Content uploaded by Khaled Elleithy
Author content
All content in this area was uploaded by Khaled Elleithy on Jan 19, 2014
Content may be subject to copyright.
Content uploaded by Khaled Elleithy
Author content
All content in this area was uploaded by Khaled Elleithy
Content may be subject to copyright.
URI - NE ASEE 2007 Conference
1
An Efficient Approach to Reduce the Energy Consumption in Wireless Sensor
Networks through Active Nodes Optimization
Khushboo Patel
1
, Chaitali Patel, Syed S. Rizvi, K. M. Elleithy
Computer Science and Engineering Department
University of Bridgeport
Bridgeport, CT 06605
{khushbop, chaitali, srizvi, elleithy}@bridgeport.edu
Abstract
A sensor network is made up of numerous small independent sensor nodes with
sensing, processing and communicating capabilities. The sensor nodes have
limited battery and a minimal amount of on-board computing power. This paper
presents a novel methodology that utilizes source and path redundancy technique
to effectively reduce the required energy consumption while at the same time
maximize the lifetime of the sensor networks. In addition, the proposed
methodology presents a strategy to optimize the number of active sensor nodes
and assign equal time slots to each sensor nodes for sensing and communicating
with the Base Station (BS). Simulation results demonstrate that the proposed
methodology significantly minimizes the energy consumption and consequently
increases the life time of the sensor nodes.
Keywords: Base station, energy consumption, sensor node.
I. Introduction
A wireless sensor network (WSN) is made of numerous small independent sensor nodes to
monitor environment at different locations. The sensor nodes, typically the size of 35mm film
canister, are self-contained units consisting of a battery, radio, sensors, and a minimal amount of
on-board computing power. Nodes must have self configuration and adaptation mechanisms to
support fault tolerance. In the past few years, the rapid development in miniaturization, low
power wireless communication, micro-sensor and small-scale energy supplies have given WSNs
a new technological vision. WSNs show great potential for increasing the information available
to people in a wide variety of consumer and industrial applications (e.g., smart buildings,
position sensing, target tracking, interactive museums, managing inventory control, and home-
automation). While a lot of research has been done on some important aspects of WSNs such as
architecture and protocol design, energy conservation, and localization, supporting Quality of
Service (QoS) in WSNs is still a largely unexplored research field. This is mainly because WSNs
are very different from traditional networks.
1
Contact author: khushbop@bridgeport.edu
URI - NE ASEE 2007 Conference
2
Communicating one bit over a wireless medium at short ranges consumes more energy than
processing that bit. With the current technology, the energy consumption for communication is
several magnitudes higher than the energy required for computation; and wireless
communication is foreseen to continue to dominate energy consumption in the near future. There
are two possible ways to decrease the energy used for communication in a sensor network:
minimize the amount of the transmitted data or decrease the communication range. The
transmission energy is proportional to d^α, where d represents the transmission distance and α
represents the attenuation exponent. Therefore, minimizing the amount and the range of
communication can significantly prolong the life of a sensor network.
In this paper, we propose a novel methodology that utilizes the source and path redundancy
technique to effectively reduce the required energy consumption while at the same time
maximize the lifetime of the sensor networks. In [2], the concept of moving the BS to its
optimum position was introduced. In addition to position optimization, the proposed algorithm
optimizes the number of active sensor nodes with respect to the location of the BS. Simulation
results demonstrate that the proposed approach significantly reduces the energy consumption
while at the same time provide a simple and efficient architecture for sensing active nodes.
Current interest of WSNs includes optimizing the performance of sensor networks for distributed
sensing applications [3]. It is well known that QoS is an overused term with various meanings
and perspectives. Different technical communities may perceive and interpret QoS in different
ways. In general, QoS is mainly concerned with the satisfaction of the users and measures the
service requirements that must be met when transporting a packet stream from the source to the
destination. In this scenario, this refers to an assurance by the internet to provide a set of
measurable services attributed to the end-to-end users in terms of jitter, available bandwidth, and
packet loss. In general, accuracy, precision, energy/power, lifetime, cost, etc. has to be taken care
of by the network in order to provide better QoS. From the network perspective, the goal is to
provide QoS services while maximize the network resource utilization. In order to achieve this
goal, the network is required to analyze the application requirements and deploy various network
mechanisms.
The sensor networks can be categorized by the periodicity of data transmission. In a time-driven
network, every node sends messages periodically, while in an event-driven, a node sends
message only when sensing a phenomenon. The third category is the query-driven approach
when the sensors transmit data only after receiving a query from the BS. In query-based sensor
systems, a user may issue a query with QoS requirements in terms of reliability and timeline and
expect a response to be returned within the deadline. This paper investigates the problem of
optimizing sensor network that has sensor nodes whose active lifetime is significantly shorter
than the required lifetime of the sensor network. In order to satisfy these QoS requirements, we
use the fault tolerance mechanisms through source and path redundancy technique which may
cause the energy of the system to be quickly depleted. We show that the utilization of source and
path redundancy technique can effectively reduce the required energy consumption while at the
same time maximize the lifetime of the sensor nodes.
The rest of the paper is organized as follows: Section II provides the state of the art research that
has done in this area. Section III presents a discussion on the proposed approach for optimizing
the number of active sensor nodes. Section IV presents the simulation results that demonstrate
the success of the proposed approach. Finally, we conclude the paper in Section V.
URI - NE ASEE 2007 Conference
3
II. Related Work
WSNs usually contains thousands or millions of sensors, which are randomly and densely
deployed (typically 10 to 20 sensors per m^2) [1]. Due to high number of node count, it is not
only impossible to keep track of each node but also not feasible to replace each node in case of
their failure. Therefore, the number of nodes more than the necessary amount to cover the area is
deployed to cope with the failure problem. This problem of WSNs is referred as redundancy [1].
This arouses the need for the sensor nodes with greater active lifetime to prevent the failure
problem. The main goal is to prolong the lifetime of the network, which can be defined in several
ways: (1) the time when the first node depletes its battery, (2) the time till a given percentage of
the sensors have enough energy to operate, and (3) the time till a given percentage of the region
is covered by active sensors [2].
Recently, much research has done in the area of energy saving issues in WSN’s. Many proposals
are put forward to minimize energy consumption in sensor networks [2, 4, 5]. Various power
saving schemes have been proposed not only for the hardware and architectural design but also
for designing the algorithms and protocols at various layers in the network architecture.
In [2], a theory is proposed to save the energy by reducing the range of communication and the
amount of data transmitted where as in [6] a model is proposed in which the sensor nodes are
forwarded to sleep mode whenever the nodes are not sensing the environment. In addition, each
location of the physical environment is kept under the examination with a set of sensor nodes
(with different sensing periods allotted) and rest of the sensor nodes go to the sleep mode. In [4],
a particular location in the environment is sensed by only one sensor node at a time. But as a
single sensor node is prone to failure, there may be errors in the sensing. In addition, a query-
based optimal path and source redundancy approach is proposes in [7]. In order to achieve the
QoS and maximize the lifetime of WSNs, the approach finds an optimal path by utilizing the
source redundancy technique.
The majority of work till now considered sensor networks to be entirely immobile. In [2], a new
concept of moving the BS to a position at which the distance between the active sensor nodes
and BS becomes optimum has been introduced. This paper proposes a novel approach in which
the BS is kept mobile where as a large majority of the sensor nodes are forwarded to sleep mode
by optimizing the number of active sensor nodes. However, for distance optimization, this paper
uses the proposed approach of [2].
III. The Proposed Method for Active Sensor Node Optimization
Before introducing the proposed algorithm, it is worth mentioning some of our key assumptions.
We assume that each node is aware of its location and it is static in nature. The BS, on the other
hand, is unaware of its location and can be moved unlike the nodes. In addition to that, the
sensing area of the nodes is assumed to be a circle of radius r with the center of the circle as the
node itself. Finally, we assume that all the nodes are synchronized with their neighboring nodes
and can communicate with each other. The first step of the proposed method is the calculation of
the optimal distance for placing the BS to its optimum position. We consider that the BS moves
relatively fast to the respective optimal location once the minimum distance is determined. After
URI - NE ASEE 2007 Conference
4
successfully computing the optimal distance, the next step is to use the proposed method that
reduces the number of active nodes to communicate with the BS.
A. Strategy for Moving the Mobile BS
In order to move the BS to the optimum position we consider the same theory as proposed by
[2]. With respect to the proposed strategy, the BS is placed to its optimum position by
determining the distances between all the sensor nodes with BS.
B. The Proposed Method for Optimizing the Number of Active Nodes
The lifetime of a sensor network has two phases: first is the initializing phase and the second is
the sensing phase. In the first phase, the sensor senses their own position and synchronizes
themselves with the neighboring sensors [6]. After the initialization phase, the sensor nodes are
ready to sense the physical environment. The proposed method introduces the strategy for
reducing the number of active nodes which sense the area and communicate with the BS. The
proposed methodology can be understood by Figure 1. As soon as the BS takes its optimum
position, only those sensor nodes which are within the range of 2r from the BS remain active. In
Figure 1, the BS is shown in middle of a grid and surrounded by the sensor nodes.
In Figure 1, only the sensor nodes which are within the range of 2r (Nodes are named as follows
A, B, C, D, E, F and G) from the BS communicate with the BS where as the rest of the nodes go
to the sleep mode. When one particular sensor node goes to sleep mode, its sensing,
communication and computation components can all be asleep and only a timer needs to work
and wake up all components according to its predefined schedule. The BS has the knowledge of
the coordinates of sensor nodes which are within the range of 2r in its memory. The placement of
the BS to the optimum position reduces the distance between the active nodes and the BS which
consequently minimizes the communication distance between them. Thus, the proposed
methodology provides these reductions in both the distance and the communication cost that
become one of the reasons for giving better energy consumption. In other words, the proposed
methodology minimizes the number of active nodes that reduces the total energy consumption
which is currently one of the QoS requirements in WSNs.
The next step of the proposed methodology is the division of the sensing period of the remaining
active nodes. Previous work has the concept of keeping only one sensor node to be active for
Figure 1. Active nodes within the range of 2r of BS
URI - NE ASEE 2007 Conference
5
sensing a particular region in the environment and so only one node consumes energy and hence
the energy is saved [4, 8]. However, one of the common problems with single node activation is
that it is prone to failure very often which may introduce errors or false alarms. In order to
overcome this problem, our proposed methodology uses the concept of a location which is
sensed by more than one node alternately for regular intervals of time. According to our model,
the sensing is always done by the active nodes. The active node sensing occurs in the form of
cycles of predetermine equal time of T
total
. Each cycle (T
total
) is divided into number of time slots
T
slots
with respect to the number of active nodes. This division provides the required time to
perform sensing for one active node. This scenario is shown in figure 2. The value of T
slots
depends on the number of optimized active sensor nodes N. This relationship can be expressed
as: T
slot
= T
total
/ N.
To understand this concept, let us assume that one cycle (T
total
) of the sensing period is of 70
minutes. For our numerical calculation, we kept N to be 7 active nodes which come under 2r
range from the BS. The use of 7 active nodes implies that we need to divide this sensing period
among the 7 sensing nodes which are named as: A, B, C, D, E, F, and G as shown in Figure 3.
Taking this scenario into account, each node has to sense for 10 minutes alternately in one cycle
of sensing period. The orders of the sensing for the nodes are decided by distance from the BS
which is calculated by the proposed methodology. The node which is nearest senses first and so
on. Consequently, each point in the target environment is covered by at least one working active
node that communicating with the BS. Hence 100% sensing coverage is achieved where as the
average lifetime of each sensor nodes is increased as the sensing period gets divided among the
active nodes. This shows that the required energy consumption is reduced by the proposed
methodology and achieved the desired QoS requirement for the WSNs.
IV.Performance Analysis of the Proposed Algorithm
T
slot
-
2
Total
-
T
T
slot
-
1
T
slot
-
N
Time
=
total
Figure 2. Sensing period of the Active nodes
A
B
C
D
E
F
G
Total
-
Time =
T
total
1
Round of
70 min
1
0 min
10 min
1
0 min
Figure 3. Sensing period for 7 Active nodes
URI - NE ASEE 2007 Conference
6
The proposed methodology is modeled and implemented in C language. For the sake of
simulation, we assume that the network is divided into small clusters with typically 450 to 500
nodes within the coverage area of a BS. The algorithm calculates the individual distances from
the various nodes and the corresponding BS. After calculating the minimum energy utilized by
the nodes, it places the BS to the optimum position. Furthermore, the algorithm determines
which are nodes are within the range of the BS. Sensor nodes that found within the 2r radius of
the BS will be kept active whereas all the sensor nodes that are out of range force to move in the
passive state. We assume the value of α to be 2. Note that the value of α may vary typically from
2 to 5. The proposed model can be explained by the flow graph as shown in figure 4. The
Figure 4. Flow graph for the proposed model
URI - NE ASEE 2007 Conference
7
proposed algorithm starts working with the occurrence of any event in the environment. Initially,
all sensor nodes become active and they synchronize themselves with the BS. The algorithm
starts calculating distance between n sensor nodes (where n indicates the number of nodes) and
the BS until it records minimum energy required to communicate with the BS. It should be note
that the energy consumption is a function of distance between the communicating BS and the
sensor nodes. The algorithm intelligently places the BS to the optimum position where the
minimum energy is consumed by the network. Furthermore, the algorithm checks the position of
each sensor node for decision making. If the targeted nodes are not within the range of 2r from
the BS, they are forwarded to sleep mode. As a result, only those sensor nodes are kept active
which are within the range of 2r from the BS.
For the sake of performance analysis of the proposed algorithm, we consider a cluster of network
of 500 nodes with the radius of the BS is assumed to be 35 mm. The energy required to transmit
one unit of data is assumed to be 2.5 KJ. The total energy of a node is assumed to be 300 KJ.
Note that the node dies as soon as all its energy is consumed. Figure 5 shows the comparison of
the energy consumption for a network having a static BS with the network having a mobile BS.
It can be seen in figure 5 that the entire network consumes 3.5 * 10^7 KJ of energy with 100
active nodes when the BS is in the passive state. On the other hand, the total energy consumption
is reduced to 1.4 * 10^7 KJ when the BS is moved to the optimum position. In order to verify the
consistency of the simulation results, we run the simulation for 1000 active nodes instead of 500
active nodes with the radius of 50mm. Figure 6 explains this scenario. This can be seen in Figure
that the energy consumed by the entire network with 200 active nodes is 7.5 * 10^7 KJ when the
BS is static. On the other hand, this energy reduces to 5.21 * 10^7 KJ after moving the BS to its
optimum position. One can, therefore, conclude that the energy consumption is reduced by more
than 45% through our proposed algorithm compared to other well known methods.
It should be noted that as the number of active nodes increases, the difference between the
energy consumption for both the static and the mobile BS increases significantly which can be
clearly evident in figure 5 and 6, respectively.
0 50 100 150 200 250 300 350 400 450 500
0
2
4
6
8
10
12
14
16
18
20
Nodes
Energy Cosumption (KJ)
Static BS
Mobile BS
Figure 5. Nodes vs. energy with no. of active-nodes = 500 and BS-radius = 35 mm
URI - NE ASEE 2007 Conference
8
Figures 7 and 8 demonstrate the simulation results of the lifetime of a particular node when there
is more than one active node exist. According to the proposed methodology, we divide the
sensing period among the optimized active nodes. By dividing the sensing period among the
nodes, the sensing time of one particular node is reduced significantly. This ensures that the
energy required by a particular node to communicate with the BS is reduced and consequently
increases the lifetime of a sensor node. Also, by dividing the sensing period among the optimized
active nodes assures that each location in the environment is sensed by at least one active node.
This ensures 100% coverage of the environment.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
5
10
15
20
25
Nodes
Lifetime
Lifetime
Figure 6. Nodes vs. life time with no. of active-nodes = 5 and BS-radius = 35 mm
0 100 200 300 400 500 600 700 800 900 1000
0
5
10
15
20
25
30
Nodes
Energy Cosumption (KJ)
Static BS
Mobile BS
Figure 7. Nodes vs. energy with no. of active-nodes = 1000 and BS-radius = 50 mm
URI - NE ASEE 2007 Conference
9
Figure 7 shows the lifetime of a sensor node where the radius of the BS is assumed to be 35mm.
On the other hand, figure 8 shows the lifetime of the sensor node with the radius of the BS of
50mm. It is observed that in figure 7, the lifetime of one node is 2.6 when the sensing period is
divided among 2 active nodes in the network with 35mm radius of BS. The lifetime increases
from 2.6 to 2.75 in figure 8 for the same parameters when the radius of BS is increased by
50mm. This is because as the radius of BS increases, the sensor node has to communicate for a
shorter distance because the effective distance between the sensor node and the BS decreases. It
can be verified that as the number of active nodes increases, the time allocated to one active node
for sensing and communicating decreases and hence the lifetime of that node increases. Thus,
this implies that our proposed model not only reduces the overall energy consumption of the
network but also increases the lifetime of the nodes significantly.
V. Conclusion
In this paper, we introduced the concept of optimizing the number of active sensor nodes after
placing the BS to optimum position. Once the optimum position of the BS is determined, we
optimize the number of the active nodes. The simulation results of the proposed algorithm
demonstrate that the energy consumption of the network is reduced by a very large factor. In
addition, our experimental verifications show that the proposed algorithm provides much longer
network life through energy conservation and balancing among sensor nodes. Finally, the
simulation results show that the proposed algorithm gives comparatively a small communication
overhead required to establish a working duty schedule among nodes.
0 1 2 3 4 5 6 7 8 9 10
0
10
20
30
40
50
60
Nodes
Lifetime
Lifetime
Figure 8. Nodes vs. life time with no. of active-nodes = 10 and BS-radius = 50 mm
URI - NE ASEE 2007 Conference
10
VI. References
[1] Hidayet Ozgur Sanli, Hasan Cam, and Xiuzhen Cheng, “An Energy Efficient QoS Protocol
for Wireless Sensor Networks,” Proceedings of SCS WMC 2004,San-Diego, CA.
[2] D. Vass, Z. Vincze, R. Vida, A. Vidacs, “ Energy Efficiency in Wireless Sensor Networks
Using Mobile Base Station,” Proc. of 11
th
Open European Summer Schooland IFIP WG6.6,
WG6.4, WG6.9 Workshop (EUNICE 2005), Colmenarijo, Spain, 6-8 July,2005.
[3] Jim Kay, “Research Proposal: Optimization of Sensor Networks for the Target Tracking
Problem under Power and Reliability Constraints,” November 2003.
[4] J. Mihaela Cardei, T. Thai, Yingshu Li, Weili Wu, “ Energy–Efficient Target Coverage in
Wireless Sensor Networks,” in proceedings of the 24
th
Conference of the IEEE
Communications Society(INFOCOM 2005), March 2005.
[5] C. E. Jones, K. M. Sivalingam, P. Agrawal, and J. C. Chen,“ A Survey of energy efficient
network protocols for Wireless and Mobile Networks,” ACM/Baltzer Wireless Networks,
Vol. 7, Issue 4, pp. 343-358, August 2001.
[6] Ting Yan, Tian He, John A.Stankovic, “Differentiated Surveillance for Sensor Networks,”
Proceedings of the 2
nd
International Conference on Mobile Systems, Applications and
Services, Boston, MA, USA, pp. 270-283 – 2004.
[7] Ahn Phan Speer, Ing-Ray Chen, “An Optimal path and Source Redundancy for Achieving
QoS and Maximizing Lifetime of Query-based Wireless Sensor Networks,” Proceedings of
14
th
IEEE International Symposium on Modelling, Analysis and Simulation, mascots, pp. 51-
60, 2006.
[8] Kaushal Mittal, Anshu Veda, Bhupendra Kumar Meena “Data Aggregation, Query
Processing and Routing in Sensor Networks,” IIT Powai, Mumbai, 2004.
AUTHOR BIOGRAPHIES
KHUSHBOO V. PATEL is a full time M.S. student of Electrical Engineering at University of
Bridgeport. She received B.S. degree in Electronics and Communication from Sardar Patel
University, India in 2006. In the past, she did research work in network security related issues in
wireless sensor networks. Her current research includes the other aspects for increasing the
lifetime of the wireless sensor nodes. Her contact information is: Department of Computer
Science and Engineering, Tech building, Room 127, 121 University Ave, Bridgeport, CT 06604.
Phone: (203) 224-0054. Email: khushbop@bridgeport.edu
CHAITALI K. PATEL is a full time M.S. student of Computer Science at the University of
Bridgeport. She received a B.E. in Information Technology from Sardar Patel University, India
in 2005. Her current research focuses on QoS of wireless sensor networks.
SYED S. RIZVI is a full time Ph.D. student of Computer Science & Engineering at University
of Bridgeport. He received a B.S. in Computer Engineering from Sir Syed University of
Engineering and Technology and an M.S. in Computer Engineering from Old Dominion
University in 2001 and 2005 respectively. In the past, he has done research on bioinformatics
projects where he investigated the use of Linux based cluster search engines for finding the
desired proteins in input and outputs sequences from multiple databases. For last one year, his
URI - NE ASEE 2007 Conference
11
research focused primarily on the modeling and simulation of wide range parallel/distributed
systems and the web based training applications. His current research focuses on the design,
implementation and comparisons of algorithms in the areas of multiuser communications,
multipath signals detection, multi-access interference estimation, computational complexity and
combinatorial optimization of multiuser receivers, peer-to-peer networking, and reconfigurable
coprocessor and FPGA based architectures.
DR. KHALED M. ELLEITHY received the B.Sc. degree in computer science and automatic
control from Alexandria University in 1983, the M.Sc. Degree in computer networks from the
same university in 1986, and the M.Sc. and Ph.D. degrees in computer science from The Center
for Advanced Computer Studies at the University of Southwestern Louisiana in 1988 and 1990,
respectively. From 1983 to 1986 he was with the Computer Science Department, Alexandria
University, Egypt, as a lecturer. From September 1990 to May 1995 he worked as an assistant
professor at the Department of Computer Engineering, King Fahd University of Petroleum and
Minerals, Dhahran, Saudi Arabia. From May 1995 to December 2000 he has worked as an
Associate Professor in the same department. In January 2000 Dr. Elleithy joined the Department
of Computer Science and Engineering in University of Bridgeport as an associate professor. In
May 2003, Dr. Elleithy was promoted to full professor. Dr. Elleithy published more than sixty
research papers in international journals and conferences. He has research interests in the areas
of network security, mobile/wireless communications, computer arithmetic and formal
approaches for design and verification.
