at least three non-collinear beacon messages.
The Scheme proposed by Dutta and Bergbreiter, estimates the distance from a sensor node to a
mobile object based on ultrasound technology [13]. Neighbouring sensor nodes cooperated to
evaluate the distance between them by exploiting common tangent concept. As long as the
node-to-node distances are available, the position of a sensor node can be measured by the
communication range of sensor nodes in range-based schemes.
Sun and Guo proposed probabilistic localization scheme with a mobile beacon [14]. The
approaches use time of arrival[TOA] technique for ranging and utilize Centroid formula with
distance information to calculate a sensor node’s position. The above approaches need to
integrate with communication range information for localization. Galstyan et al. proposed a
coarse-grained range-free localization algorithm to lower the uncertainty of node positions using

132
radio connectivity constraints [15]. With each received beacon, the receiver’s location lies
within the transmission area of the sender. Another localization scheme proposed in [16] in
which mobile anchor points are used as reference points for localization in wireless sensor
network. The scheme presented in this paper makes use of the communication range of the
sensor nodes as a reference for locating the points of mobile anchor point. The mobile beacon is
assumed to know its location (through GPS). However, in their scheme three distinct beacon
points are needed and a well-defined mobility of the mobile beacon is required. Further, our
scheme makes use of only two beacon points in estimating the location of the sensor nodes.
Another scheme proposed by Saad Ahmed Munir, Yu Wen Bin and Ma Jian, the approach was
based on the communication range of the sensor nodes [17]; and with the help of the motion of
a mobile sink in a straight line, the localization of the sensor nodes is estimated. A geometrical
topology representation of the sensor nodes was implemented by the mobile sink based on the
communication range of each sensor node. The inter-node distance was calculated by the mobile
sink with reference to its line of motion.
Bin Xiao, Hekang Chen and Shuigeng Zhou presented two distributed range-free localization
methods (ADO and RSS) that use only one moving beacon within a sensor network [18]. The
basic idea of the ADO method is to narrow down the possible location of a node by using
arrival and departure constraint areas derived from the moving beacon. To estimate the position,
a node should obtain four critical positions of the moving beacon: pre arrival position, arrival
position, departure position and post-departure position to compute its ADO. The RSS method
seeks the location of a sensor from a sequence of RSS broadcasted from the moving beacon.
Both methods employ the range-free techniques in the sense that they do not rely on the direct
distance measurement from RSS but on a sequence of constraints. Ecolocation for the ideal
scenario considered zero multi-path fading and shadowing effects [19], Relative location
estimation algorithms are implemented in a wireless sensor network deployed in indoor and
outdoor environments [20-21].
The Remainder of this paper is organized as follows. Section 2 identifies the objectives and the
requirements of localization algorithm. Section 3 presents Factors affecting location estimation.
Section 4 presents the Location Estimation Algorithm using varying speed mobile sink. Section
5 presents the performance analysis. Section 6 presents the conclusion.
2. Objectives and Requirements
An efficient and accurate range-based scheme is proposed here, with the help of a mobile sink.
This comes up with the near-accurate co-ordinates of sensor nodes within a wireless sensor
network. Simulations are conducted for varying sink velocity and for different packet generation
interval for obtaining minimum localization errors. Similarly, the performance is demonstrated
by varying radio communication range error.
In this section, localization mechanism for estimating the location of sensor nodes within a
wireless sensor network is described. This approach is based on the communication range of the
sensor nodes; and with the help of the motion of a mobile sink, the localization of the sensor
nodes is achieved. Some of the assumptions that have been made are;
1. The mobile sink moves at a randomized speed initially.
2. It changes its speed when any event occurs, according to the position of the information
driven sensor node.
3. The velocity of the mobile sink does not remain constant while moving through the sensor
network.

133
This scheme is independent of the use of RSSI or GPS in finding the location of sensor nodes
and relies only on the communication range of the sensor node. This implementation is not only
cost effective but is also simple in implementation and can locate the sensor. A mobile beacon
has been used in many proposed localization schemes wherein reference points are used for
estimation of sensor.
2.1 Proposed Scheme
In this paper, a novel range based scheme for efficient localization of sensor nodes within a
network with a varying speed mobile sink is explained in detail. The scheme is based on the
mobility of a mobile sink introduced within the network. It is assumed that the mobile sink
moves in a randomized fashion initially and then the speed and the direction of the mobile sink
changes according to the event of interest. With the known communication range of the sensor
nodes, the mobile sink is able to estimate the location of each and every node in the
geographical field.
2.1.1 Event Detection Mechanism
A randomly moving target is introduced as an event, which has to be detected by the set of
sensor nodes in the path of the target. The occurrences of this event must be made known to the
mobile sink by the sensor nodes. And this will be achieved by the collaboration among the
sensor nodes. It is not necessary that the mobile sink should be near to the event driven node.
Information driven sensor querying mechanism is followed by the sensor nodes in sensor
selection criteria. Information-driven sensor querying (IDSQ) is a sequential tracking scheme
where, at a given point of time t, there is only one sensor active. All the other nodes remain in
power-conserving sleep states. The active sensor takes a measurement and updates the belief
p(xt
|zt
). It then decides which sensor in its neighbourhood is the most informative, hands the
belief off to that sensor and returns to the sleep state. The sensor receiving the handoff becomes
active and this operation repeats. Intuitively, by selecting the most informative neighbour, the
active sensor is seeking good quality data.Our aim is to locate the sensor nodes which are
tracking the event of interest. The mobile sink has to travel nearer to the location of the event
driven sensor, so that the sensor can estimate its location and combine it with the target tracking
information. Therefore an autonomous speed mobile sink is introduced to estimate the location
of the sensor node.
2.1.2 Varying Speed Mobile Sink Mechanism
The mobile sink speed is not necessarily the same as that of the target speed. It must be lesser
than or equal to the speed of the target. Once the event is known to the mobile sink, the varying
speed mechanism is adopted. It varies its speed according to the location of the event driven
sensor. Mobile sink is assumed to have prior knowledge about the sensor field. The mobile sink
is able to move by them selves or other carriers such as RC plane. The following figure shows
the mobile sink broadcasting its Information Driven Sensor Querying message [IDSQ]to know
the event driven node’s location according to its communication range. Sensor nodes were
expected to give acknowledgement containing the information about the event that occurred
within the sensor field.
Figure 1: Mobile sink broadcasting its IDSQ message to all sensors, to identify the information
driven sensors in the sensor field.

134
If any node detects the target, it starts sending continuous periodical target information
messages to mobile sink. The neighbour nodes within the area of that event driven node will
also receive the target information. The neighbour nodes will also start sending the same
message. Inter node collaboration is executed. Two possibilities are taken into consideration for
analyzing our algorithm. In the first case, if mobile sink is nearer to the event driven node, then
it can come close and collect sensor data. In the second case, if the mobile sink is far away from
the event driven sensor node, called as source node, then mobile sink gets the notification about
the target through the other nodes. To illustrate this concept, in figure 2, the target was detected
by the node S4, and takes a measurement. Now S4 wants to give notification to mobile sink by
acknowledging to its beacon message. The acknowledgement message contains target
notification data and its information. It then decides which sensor in its neighbourhood is the
most informative to handoff the acknowledgement, say S3. Now S3 appends its information to
the received packet from S4 and hand off to the next informative neighbourhood. This process
repeats until it reaches the mobile sink. Once the sink detects the notification, it instantly
changes its speed level and move towards the even driven node S4, since the mobile sink knows
the path towards the source node by means of received acknowledgement packet. Thus the
speed of the mobile sink will change from one level to another level according to the event of
occurrence in the sensor field. If there is no event detection, then the mobile sink is in one speed
level. If any detection of event takes place, its speed will be in another level.
Figure 2: The event driven node S4, gives notification about the target to mobile sink, by
deciding which sensor in its neighbourhood is the most informative in order to handoff the data
in the sensor field..
Our algorithm, comes up with area localization within a wireless sensor network and has an
advantage of localizing the desired area. The desired region to be localized is traversed by the
mobile sink. This scheme also facilitates the location estimation of the places of interest and
thereby minimizes the energy and time it takes to localize the whole network.
Assumptions:
• The mobility is controllable and thus predictable,
• The pause time of the mobile sink along its moving trace is negligible.
Localization compensates all the existing problems through the following characteristics:
1. A varying speed varying mechanism is introduced in the mobile sink.
2. A variable RF communication range is taken into consideration.
3. The number of beacon points is reduced to two for the mobile sink.
4. The path of the mobile sink is also varied so as to overcome the limitation of the
existing localization scheme.

135
5. Sensor nodes placed farthest from the mobile sink should not face greater error if they
lie within the communication range.
6. Line error should be reduced for the sensor nodes placed on the path of the mobile sink.
7. Only one reference point (Mobile Sink) is introduced which knows the entire
heterogeneous network topology.
8. Node mobility is kept uniform.
9. Target speed is not a criterion to collect the optimized information about the target
10. Sensors must record the communication details of the mobile sink.
3. Factors Affecting Location Estimation
The following parameters were varied to check the performance of MOBISPEED.
3.1 Communication Range
The communication range has considerable effect in the estimation of localization. For varying
communication ranges the difference between actual and estimated locations of sensor nodes
were calculated and tabulated.
3.2 Packet Interval
The periodical intervals between packet transfer from mobile sink and sensor node are also
varied. As a performance measure, the intervals between the packets also have considerable
effect.
3.3 Varying Mobile Sink Velocity
The accuracy of location estimation increases as the number of times the mobile sink visits the
unknown sensor nodes. For varying mobile sink velocity, the algorithm gives better results.
3.4 Radio Frequency
One of the problems of using RF to locate objects is the inconsistency of the signal strength
reception. This can primarily be due to the environment and the device itself. In most cases, the
environmental factors always have the most impact on the accuracy and maximum detectable
range. For convenience the environmental factors are not considered in Mobispeed, as it uses
only the distance to estimate location of sensor nodes.
3.5 Varying the Path of Mobile Sink
The path of the mobile sink within the sensor field is varied. The impact of path variation on
localization error is made negligible by making the mobile sink to move nearer to the event of
interest in a randomized fashion.
4. Location Estimation
4.1 Assumptions And System Environments
Fig. 1 and 2 illustrates the system environment where a sensor network consists of sensor nodes
and mobile sink. The sensor nodes are distributed randomly in the sensing field. Once the nodes
are deployed, they will stay at their locations for sensing tasks. The sensor nodes can receive
messages from both other nodes and from mobile sink. The mobile sink is able to traverse
through the sensor field with sufficient energy and broadcast beacon messages during the

136
localization process. The mobile sink is able to move by themselves or other carriers such as RC
plane or vehicles.
For simplicity the network environments are given below.
a. Sensor nodes are S1 to Sn..
b. One Mobile Sink M.
c. One Moving Target for Event of Interest.
d. Communication range of sensor nodes is r.
e. All sensor nodes form a sensor field E.
f. Sensors can communicate with each other within the communication range of r.
g. Mobile sink communicate with all sensor nodes, in turn sensor node
communicate to mobile sink, only when the sensor start sending the
acknowledgement packet to the mobile sink.
4.2 Localization Mechanism
The center of the circle is the location of the sensor node; the radius of the circle is the largest
distance where the sensor node can communicate with the mobile sink. The endpoints of the
chord are the position where the mobile sink passes through the circle (Entry and Exit point). In
the mechanism, two endpoints (Entry & Exit) on the circle must be collected for establishing a
chord. Mobile sink periodically broadcast beacon message when it moves in the sensor network.
The beacon message contains the mobile sink id, location and timestamp.
4.3 Location Calculation
As, Mobile Sink knows the entire topology information, based on the entry & exit of the mobile
sink over the communication range of the sensor node, it calculates the location of the sensor
node based on the simple geometrical equations. Suppose, let (a, b) be the location coordinates
of the sensor node to be estimated. Let (x1 ,y1) be the entry location point of the mobile sink and
(x2, y2) be the exit location point of the mobile sink based on the communication range of the
sensor node, whose location (a, b) to be estimated. Then the following equations will help to
estimate the (a, b) as follows,
Where r represents the radius, which is communication range of sensor node. Solving the above
two equations will give the two possible locations. Sensor node may be resided at either one of
the estimated location. The most informative sensor must provide the target information to the
mobile sink since the sensor nodes which sense the target information have to inform to the
Mobile Sink. The exception of locating the most Informative sensor lies on both parts of the
chord, as per the mathematical calculation. The exception to be overcome is identified by the
location of the most informative sensor. As the information received by the mobile sink through
the event driven sensor detects the location of the sensor and also the location of the target.
Orientation of the target will be analyzed for every event occurrence (target movement). The
stored information in the mobile sink and the sensor node give the sensor location.
Algorithm:
Mobile Sink: 1. Generate Location Information Packet Li.
2. Send it to the sensor nodes S1 to Sn. (To all nodes)
3. Receive Ack from (Si-Sj)
(From a set of nodes which can hear)
Sensor Node: 3. Save Location Information Li as Ei
(Entry Point).

137
(Si - Sj) 4. Send Ack to Mobile sink
(Sensing region).
5. Update Li to Li+1.
6. Go to step 4
7. If Update = Lost
Stop Ack..
8.Store Last Li as Ej (Exit Point).
9.Use distance formula to estimate
location coordinates.
Corollary:
The most informative sensor node may lie either above or below the chord. To identify the
Location of the most informative sensor, the following steps have been executed.
1. There may be two sensors in two possible locations.
2. Say S1 is in the upper part and S2 is in the lower part of the chord.
3. If the target is in the upper part of the sensor, S1 will transmit the signal along with its
id and target information to the mobile sink.
4. Mobile sink receives this information in due course of time as the mobile sink moves in
the sensor field with varying speed capability.
5. As mobile sink captures the information from the sensor nodes directly and it is able to
capture the target information from the most informative sensor S1.
6. Else it has to capture from S2.
Thus the localization algorithm comes up with efficient localization of sensor nodes and our
mechanism reduces needed power consumption compared to other schemes
This mechanism only requires mobile sink to broadcast beacon messages. The ordinary sensor
nodes do not spend energy on neighbouring interaction for localization.
4.4 Beacon Points and Information Base in Mobile Sink
This mechanism utilizes mobile sink that move around in the sensing area and periodically
broadcast beacon messages, including the current location information. After sensor nodes
receive the beacon message, it should give acknowledgement packet containing, its id, location
of the mobile sink found in beacon message. The acknowledgement packet is received by the
mobile sink when it enters into the communication range of that sensor node.
The point it enters into the communication range and the point it exit the communication range
of the sensor node is taken into account. In between entry and exit, there may be periodical
communication between mobile sink and sensor node take place. But, only the entry and exit
point were considered and thus the valid beacon points and chords will be determined. Here
only these two beacon points act as input. And these two endpoints on the circle were stored in
the information base by the mobile sink for efficient localization of the sensor node in the future
while it visits the same sensor node once again.
Entry and exit points:
The sensor node once it heard the signal from the mobile sink, it starts acknowledging the
mobile sink. The entry point is ascertained by sending the first acknowledgement to the mobile

sink. After the first acknowledge
as long as it is receiving the lo
information from the mobile sink
instant the sensor is not able to
received packet from the inform
the sensor ascertains the exit poin
The mobile sink is designed to
sensor node in one complete r
possibility that the mobile sink c
it visits a particular sensor node
that estimated information into the
Figure 3: Selection of Beacon Po
If suppose mobile sink found tha
possible locations will be estimat
in near ranges and these only two
coordinates mostly differ in far
come up with a conclusion in esti
Let us see the illustration. Consi
communication range of the sen
beacon points. The beacon point
communication circle. In the fig
(x2, y2) as exit point. The entry
node can communicate with ea
communication between them ta
circle, the mobile sink can able
the mobile sink, it can get two
only one of the two possible l
information base by the mobile s
ement is over, the sensor keeps on sending acknow
cation information packets from the mobile sink. T
nk is stored in an information base of the sensor
o receive the packets from the mobile sink, it selec
mation base, as the departure packet of the mobile si
nt.
traverse the sensing area in such a way that it sho
round it takes. During its one complete round, th
can visit some of the sensor nodes one or more time
e, it will estimate the location of the sensor node and
he information base.
oint (Entry & Exit of mobile sink) - first possibility.
at it visited a sensor node two times, then by our alg
ted. Out of four, two location coordinates may be sa
o points were considered by the mobile sink. Other
ranges and it will be skipped. Thus the mobile sink
timating the exact location of the sensor node.
ider the figure 3, where the mobile sink is traveling
nsor node S. The mobile sink is designed to maint
t is considered as an approximate endpoint on the se
gure 3, the mobile sink has chosen (x1, y1) as entr
y point is the point from which the mobile sink and
ach other and the exit point is the point at wh
akes place.With the chosen two points on the com
to estimate the possible location of sensor nodes. A
possible locations and the sensor node is expected
locations. These estimated locations were then st
ink with sensor node’s id.
138
wledgements,
The location
node. The
ects the last
ink and thus
ould visit all
here may be
es. Each time
nd will store
gorithm, four
ame or differ
two location
nk is able to
ng through the
tain a set of
ensor node’s
ry point and
nd the sensor
hich the last
mmunication
According to
d to reside in
tored in the

Let the two possible estimated
location is identified in one of
node once again or by the capab
section. (refer corollary in previ
sink visited the same sensor node
Figure 4: Selection of Beacon Po
again – second possibility.
This time the entry and exit poin
sensor node S using our mechan
In order to come into conclusion
same sensor node before during
take both the previously estimate
possible estimated locations, two
these two will be taken into acco
be differ with wide ranges and th
up with exact location of the se
mentioned mechanism will be
assumed to be static. For the situ
estimate the location of the senso
there is a possibility to reside on
Selection of the beacon points is
localization of the sensor nodes
the communication circle. How
environments, incorrect beacon
intervals. The chords generated us
sensor. Figure 5 displays selecti
errors. Almost this can be avo
increasing the number of beacon
location by the mobile sink be (a, b) and (c, d)
two ways. Either when the mobile sink visits the
bility of most informative sensor node as mentioned
ious section). The next possible assumption is that
node S, as shown in the following figure 4.
oint when the mobile sink visits the same sensor node
nt will differ and the mobile sink can estimate the loc
nism. Again it can get two possible location of the
n, it will check the information base whether it had
g the same current round. If it found that it had vi
ed two possible locations and the present estimation.
o location coordinates must be same or differ in few
ount by the mobile sink and the remaining two coor
hus it will be skipped. Finally, the mobile sink can a
ensor node, possible with few estimation errors. And
suitable only for the situation where the sensor
uation where the sensor nodes are mobile, the mob
or node by the capability of most informative sensor
both side of the chord.
s the most important criteria in this localization algo
will be accurate when the selected beacon points ar
wever there are situations, particularly, in case
n points could be chosen due to collision or i
using the beacon points thus fails to estimate the po
tion of incorrect beacon points that leads to large
voided by decreasing the beacon broadcasting i
messages.
139
). The exact
same sensor
d in previous
t the mobile
node S once
cation of the
sensor node.
d visited the
isited, it will
. Out of four
w meters and
rdinates must
able to come
nd the above
nodes were
obile sink will
r node when
gorithm. The
are exact on
of practical
inappropriate
osition of the
localization
interval and

Figure 5: Incorrect selection of b
The more the beacon messages
selection of entry and exit point
enhanced beacon selection mech
radio transmission due to obstacl
5 Performance Evaluation
The sensor field for simulation
field 12 sensor nodes were place
randomly in the field. The mob
field at the beginning, as illustra
sensing area. It will move through
change according to the situation
autonomous speed varying mec
communication range of the sens
A sensor network as shown in fi
traverse through the sensor field
the direction of the mobile sink in
Fi
beacon points leads to large localization errors – third
and less the beacon broadcasting interval will lead
t, that is, beacon points on the communication circ
hanism, Localization is performed accurately in spite
les.
on
was a square of 100 * 100 m2. For simplicity, in
ed. In the practical environment, the sensor nodes we
bile sink was randomly introduced at any corner of
rated in figure 6. The mobile sink cannot be placed
gh the sensor field with a different speed. The level o
on in the sensor field. Thus the mobile sink is design
chanism. The localization scheme was analyzed w
sor node.
igure 7 is implemented wherein a mobile sink was in
in a random direction at various levels of speed. Fo
nitial movement is shown in the figure.
igure 6: Simulation Environments
140
d possibility
d to efficient
cle. With the
e of irregular
n the sensing
ere deployed
f the sensing
d within the
of speed will
gned with an
with different
ntroduced to
or simplicity,

141
Figure 7: Area localization for sensor nodes with a mobile sink
Figure 8:Node11-Sink-Node12
Figure 9: Ns2 NAM of MobiSpeed 4

142
Node 1 Sink
Figure 10:Ns2 NAM of MobiSpeed 4
Node11 Sink
Figure 11:Ns2 Simulation of MobiSpeed4

143
Figure 12: Packet Drop
Figure 13: Packet Drop-Minimum
Performance results are presented for the difference in location estimation based on our
implementation and the actual location of the sensor nodes. Simulations were conducted to
study the effect of two values, that is, the sink speed and induced radio communication range
error. Simulations were conducted for different radio communication range errors are presented.
Before that, the information base maintained by the mobile sink about the location estimation of
the twelve sensor nodes of our simulated environment of figure 7 is shown in Table 1 and also
Table 2 for different radio range.he location estimation error (e) is calculated as follows:
Distance between the sensor node’s
real coordinates (x0, y0) and the computed Coordinates (x, y), given by

144
The average location estimation error is calculated as,

F
Location Estimation of Sensor Node
as pe
Similarly, simulation results for
figures 8, 9, 10 and 11. These i
range error as part of the theoreti
was set at 10, 20, 25 and 30m. T
of the localization of sensor nod
performance degrades and is wo
range error, that is 0.15, the perf
schemes. Simulation of the prop
farthest from the mobile sink pa
have a fewer error as shown in al
Figure 14: Simulation Parameters
Figure 15: Simulation Results
de S1, based on Entry & Exit point of Mobile Sink in the
er the simulation, (10, 108) & (10, 52)
r different radio communication range error are p
induced errors correspond to varying degree in comm
ical known communication range. In our simulation,
The induced error, of course has some effect on the p
odes. With an increased induced communication range
orse for the high value of range error as 1.22. For a l
rformance of our proposed scheme is greatly better t
posed localization scheme shows that the sensor nod
ath and the sensor nodes placed on the path of the
ll graphs.
145
e sensor field
presented in
mmunication
, radio range
performance
ge error, the
low value of
than existing
nodes placed
mobile sink

146
Thus the distance of the mobile sink from the sensor nodes is also a factor to have
little influence upon the accuracy of the location estimation of sensor nodes. Finally,
communication overhead in localization algorithm is only included the broadcasting
information driven sensor querying (beacon) message.Figure 12-19 exhibits the localization
error graphs.
Figure 16: Localization of sensor nodes for radio communication range error 0.15
Figure 17: Localization of sensor nodes for radio communication range error 0.20

147
Figure 18: Number of generated packets at all the nodes. x-source node, y-destination node
Figure 19: Number of dropped packets at all the nodes. x-source node, y-destination node
Analysis
The speed of the mobile sink is varied for different packet intervals and the performance
analysis is made. The increase in mobile sink speed and the increase in packet interval size
made the localization error also to increase.
6. Conclusion
In this paper we have presented a range based localization scheme, MOBISPEED, without using
constraints or angle information. The scheme was designed to come with efficient localization
of sensor nodes within a sensor network. With the help of mobility of mobile sink presented
within the network, location estimation is performed. Based on the location information from
mobile sink and the principles of elementary geometry, the mobile sink can compute the sensor
nodes position without additional interactions. It is assumed that the mobile sink moves in a
network in random fashion. It can vary its speed level according to the situation of event
occurrence in the sensor field. Based on the selection of beacon points, the mobile sink can
comes up with an efficient localization of the sensor nodes.
The localization mechanism that proposed here reduces needed power consumption compared
to other schemes. Our mechanism only requires that mobile sink broadcast beacon messages.
The ordinary sensor nodes do not spend energy on neighbouring interaction for localization.
Simulations conducted in TOSSIM [22-23] and NS2 verify that MOBISPEED performs
efficiently and is accurate in its functionality. As in” An Efficient Directed Localization
Recursion Protocol for Wireless Sensor Networks” a different scheme in which only two
reference points are required in order to estimate a position was discussed.[24]This can also be

Mobile Sin
(m/
Packet
Interval
(ms)
0.3
millisecond
0.6 milliseco
useful in predicting location coord
Bahi, J.M et al proposed that
optimum mobile beacon trajecto
proposed a simple localization
Critical to the location acc
transmission range of the be
Theoretical analysis shows tha
estimation error when the traver
B.Xiao et al in [26]. Simulation
various communication ranges
mechanism can be shortened if
interval and increasing the nu
with an efficient localization
localization schemes.
Table3: Average Estimation
Our future work involves the si
due to the number of dropped pac
References
[1] Feng Zhao, Leonidas Guibas, "
Kaufmann publishers, 2004.
[2] J. Hightower,G. Boriello, and
based on RF signal strength,”
[3] Radu Stoleru John A. Stankov
Networks” EmNets'07, June 2
[4] P. Bahl and V. N. Padmanabh
system,” in INFOCOM, Mar 200
[5] Nissanka B. Priyantha, Anit C
6th ACM International Con
Boston, MA, August 2000.
[6] R. Klukas, M. Fattouche, “Line
Environment”, IEEE Transactio
[7] N. Bulusu, J. Heidemann, and D
devices,” IEEE Personal Comm
[8] D. Niculescu and B. Nath, “D
Systems, January 2003.
[9] T. He, C. Huang, B. M. Blum,
large scale sensor networks,” in
2003.
[10] Jagoba Arias *, Jesu´s La´zaro
algorithm for wireless sensor n
nk
/s)
10 meter/second 4 meter/second
1.004167 0.3725
ond 1.065833 0.40833
dinates of an unknown sensor node.
t a solution has been given to the question "W
ory and when should the beacon packets be sent?"
on algorithm which was based on one mobile b
ccuracy of sensor nodes are two parameters,
eacon, and how often the beacon broadcasts it
at these two parameters determine the upper bou
rse route of the beacon is a straight line had been ex
ons were conducted with varying mobile sink veloc
of sensor nodes. The execution time for the l
the moving speed is increased and by reducing t
umber of beacon messages. Our localization sche
with minimum error as compared to other r
on Error at Different Mobile Sink Speed and Packet In
ignal strength effects and the impact of localization
ckets by the sensor node.
"Wireless Sensor Networks - an information processing approach
d R.Want, “SpotON: An indoor 3D location sensing technolo
Feb. 2000.
ovic Sang Son “ Robust Node Localization for Wireless Sen
25-26, ACM ISBN 978-1-59593-694-3/07/06, 2007.
han, “RADAR: an in-building RF-based user location and tracki
2000.
Chakraborty, and Hari Balakrishnan” The Cricket Location -Supp
nference on Mobile Computing and Networking (ACM
e-of-Sight Angle of Arrival Estimation in the Outdoor Multipat
ons on Vehicular Technology, July 2003.
D. Estrin, “GPS-less low cost outdoor localization for very sma
munication., Oct 2000.
DV based positioning in ad hoc networks,” Telecommunicatio
, J. A. Stankovic, and T. Abdelzaher, “Range-free localization
n Proc. ACM Int. Conf. Mobile Computing Networking (MOB
o, Aitzol Zuloaga, Jaime Jime´nez, Armando Astarloa “GPS
networks” Computer Communications 30 (2007) 2904– 2916.
148
What is the
", and then
beacon.[25]
the radio
ts position.
und of the
xplained by
city and for
localization
the beacon
eme comes
range-based
nterval
n accuracy
h", Morgan
ogy
nsor
ing
upport System”
M MOBICOM),
th
all
on
n schemes for
BICOM), Sep.
S-less location

[11]A. Nasipuri and K. Li, “A dir
in First ACM Internationa
2002.
[12] M. L. Sichitiu and V. Ramad
Center for Advances Computing
[13] P. Dutta and S. Bergb
bsac.eecs.berkeley.edu/~sbergbr
[14]G.-L. Sun and W. Guo, “Comp
with a mobile beacon”, Network
[15] A.Galstyan, B. Krishnamachar
networks using a mobile target”
[16]Kuo-Feng Ssu, Chia-Ho Ou, H.
Sensor Networks”, May 2005.
[17] Saad Ahmed Munir, Yu Wen B
Sensor Network with a Mobile
IEEE, 2007.
[18]Bin Xiao, Hekang Chen and
Wireless Sensor Networks”, IEE
[19]Kiran Yedavalli,Bhaskar Krishn
Sequence Based Technique fo
Technical Report number 2004-
[20] Neal Patwari, Alfred O. Hero,
Estimation in Wireless Senso
AUGUST 2003.
[21]Mohamed Youssef and Naser
Deployment Issues” Fifth An
(CNSR'07) IEEE 2007.
[22]Philip Levis and Nelson Lee, "T
[23]Philip Levis, Nelson Lee, Mat
Simulation of Entire TinyOS
USA.
[24]H. A. B. F. de Oliveira, “An E
Sensor Networks,” IEEE Transa
[25] J. M. Bahi, A. Makhoul, “A
International Conference on Wi
[26]B. Xiao, H. Chen, S. Zhou, “
Sensor Networks,” IEEE Trans
2008
Authors
B.Amutha received her M.E.degree from Mad
Context Aware Computing in Wireless Senso
years of teaching experience. She is now serv
Engineering Department, SRM University, C
bamutha62(at)gmail.com
Murugesan Ponnavaikko is a graduate in El
Engineering from the college of Engine
Distribution System Planning from I.I.T (
Technologist, Administrator, Academic and
Board in 1972 and had seventeen years of ev
Power Sectors in India and abroad.Presently
ponnav@gmail.com
N.Karthick is a B.E[CSE] student and interested in the field of wireless sensor networks. His interests
include positioning,tracking and navigation using WSN.
Matlab software.
b.d.best4ever@gmail.com
M.Saravanan is currently working on localization using mobile sinks. Hia current interests
with centimenter accuracy by varying the packet size and packet interval
Wellwih1989@gmail.com
rectionality based location discovery scheme for wireless sens
al Workshop on Wireless Sensor Networks and Appli
durai, “Localization of wireless sensor networks with a mob
g Communications, Jul. 2003.
breiter, “MobiLoc: Mobility enhanced localization”,
re/publications/CS294.pdf
parison of distributed localization algorithms for sensor networ
king, Sensing and Control, March 2004.
ri, K. Lerman, and S. Patterm, “Distributed online localizati
”, Information Processing in Sensor Networks, 2004.
H. C. Jiau, “Localization with Mobile Anchor Points in Wireles
Bin, and Ma Jian, “Efficient Minimum Cost Area Localization
e Sink”, International Conference on Advanced Networking and
Shuigeng Zhou, “A Walking Beacon-Assisted Localization i
EE Communications Society, ICC 2007 proceedings.
namachari, Sharmila Ravula, Bhaskar Srinivasan“Ecolocation: A
for RF Localization in Wireless Sensor Networks” USC CEN
-16.
Matt Perkins,Neiyer S. Correal, and Robert J. O’Dea,” Rela
or Networks” IEEE Transactions on signal processing, VOL
r El-Sheimy” Wireless Sensor Network: Research vs. Realit
nnual Conference on Communication Networks and Servi
TOSSIM: A Simulator for TinyOS Networks", September 2003.
att Welsh, and David Culler, “TOSSIM: Accurate and Scalab
Applications", SenSys’03, November 5–7, 2003, Los Angele
Efficient Directed Localization Recursion Protocol for Wireles
actions on Computers, vol. 58, no. 5, pp. 677-691, May 2009.
Mobile Beacon Based Approach for Sensor Network Localiz
ireless and Mobile Computing, Networking and Communication
“Distributed Localization Using a Moving Beacon in Wireles
sactions on Parallel and Distributed Systems, vol. 19, no. 5, Ma
dras University in 2002. She is doing research in
or Networks. She has 2 years of industrial experience and 23
ving as an Assistant Professor, Computer Science and
Chennai, Tamilnadu, India.
lectrical Engineering and a post graduate in Power System
neering, Guindy He received his PhD degree in Optimal
(Delhi) in 1983. He has proven record of excellence as
Leader. He started his carrier from Tamil Nadu Electricity
eventful service as Consultant and Administrator in the Public
he is serving as Provost at SRM University, Chennai
N.Karthick is a B.E[CSE] student and interested in the field of wireless sensor networks. His interests
and navigation using WSN. Also he is working on Object classification using
M.Saravanan is currently working on localization using mobile sinks. Hia current interests Include positioning
with centimenter accuracy by varying the packet size and packet interval
149
sor networks,”
lications, Sept.
obile beacon,”
, http://www-
rk
tion in sensor
ss
n for Wireless
Applications,
in
A
NG
lative Location
L. 51, NO. 8,
ality Design and
ices Research
ble
es, California,
ss
alization,” IEEE
ns, 2007.
ss
ay

LOCALIZATION ALGORITHM USING VARYING SPEED MOBILE SINK FOR WIRELESS SENSOR NETWORKS

More Related Content

What's hot (17)

Similar to LOCALIZATION ALGORITHM USING VARYING SPEED MOBILE SINK FOR WIRELESS SENSOR NETWORKS (20)

Recently uploaded (20)

LOCALIZATION ALGORITHM USING VARYING SPEED MOBILE SINK FOR WIRELESS SENSOR NETWORKS