International Journal of Pure and Applied Mathematics

Volume 118 No. 18 2018, 2063-2070

ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version)
url: http://www.ijpam.eu
Special Issue
ijpam.eu

FACE TRACKING ALGORITHM FOR TRACKING TARGET IN WSN

C. Selvarathi1 R.Sujatha2
1
Department of Computer Science and Engineering
M. Kumarasamy College of Engineering, Karur.
2
Department of Information Technology
M. Kumarasamy College of Engineering, Karur.

ABSTRACT

Target tracking in wireless sensor network (WSN) is most important in surveillance

applications. In existing work, a tracking framework called Face Tracking is used. A polygon
region called face is constructed. The nodes that are grouped inside a face can communicate only
among them. Brink detection algorithm is used to find a edge in which two nodes are connected
has the best coverage area. An Optimal selection algorithm is used to select the nodes which can
track the target with less energy usage. But if the target moves out of coverage area or if energy
of the tracking node becomes low then the target may not be tracked exactly. To overcome this
problem, a tracking scheme, called t-Tracking is proposed with an objective to achieve quality of
tracking (QoT). Distributed tracking algorithm sends queries about the energy level and coverage
area, to all the nodes in the face in which the target has to be tracked next. Based on the reply
from all nodes, a node with best energy level and coverage area node will be selected for target
tracking. Since a best node is selected, target can be tracked with accuracy.

Keywords - Wireless Sensor Network (WSN), Distributed Tracking Algorithm and t-Tracking.

INTRODUCTION Currently, wireless sensor networks are beginning to

A wireless sensor network is a collection of be deployed at an accelerated pace. It is not
nodes organized into a cooperative network. Each unreasonable to expect that in 10-15 years that the
node consists of processing capability, may contain world will be covered with wireless sensor networks
multiple types of memory (program, data and flash with access to them via the Internet.
memories), have a RF transceiver (usually with a
single unidirectional antenna), have a power source This can be considered as the Internet

(e.g., batteries and solar cells), and accommodate becoming a physical network. This new technology is

various sensors and actuators. The nodes exciting with unlimited potential for numerous

communicate wirelessly and often self-organize after application areas including environmental, medical,

being deployed in an ad hoc fashion. Systems of military, transportation, entertainment, crisis

1000s or even 10,000 nodes are anticipated. Such management, homeland defense, and smart spaces.

systems can revolutionize the way live and work. Since a wireless sensor network [5] is a distributed

2063
International Journal of Pure and Applied Mathematics Special Issue

real-time system a natural question is how many environmental properties, from geographical areas of
solutions from distributed and real-time systems can interest. The applications of WSNs can be found in
be used in these new systems? Unfortunately, very diverse fields such as survivable military surveillance
little prior work can be applied and new solutions are systems (e.g., battle field surveillance),
necessary in all areas of the system. environmental protection (e.g., habitat monitoring),
industrial monitoring (e.g., machine equipment
The main reason is that the set of monitoring), monitoring healthcare, personnel
assumptions underlying previous work has changed monitoring, home automation, and so on. One of the
dramatically. Most past distributed systems research most important areas where the advantages of WSNs
has assumed that the systems are wired, have can be exploited is tracking mobile targets. Methods
unlimited power, are not real-time, have user of tracking mobile targets have been gaining more
interfaces such as screens and mice, have a fixed set and more attention due to their importance in
of resources, treat each node in the system as very employing wireless sensor networks (WSNs) for
important and are location independent. In contrast, surveillance applications. Sensors deployed for
for wireless sensor networks, the systems are tracking schemes are capable of deducing kinematic
wireless, have scarce power, are real-time, utilize characteristics, such as position, velocity, and
sensors and actuators as interfaces, have dynamically acceleration of single or multiple targets of interest.
changing sets of resources, aggregate behavior is
important and location is critical. Many wireless 2.1 Quality of Tracking
sensor networks also utilize minimal capacity devices In many practical scenarios similar, the
which places a further strain on the ability to use past movements of a target are relevant only to local areas
solutions. and for a short period of time. This implies that such
a scenario requires fast tracking operation and also
2. PREVIOUS WORKS
the high quality of tracking (QoT). This QoT can be
A sensor is a low cost device that detects used as the quality of service (QoS) in a tracking
changes in the environment and records the changes. system. On the one hand, if one wishes to achieve
It is typically capable of sensing, computing, and tracking a target by sensor nodes that are already
communication. A large number of sensors can organized into local groups (unlike dynamic clusters
collaborate to form a wireless sensor network or trees) before the network starts tracking operation,
(WSN), which can be used to monitor large areas the tracking will be energy-efficient. Because such
effectively. Sensor nodes in a WSN constitute a tracking operation do not require a central server
wireless ad-hoc network, with one or a few sink interaction in tracking. On the other hand, in order to
nodes as the collection point(s) and bridge(s) to the track (or capture/reach the target) timely in a
central server (called the sink). Every node in the surveillance application, the systems should demand
network may create data periodically, on demand of a low tracking time (or capturing time). Regarding
the sink, or triggered by events of interest. At the these in prior work, argue that there still remain
same time, every node may forward data that it concerns for energy-constrained WSNs to be
receives toward sink nodes, which are often multiple addressed.
hops away. WSNs are increasingly being envisioned
for collecting data, for example, physical or

2064
International Journal of Pure and Applied Mathematics Special Issue

2.2 COMMUNICATION CONCERNS graph planarization, the WSN is organized into non-
The network area is usually divided in regions, cells, overlapping areas, which is usually carried out in
grid, clusters, trees, etc., to track the target in a localized geographic routing (particularly, in face
distributed manner. Such division obviously incurs routing).Each face, comprising a number of nodes,
extra frequency of interactions and energy corresponds to a local area of the WSN.
consumption. Sensor nodes near the sinks consume The tracker intends to follow a target, it
more energy and die first such that the tracking task queries the WSN. The nodes in the WSN are
may become critical and slow. periodically clock synchronized to be in an awake,
active, or inactive state. Each node has the capability
2.3 LOCALIZATION CONCERNS of sensing, computing, and communicating. When a
The localization suffers from secular biases node of a face receives a query request, it checks with
due to effects of shadowing or multi-path its neighboring nodes whether or not it is the closest
propagation, radio occlusions, and decalibration, as to the target; if it is, it is elected as a monitor and one
well as large unbiased errors due to measurement of its neighbors is elected as a backup.
noise. The inaccuracy in the localization cannot be The monitor then works at the request of
eliminated even with a plenty of observation data. the tracker and sends information about the monitor
itself, the backup, and the target, while the target
2.4 ROBUSTNESS CONCERNS traverses through the face. In the case that the
Target tracking is made more efficient by monitor has any problem due to any reason, the
exploiting a mobile sink which can always be around backup takes the role of the monitor. Target detection
the detecting sensors or a little distance away (single and localization is mainly performed by the
to multi hop). The objectives are to reduce the cooperation between the monitor and the backup.
capturing time (a new metric that measures the total The tracker then moves toward the monitor
tracking time required to get around of a target within and queries for an update. If the target is still within
a certain distance), to enhance the energy efficiency the face, the monitor keeps tracking the target; at the
of the WSN, and to ensure the QoT. same time, the monitor elects the next possible
monitor and backup to be the new monitor and
2.5 t-TRACKING backup by using prediction method. If the target has
t-Tracking an entity, e.g., a respective already moved out of the area of the face, the monitor
authority that intends to follow a target, is called a informs the tracker about the new monitor and
tracker, which can also be called a mobile sink backup, and the tracker moves toward them.
because it traverses through the network. A tracker is Monitor and backup are two common
assumed to be a single generic source such as a sensors of the current face and one of its adjacent
mobile user or a respective authority. A target can be faces. When the monitor finishes its task, it changes
any mobile entity such as an enemy vehicle or an its state to the inactive state. This is also true for the
intruder. Thus, two mobile nodes, “Target” and backup. In this way, a special linked list of monitors,
“Tracker”, are implemented. A WSN composed of a backups, and other nodes in a face is formed as time
set of static sensor nodes is deployed in a plane, goes on. If both the monitor and the backup are
where the target moves in dynamic patterns. Through viewed as one logical node at each time step of the

2065
International Journal of Pure and Applied Mathematics Special Issue

tracking, this special linked list is simply a linear link from tracking. Once the nodes have been selected for
of logical nodes. tracking, it will start tracking immediately when the
3. SYSTEM ARCHITECTURE target enters into its corresponding polygon. Likewise
when the polygon is about to move to the next face
the same procedure will be followed in that face for
tracking target.

Fig 3.2. Energy Consumption

Fig.3.1. Working procedure of t-Tracking

4. PROBLEM STATEMENTS
For tracking the target polygon region (face)
3.1 WORKING PROCEDURE OF t-TRACKING
is designed with the fixed sensor nodes in the
First the network will be analyzed and the
network region. All the possible movement of target
polygon region will be formed without overlapping.
from one region to another will be indicated by the
Then brink detection algorithm will be used to find
Brink Detection. Then the target may move through
the edge in which the coverage area of the connected
any one of the brink (edge) to the next region. The
nodes is mostly combined. Next the queries about
brink through which the target is moving will be
coverage area and energy level, will be sent to all the
connected by two sensor nodes. The problem in the
nodes in the edges selected by the brink detection
existing system is that the target must be in any one
algorithm. After receiving reply from all nodes, the
of the sensor nodes coverage area and both the nodes
nodes with maximum coverage area will be selected
must have the energy to track the target until it moves
for tracking and the remaining nodes will be rejected
to the next region. If the nodes do not have the

2066
International Journal of Pure and Applied Mathematics Special Issue

energy until the target moves to the next region the are in a low-power mode and wake up at a pre
tracking will be stopped or else if the target moves defined period to carry out the sensing for a short
away from the coverage region of both the nodes then time.
the tracking will be stopped. 5.3 Node Selection
The target node element is the mobile
5. IMPLEMENTATION AND RESULT
target that can emit various forms of signals and has
5.1 Network Formation mobility function. In addition to the mobility
A wireless sensor network (WSN) consists function, the tracker element is equipped with a
of a large number of small sensor nodes that are communication radio to interact with the WSN. The
deployed in the area in which a factor is to be communication element of a sensor node interacts
monitored. In wireless sensor network, energy model with neighboring nodes and the tracker. The element
is one of the optional attributes of a node. The energy follows the rules of energy-efficient state transition
model denotes the level of energy in a mobile node. and its duty-cycle. The WSN planarization is carried
The components required for designing energy model out after its deployment. Once the WSN starts
includes initial Energy, txPower, rxPower, and tracking operation, t’s detection element commences.
idlePower.
The “initialEnergy” represents the level of 5.4 Target Tracking
energy the node has at the initial stage of simulation. Network failure occurs when the monitor
“txPower” and “rxPower” denotes the energy relays a message, but does not receive an
consumed for transmitting and receiving the packets. acknowledgment in time. This is a result of losing an
If the node is a sensor, the energy model should acknowledgment. A node may also fail during
include a special component called “sensePower”. It operation because of fault, energy depletion, or some
denotes the energy consumed during the sensing other reasons. Prediction failure occurs if the
operation. Apart from these components, it is prediction of the next location where t is approaching
important to specify the communication range is incorrect, e.g., t may suddenly change its route or
(RXThresh) and sensing range of a node take a U-turn towards the opposite direction, which
(CSThresh_). The sample 18.tcl designs a WSN in results in an inaccurate predicted location. A
which sensor nodes are configured with different connectivity hole or physical obstacle is an area
communication and sensing range. Base Station is where the edges between nodes form a closed face
configured with highest communication range. Data without direct links between nodes that are not
Transmission is established between nodes using adjacent on the perimeter of the face. If t’s detection
UDP agent and CBR traffic. failure occurs, a monitor instantly issues a message to
all face neighbors about changing their state, which
may result in a short delay in tracking.
5.2 Polygon Formation 6. CONCLUSION
The system initialization, including initial Proposed a new tracking scheme in WSNs,
polygon construction in the plane. A node has all of t-Tracking. Its ideas such as (i) fast tracking
the corresponding polygons’ information after the operation through face detection and prediction (ii)
WSN planarization. Initially, all nodes in the WSN reducing the interactions between a mobile tracker

2067
International Journal of Pure and Applied Mathematics Special Issue

and monitor/backup (iii) reducing the dependency on Processing in Sensor Networks (IPSN), pp.
the need of strict localization accuracy in tracking, 404-405.
can all be useful to many surveillance applications. [4] L.M. Kaplan (Jan. 2006), “Global Node
Particularly, those applications that require keeping Selection for Localization in a Distributed
an eye on a mobile entity and pursuing the entity for Sensor Network,” IEEE Trans. Aerospace
different purposes, such as safety, investigation. T- and Electronic Systems, vol. 42, no. 1, pp.
Tracking can track the entity by accomplishing 113-135.
cooperation between the sensor nodes and tracker in [5] R.Karthikeyan and C.Jothi Kumar,
a real-time manner and by requiring very little “Improved Reputation Systemfor Wireless
processing and communication energy requirements Sensor Networks (Wsns)”, International
on the nodes. Since the best node is selected only Journal of Innovations in Scientific and
based on reply from all nodes in a face it may Engineering Research
consume more time. Some nodes may be in tracking (IJISER),vol.1,No.3,pp.191-195,2014.
in such cases those nodes will reply one when the [6] X. Wang et al (Apr. 2012), “Target
target goes out of its coverage area. Due to this the Tracking in Wireless Sensor Networks
target may not be tracked continuously. In future an Based on the Combination of KF and MLE
efficient algorithm may be developed, which may Using Distance Measurements,” IEEE
keep on updating the energy and coverage area of Trans. Mobile Computing, vol. 11,no. 4, pp.
each nodes in the sink node from where the 567-576.
information may be collected when needed. Another [7] Z. Wang et al (2010), “A Novel Mobility
possible aspect is to verify the proposed scheme in Management Scheme for Target Tracking in
different tracking situations. Cluster-Based Sensor Networks,” Proc.
Sixth IEEE Int’l Conf. Distributed
REFERENCES
Computing in Sensor Systems (DCOSS), pp.
[1] Guojun Wang et al (April 2014), “Detecting 172-186.
Movements of a Target Using Face Tracking [8] Y. Wang et al (2011), “Analysis of Event
in Wireless Sensor Networks,“ IEEE Trans. Detection Delay in Wireless Sensor
On Parallel And Distributed Systems, Vol. Networks,” Proc. IEEE INFOCOM, pp.
25, No. 4. 1296-1304.
[2] T. He et al (2006), “Achieving Real-Time [9] W. Zhang and G. Cao (Sept. 2004),
Target Tracking Using Wireless Sensor “Dynamic Convoy Tree-Based
Networks,” Proc. 12th IEEE Real-Time and Collaboration for Target Tracking in Sensor
Embedded Technology and Applications Networks,” IEEE Trans. Wireless Comm.,
Symp. (RTAS), pp. 37-48. vol. 3, no. 5.
[3] O. Kaltiokallio et al (2010), “Distributed [10] Y. Zhou et al (Feb. 2010), “Posterior
RSSI Processing for Intrusion Detection in
Cramer-Rao Lower Bounds for Target
Indoor Environments,” Proc, Ninth
Tracking in Sensor Networks with Quantized
ACM/IEEE Int’l Conf. Information
Range-Only Measurements,” IEEE Signal

2068
International Journal of Pure and Applied Mathematics Special Issue

Processing Letters,vol. 17, no. 2, pp. 377- 388.

2069
2070