Content uploaded by Khaled Elleithy
Author content
All content in this area was uploaded by Khaled Elleithy on Aug 31, 2014
Content may be subject to copyright.
An Energy Efficient Algorithm for Object-Tracking
Wireless Sensor Networks
Majed Elbishti and Khaled Elleithy
Computer Science and Engineering Department
University of Bridgeport, CT, USA
melbisht@my.bridgeport.edu, elleithy@bridgeport.edu
Laiali Almazaydeh
Department of Software Engineering
Al-Hussein Bin Talal University, Ma’an, Jordan
lalmazay@my.bridgeport.edu
{melbisht@my.bridgeport.edu, elleithy@bridgeport.edu}
Abstract— In this paper we propose and simulate an energy
efficient protocol for object-tracking over Wireless Sensor
Networks. The proposed algorithm aims to use virtual
clustering among the observation region to initiate duty-cycle
across the border nodes and the nodes in the inside region. The
mean idea is to keep the outer nodes active and the interior
nodes sleep for energy saving until an object is detected by the
outer nodes. Castalia 3.2 wireless sensor network simulator is
used to simulate the proposed protocol. Results indicate that an
improvement of 42% is achieved in terms of energy
consumption. Furthermore, it is demonstrated that S-MAC
protocol is a better choice in high load traffic applications in
terms of energy consumption.
I. INTRODUCTION
In 1999, The BusinessWeek expected that the Wireless
Sensor networks (WSNs) to be one of most important
technologies [1]. The applications of wireless sensor
networks are unlimited; it includes all fields where
monitoring of a physical quantity is difficult to be performed
by humans due to location, time or environmental conflicts.
Surveillance systems and continues monitoring, forest fire
detection, and chemical plant observation system are few
examples of such applications[2-4].
This paper aims to propose and simulate an application
specific energy-efficient protocol for object tracking wireless
sensor network systems. In this type of applications it is very
important to tradeoff between power consumption and other
performance metrics because of the life-time requirements of
these systems. The simulation is used for extensive testing
because the nature of the WSN makes it impractical to use
testbeds for hundreds or thousands of nodes [5-7].
In object-tracking sensor networks, the nodes remain
awake to sense any target, although, this is the main goal, but
it is unpractical to keep all the nodes awake from the energy
consumption point of view. Virtual clustering and Regional
clustering are not the best solutions when considering the
unpredicted object entering the field. Trigged wakeup
techniques allowing the nodes to wake up as required, made
it possible to put nodes on sleep for longer time till tracking
event occur. The need to propose object-tracking specific
protocols to benefit from these new techniques has been
established and proposed [8-15].
Energy consumption is very important constraint
because the node is used once and the life-time of the node
must be long enough to support the application. The radio
unit of the sensor consumes most of the energy compared to
the energy consumed by central processing unit or the
sensing device [16]. The idle listening mode in wireless
nodes consumes most of the energy likewise transmission or
receiving modes. Thus, efficient MAC protocol designs are
essential to reduce energy consumption. Scheduled-sleep
active/sleep MAC Protocols were introduced in literature to
reduce the wasted power in idle listening. S-MAC and
T-MAC are the most commonly used to achieve this purpose
[17, 18].
II. PROBLEM IDENTIFICATION
One of the most important applications in WSN is
tracking of mobile objects [19]. Tracking of enemy, animals,
humans and cars in highways are few examples of object
tracking. The energy saving techniques may vary according
the application requirements. For example, the latency
constraint in observing animals is much less compared to
observing enemy targets in battlefield. Thus, there are
application specific techniques and protocols which can be
applied to save energy depending on the nature of the
application [20].
In object-tracking, sensor nodes are deployed in
different regions. The target may not navigate through all of
them. Therefore, there is a need to reduce the energy
consumption in those inactive regions where tracking the
system is unnecessarily active at all times. Some researchers
have proposed Wake up and Sleep mechanics such as
clustering and filtering. Clustering leads sometimes to black
hole issue while filtering involves very complicated
calculations not well suitable to sensor nodes from energy
point of view [21].
This paper proposes an improvement on the target
tracking algorithm proposed by Wang Duoqiang et al. [21].
The original algorithm of Wang Duoqiang et al. proposes an
approach in which all the boundary nodes are active at all
times, and if an intruder is detected, the inside nodes cluster
is activated by using a triggered-wake up technique. This
allows the inter nodes to sleep when no object is available to
track to save energy. The authors assumed that localization
protocol and wake up technique are both available. Although
the authors tried to reduce the energy consumption of the
wireless sensor network by switching on only the nodes
around the field, the technique is not saving any energy for
the boundary nodes. Also there is no contingency plan if any
of the boundary nodes is dying. The proposed changes in this
paper make the algorithm more reliable and make the
life-time of the boundary nodes longer.
III. PROPOSED SOLUTION
In [21], the authors save energy based on the idea of a
sleep schedule for the nodes. All the nodes inside the
monitoring region will remain sleep until they receive a
signal from the boundary/head nodes which will be active all
the time for object detection. When an object enter the region
through boundary nodes, then the major node, which is the
nearest node to the object will send a message to the nodes in
its area of detection to wake up for some period of time then
they go to sleep mode again after the object has left their
sensing area.
We propose an automated sleep and wake-up schedule
for the boundary nodes itself to make their life-time longer,
which can be achieved using T-MAC protocol. Also we
propose an out-of-energy algorithm to be automatically
triggered whenever a boundary node battery is less-than or
equal-to 5% of its full energy level. The algorithm main
function is to replace a dying boundary node with the nearest
inter node.
A. Proposed Algorithm
This paper focuses on an automated sleep schedule for
boundary nodes. T-MAC protocol is used so that the
boundary nodes will automatically start a scheduled sleep
based on their location as boundary node, while the interior
nodes will remain sleep till they receive a wakeup call from
one of the outer nodes in case of object detection as shown in
Figure 1. The red circles indicate the active nodes, while the
green circles are the sleep nodes.
Fig 1 Virtual- Clustering Automated sleep schedule for boundary nodes
We also propose an algorithm that acts as a contingency
plan for any boundary node that is dying. Hence we are
protecting the whole system from being destroyed when
losing boundary nodes. For this purpose an algorithm was
developed to transfer the responsibility of the dying node to
the adjacent node in its sensing area while it passes on the
information to the nodes in that area. The algorithm works as
follows:
If (E ≥ 0.05*Ef AND i is a boundary node)
{
Send WM to Ni nodes
localize Ni , and calculate di
loop (K=1 to k= n , where n is Number of Ni nodes)
If (Jn ≠ a boundary node & di= Min(dk, dk=1, dk=2..dn)
{ send, New ID = ID , Tw, die };
Else { ignore , k=k+1}
}
where, Battery level can be measured
~ Ef is max level
~ Ni is the nodes in sensing area of node (i)
~ di 1 ,2 ,3…K is the distance between node i and it’s
sensing radius neighbors i(n)
The minimum number of nodes in the boundary
coverage is given by [21]:
Nmin= L/R
Where, L is the length of the boundary of the
monitoring region and R is the sensing radius. In this paper
when using the T-MAC model at any instant of time, the
minimum number of active nodes in the boundary will be
approximated:
Nmin= L/2R
This means that the number of active boundary nodes at any
instant of time is 50% less. According to [21], the proportion
of boundary nodes to all nodes is:
P = Nmin/ N= (4L/R)/(pL²) = 4/LpR
According to our proposed improvement this will be:
P=2/(LpR), where p the node density
Based on this mathematical model, we expect to save more
energy in the boundary nodes and make their life-time longer
with 40% more.
IV. SIMULATION
Castalia WSNs simulator has been selected to demonstrate
the performance of the proposed algorithms. 136 nodes were
deployed on field of 10x10 meters according to the
configuration given in Table 1.
TABLE I NODE DEPLOYMENT CONFIGURATIONS TABLE
Node
ID
Deployment type
Code used
0 to 99
Gird
SN.deployment =
“[0..99]->10x10”
100 to
135
Manually located
around the field
SN.deployment =
“[100..135]->”
SN.node[ID].xCoor = X
SN.node[ID].yCoor = Y
The configuration in Table 1 creates a deployment that is
shown in Figure 2. The outer nodes are manually allocated
by using Cartesian coordinates to simulate the assumption
that automatic allocation algorithm is implemented in the
nodes. By grouping the nodes according to their ID to two
groups; Outer and Interior nodes; a virtual clustering is
created within these two groups. The main goal of this
simulation is to compare the power consumption when no
energy saving protocol is used and the case when an
approximate configuration to the proposed protocol is used.
Fig 2 Node Deployment Setup
In the first scenario we simulate the case when no energy
saving protocol is used. This means that all the nodes in the
field are active all the time. The assumptions used in this
case are:
- All nodes are active.
- No energy saving MAC is used.
- CC2420 Radio parameters will be used.
- Throughput Test application is used.
In the second scenario we simulate the case using our
proposed protocol. The assumptions used in this case are:
- The interior nodes are sleep until an object is detected
after sometime then they start to be active.
-Throughput Test is used.
- The outer nodes are active at the beginning of the
simulation.
- TMAC protocol is enabled for the outer nodes to
simulate the duty cycle sleep schedule within the outer
nodes.
- CC2420 Radio parameters are used.
A. Simulation Results
Figure 3 shows the improvement in energy
consumption using OTP (Object-Tracking Proposed
Protocol) compared with no energy saving protocol.
Figure 4 shows the effect of using different MAC
protocols on energy consumption, with high traffic
application throughput test. Figure 5 shows the effect on
energy consumption of using different power transition
levels for each MAC protocol used in conjunction with the
proposed protocol.
Figure 6 shows that TMAC has better performance over
SMAC for 1000 maximum MAC layer packet size, while
keeping the payload at 2000. Also, it demonstrates that
changing the payload and the packet rate has significant
effect on the type of MAC used in terms of energy saving.
Fig 3 Power Consumption VS Time with and without using the proposed
protocol
Fig 4 Energy Consumption VS Time with use of different MAC Protocols
Fig 5 Energy consumption VS TX Power for of different MAC protocols,
2500 Max packet size
Fig 6: Energy consumption VS TX Power for of different MAC protocols,
1000 Max packet size
Figure 7 shows the effect of using different payload on
energy consumption for each MAC. Figure 8 shows that
there almost no effect on energy consumption while
changing the data packet rate in the application we used,
because all the nodes are sending to the sink node in this
application. Although there is low packet rate for each node,
the overall channel will be busy.
Fig 7 Energy consumption VS payload for each MAC used with OTP.
Fig 8 Energy Consumption vs Data Packet Rate per second, with 1000
payload, 2000 Max MAC Packet size
V. CONCLUSION
This paper proposes an application specific protocol to
be used for object-tracking systems. The protocol is based on
virtual clustering by dividing the nodes into two groups;
interior and boundary nodes. The results demonstrate an
improvement of performance in terms of energy
consumption. The simulation establishes the following
results:
- the proposed protocol improves energy consumption,
- SMAC has better performance over TMAC,
- changing the payload size affects only TMAC’s
performance and it is better to use TMAC with larger
payload, and
- changing the data rate has no effect on the energy
performance in case of using an object tracking
application.
These results emphasize that an application oriented
energy efficient protocol is better solution for the energy
problem in object-tracking systems.
REFERENCES
[1] D. Orfanus, J. Lessmann, P. Janacik and L. &Lachev, "Performance of
wireless network simulators: a case study," in 3nd ACM workshop on
Performance monitoring and measurement of heterogeneous wireless
and wired networks, Vancouver, British Columbia, Canada, 2008.
[2] P. L.-H. S. Roy, "Low-Power Wake-Up Radio for Wireless Sensor
Networks.," Mob. Netw. Appl, pp. 15(2), 226-236, 2010.
[3] J.-M. Bohli, A. Hessler, O. Ugus and D. &Westhoff, "A secure and
resilient WSN roadside architecture for intelligent transport systems,"
in the first ACM conference on Wireless network security, Alexandria,
VA, USA., 2008.
[4] U. M. Colesanti, C. Crociani and A. &Vitaletti, "On the accuracy of
omnet++ in the wireless sensornetworks domain: simulation vs.
testbed," in the 4th ACM workshop on Performance evaluation of
wireless ad hoc, sensor,and ubiquitous networks, Chania, Crete Island,
Greece, 2007.
[5] J. Glaser, D. Weber, S. A. Madani and S. Mahlknecht, "Power Aware
Simulation Framework for Wireless Sensor Networks and Nodes,"
EURASIP Journal on Embedded Systems, vol. doi: 10.1155, no.
2008/369178, pp. 1-16, 2008.
[6] T. Olivares, P. J. Tirado and L. & Orozco-Barbosa, "Simulation of
power-aware wireless sensor network architectures," in the ACM
international workshop on Performance monitoring, measurement, and
evaluation of heterogeneous wireless and wired networks,
Terromolinos, Spain., 2006.
[7] P. Le-Huy and M. & Sébastien Roy, "Low-Power 2.4 GHz Wake-Up
Radio for Wireless Sensor Networks," in IEEE International
Conference on Wireless & Mobile Computing, Networking & Comm,
2008.
[8] N. Xu, S. Rangwala, K. K. Chintalapudi, D. Ganesan, A. Broad, R.
Govindan and D. &Estrin, "A wireless sensor network for structural
monitoring," in the 2nd international conference on embedded
networked sensor systems, Baltimore, MD, USA., 2004.
[9] R. P and P. M. & P, "Wireless Sensor Network for Continuous
Monitoring a Patient's Physiological Conditions Using ZigBee,"
Computer & Information Science, vol. doi: 10.5539/cis.v4n5p104, pp.
104-110, 2011.
[10] W. Hu, N. Bulusu, C. T. Chou, S. Jha, A. Taylor and V. N. & Tran,
"Design and evaluation of a hybrid sensor network for cane toad
monitoring," ACM Trans. Sen. Netw, vol. 5, no. 1, pp. 1-28, 2009.
[11] Q. I. Ali, A. Abdulmaowjod and H. M. & Mohammed, "Simulation &
Performance Study of Wireless Sensor Network (WSN) Using
MATLAB," Iraqi Journal for Electrical & Electronic Engineering, vol.
7, no. 1, pp. 112-119, 2011.
[12] Z. Rasin and M. R. & Abdullah, "Water Quality Monitoring System
Using Zigbee Based Wireless Sensor Network," International Journal
of Engineering & Technology, vol. 9, no. 10, pp. 24-28, 2009.
[13] Z. Jingcheng, H. Allan, Y. Qinzhong and G. &Shaofang, "Design of
the Remote Climate Control System for Cultural Buildings Utilizing
ZigBee Technology," Sensors & Transducers, vol. 118, no. 7, pp. 13-27,
2010.
[14] E. E. Egbogah and A. O. &Fapojuwo, "A Survey of System
Architecture Requirements for Health Care-Based Wireless Sensor
Networks," Sensors (14248220), vol. 11, no. 5, pp. 4875-4898, 2011.
[15] I. Akyidiz and M. & Vuran, in Wirless Sensor Networks, John Wiley &
Son, Ltd, 2010, p. 44.
[16] W. Ye, J. Heidemann and D. & Estrin, "An Energy-Efficient MAC
Protocol for Wireless," 2010.
[17] T. Dam and K. & Langendoen, "An Adaptive Energy-Efficient MAC
Protocol," in ACM, 2003.
[18] M. Naderan, M. Dehghan and H. & Pedram, "Mobile object tracking
techniques in wireless sensor networks," in Ultra-Modern
Telecommunications & Workshops, 2009. ICUMT '09. International
Conference , St. Petersburg, 2009.
[19] U. R. a. A. S. Petcharat Suriyachai, IEEE Communications Tutorials,
vol. 14, no. 2, pp. 240-262, SECOND QUARTER 2012.
[20] W. Duoqiang, L. Mingke and Q. Qi, "An Energy-efficient Target
Tracking Algorithm in Wireless Sensor Networks," in Wireless
Communications, Networking and Mobile Computing (WiCOM),
2011 7th International Conference on, Wuhan, 2011.
[21] B. L. Titzer, D. K. Lee and J. &Palsberg, "Avrora: scalable sensor
network simulation with precise timing," in the 4th international
symposium on Information processing in sensor networks, Los
Angeles, California, 2005.
