New Clustering Technique in a Sensor Network for a Connection Durability Improvement

Abstract

The LEACH technique to be representative in much routing technique that existing, was suggested in sensor network environment is a study method of a clustering base. This is suggested technique to manage the sensor network group which consisted of many sensors in efficiency. However, a destroyed problem of data transmission of a message cluster group in the way can occur by the fault that was not able to consider quantity of energy leftover for a cluster header to manage a group of a cluster unit then. We are going to propose to solve such a problem by this paper for new clustering technique to accomplish management for a cluster group. An experiment result shared that prolonging it did not occur at speed to handle connection intentionality of the cluster group which tried to use the technique that suggested by an article of a book as well as improving.

Full text

Page 1

New clustering technique in a sensor network for a connection durability
improvement∗ ∗


Dae Hyun Kim, Young Han Lee, Jin Mook Kim, Kyung Oh Lee
Dept. of Information and Computer Science, Sunmoon University, Korea
E-mail : perseus0080@hanmail.net, hans0209@naver.com,
{calf0425, leeko}@sunmoon.ac.kr



∗ This research was supported by MIC(Ministry of Information and Communication), Korea, under the
ITRC(Information Technology Research Center) support program supervised by the IITA(Institute of Information
Technology Advancement).
Abstract

The LEACH technique to be representative in much
routing technique that existing, was suggested in
sensor network environment is a study method of a
clustering base. This is suggested technique to manage
the sensor network group which consisted of many
sensors in efficiency. However, a destroyed problem of
data transmission of a message cluster group in the
way can occur by the fault that was not able to
consider quantity of energy leftover for a cluster
header to manage a group of a cluster unit then. We
are going to propose to solve such a problem by this
paper for new clustering technique to accomplish
management for a cluster group. An experiment result
shared that prolonging it did not occur at speed to
handle connection intentionality of the cluster group
which tried to use the technique that suggested by an
article of a book as well as improving.

1. Introduction

While much interest is collected recently for a
sensor network, it is worse studies environmental
monitor, weather measurement, lively militarily in
many fields having a specification purpose fields.
Sensor nodes send the information that collected
persistently in sink node and accomplishes a purpose
of an applicable field here. The sensor nodes have
limited energy that was an influence on life and
performance of a general network and uses such an
energy in how much efficient. After all a consideration
matter such as range of a network group-related, load
is herewith a demand matter to be important in a sensor
network.
There is LEACH[1] which used cluster ring
technique in known technique in various technique
most widely of a sensor network. However, LEACH is
not energy efficient because I elect a cluster header
with a probability base without considering head
election time energy leftover quantity. When I am the
worst, the phenomenon that a cluster group by the
energy drying up of data transmission of a message
cluster header in the way is destroyed can occur by
this.
We are going to examine the propriety for the study
that suggest the new cluster ring technique that may be
able to try to elect one candidate node and two head
node and maintain a connection of a network in energy
efficient and watched through simulation for this to
make up for a fault of such LEACH by an article of a
book here.

2. Related Work

2.1. Cluster-Based Routing Protocol

There is the good point that a routing protocol of a
cluster base decreases energy consumption by
repetition transmission of similar information of node
which contacted with, and can raise energy efficiency.
In other words, It forms cluster for sensors, and can
lack in similar event information in cluster head
through data unification process, and it is possible for
routing which is energy efficient to depend by starting
it. It can express constitution of a routing protocol of a
cluster base with Figure 1.
2008 International Conference on Information Science and Security2008 International Conference on Information Science and Security
0-7695-3080-X 2008
U.S. Government Work Not Protected by U.S. Copyright
DOI 10.1109/ICISS.2008.27DOI 10.1109/ICISS.2008.27
1811810-7695-3080-X 2008
U.S. Government Work Not Protected by U.S. Copyright

Page 2

Figure 1. Cluster-Based Routing Protocol
It was suggested the purpose that was going to grow
vitality of a network by because filtering did the event
data that it was repeated by a protocol of a cluster base
tied up node with a cluster unit with Figure 1. and
elected header node of this middle one and stayed, and
managing node.
There is LEACH in a representative protocol in a
protocol of such a cluster base. The LEACH chooses
cluster head stochastically and gives node of all
clusters inner wears the opportunity that can become
cluster head. The LEACH consists of a time unit such
as round, and each stage consists of a stage to
constitute a stage and the TDMA schedule that cluster
consists of it. It chooses cluster head at the time of
cluster constitution and if head is chosen, it tell node
through an advertisement message and it make a
TDMA schedule and tell cluster constitution node. The
constitution node sends data according to a decided
schedule afterward and stays and changes it and
reduces energy consumption by a sleep mode in time.
However, because it choose head without considering
quantity of energy of node, it is energy efficient, and it
cannot be possible and, in addition, it appear by the
energy drying up of head and a network connection is
non-normal and holds the danger that can be over.
A protocol of a cluster base is LEACH-C[2] and
HEED [3] besides LEACH It is a method it is different
from LEACH-C in HEED, and each node calculates a
oneself critical value in the technique that I improved
BS calculates quantity of leftover energy of node, and
to elect that LEACH-C elects it for time probability of
the head election in LEACH, and to elect head by
communication expense.

2.2. NS2 : Network Simulator

This paper used NS2 simulator for an experiment.
At 1988 a research in base and the simulation tool
which calls and it developed the network simulator
which is REAL from UC and Berkeley at 1989 the
network research group of LBNL (Lawrence Berkeley
National Laboratory) [7] this REAL [8] it started ns-1
(network simulator) as after and this it improved ns-2 it
developed it developed.
The ns is TCP, routing protocol, multicast protocol ,
RTP(Real Time Protocol), SRM(Scalable Reliable
Multicast) It is the network simulation tool which it
offers appropriate much environment to accomplish
simulation for the Internet protocol that tool is various,
and now it is wide, and is used.

3. Proposed Algorithm

3.1. Model

This paper suggests to design LEACH there is you
as a base, and to make up for a weak point of LEACH.
The LEACH the problem point which is the
possibility without of being able is caused by with
head elections which are probabilistic and energy
efficiency intelligence it holds. In addition, an
impossible problem of transmission of the data that cut
a network connection when wrong head is elected, and
there is reliability occurs.
For example, if leftover energy of time node that
one node was elected in head is the state that does not
stay very much, cluster which the head belonged to
gets impossible to send data to BS directly. As for the
reason, head is early death. The head will not be able to
transmit information which is collected from the case
cluster where death becomes with BS. Also even until
information which are useful is collected from within
cluster it will not be able to uses. So the node capacity
head with base and with the energy quantity which it
has it gets the location information of the respectively
node fields and it elects character it does.
The BS calculates the position information that
leftover energy of node and each node have and elects
head. The node which BS sends Query to each node,
and caught Query sends each energy and data for
position information to BS with an answer message.
And the BS chooses candidate node of head that one
data collect these data as a base and routing head of
one and one. When 3 nodes is elected like this, node
where it is not elected this head node field and
candidate node who listen it forms cluster in the center
and the Figure 2.(a) and like cluster undergarments
head and candidate works each other communication
as the box confirms a respectively existence.

Figure 2. proposed algorithm
If confirmation work is over between two head node
and candidate node, data collection head makes a
TDMA schedule and conveys it in all node. This time
routing head and it draws up the schedule of separate
182182

Page 3

way even in candidate and node it transmits. These
three node do and the energy remainder quantity
according to the schedule which it comes to decide
communication they check.
If schedule making is over, data movement right
occurs. When the movement of data occurs, the data
acquisition head the Figure 2.(b) from and like receives
composition node capacity sensing information of
circumference and it integrates it compresses the role
which it accomplishes a data. When the Figure 2.(c)
from like it transmits the data which, are integrated like
this with routing head routing head from data
acquisition head it transmits the data which are
transmitted in BS.
The candidate node maintains a sleep state in usual
times. As spare node data collection head node energy
are consumed below critical and they cannot and in
order to substitute this it exists a role accomplishment.
Neither consequently to normal any role accomplishes
not to be, the collection node wants a shift, nor, from
collection node schedule time even to passing, with
data acquisition node it shifts a role in case there is not
answer back.
When trying to lead, system to be stabilized when
with existing LEACH comparing, minimum 3 times
about high efficiency and it is forecast the contents
above with the fact that it will have lifetime.

3.2. Cluster Head Election

It elects head and in order to shift a specific critical
value it stands it uses. This time the critical value
which it uses calculates and lower part (1) with (2) the
price which it comes to get from two it compares and T
price is big and it elects with heads.

T = Eresi /EINIT * CHpnt


T` = ∑Eresi/∑EINIT * CHpnt

Eresi is quantity of early stage of energy that node
had EINIT with quantity of leftover energy of node. Also
CHpnt shows the ratio of data acquisition head of the
entire network. It calculated the ratio of most suitable
cluster head at 5% in a network.[2]
(1) in compliance with the critical value which is
calculated (2) it leads and critical value of the entire
network which is calculated and it compares and (1)
the price is small and like Figure 3. and it transfers
head roles in candidate node.
(1)
(2)

Figure 3. turn over cluster head
It convey that it was it in head that the data that
oneself is new collect authority of data collection head
in node that candidate committee acquisition by
purchase figure node constitutes a cluster inner wear
and accomplish a duty of head.

4. Result

We used NS2 Simulator for performance evaluation
of suggested algorithm and compared network life with
LEACH, LEACH-C, HEED.
The simulation environment set the early stage of
number of node in 100 in an area of 100 * 100 sizes
and increased the number of node by 100 units and
tested it. BS posted at a position (0, 0). In addition, the
initial energy price that each node had set it in 2J.
It calculated the energy consumption model
occurred at the time of the data transmission of a
message between node and node in a lower expression
suggested in LEACH.[1]


ETx(k, d) = Eelec * k + ∊ amp * k * d2

ERx(k) = Eelec * k

It shows (3) a energy dissipation when transmitting
which occurs, it shows (4) a energy dissipation when
receiving a data.
Eelec the data transmitting and receiving at the time
of as the exhaustive energy which occurs has 50nJ/bit
and ∊ amp amplifies a transmission signal as the energy
for has 100pJ/bit/m2. k is data size and d is a node
between distance.
Figure 4 is the graph which compares the survival
rate of nodes from respectively protocol.
(3)
(4)

Figure 4. node survival rate
When protocols of existing and comparing, above
about 30% there is a possibility of knowing the fact
that it is visible node survival rates which improve.
183183

Page 4

This leads and cluster head loads candidate the
dispersive summer solstice node when head energy are
depleted, from the first phase it does not enforce not to
be, the cluster which will grow sprouting and from
within the head and candidate node role shift with the
box it is in order to reduce a energy dissipation.

Figure 5. energy dissipation

Figure 5. is the graph which compared quantity of
energy consumption of node by unit time. The
algorithm to suggest in energy consumption of node
understands a thing with a little quantity of
consumption in comparison with existing algorithms.

Table 1 FND, LND compare
100

FND
(round) (round) (round)
LEACH 708 1387 1634 3209 2390 4608
LEACH-C 872 1698 1803 4145 2601 5106
HEED 905 1806 1912 4707 2809 5379
Proposed 1150 3059 2055 6098 3101 8956

It is table which table pours FND in each protocol
and LND into the number of node, and it compared.
About 10-30% comparison with about 20-50%, HEED
shows that it improved in comparison with about 30-
60%, LEACH-C in comparison with LEACH. The
above-mentioned experiment tends and is for energy
efficiency and network lifetime maintenance than the
algorithm that existing, was suggested and we are
stable all the time and show the performance that
improved.

5. Conclusion

What is represented in the algorithm that used
cluster ring technique in technique of a sensor network
is LEACH. The problem that a connection with cluster
of pertinence can cut if energy of head dries up
because I elect it LEACH consists of a cluster base and
prevents repetition of data, and to do random in energy
efficient and the head election occurs.
It uses two from head present papers and also it
disperses the load which is caught to head this and
candidate node one compared to as with difference, the
200
LND
(round)
300
LND
(round)
LND FND FND
(round)
fact that the connection of the network which is whole
disappears on center the algorithm it proposes. These
three node it is selected the energy which it listens and
a respectively location
compositions stable and consequently is a possibility to
maintain the connection of the entire network well.
The hereafter subject will be made to have and data
acquisition head multiple candidate node in about
candidate node and it sees a data maintenance expense
it minimizes with it will have it researches about in
case and performance appraisal it tries to lead to
propose the technique which is improved.

Reference

[1] W.R.Heinzelman, A.Chandrakasan, and H.Balakrishnan,
"Energy-efficient communication protocol for wireless
microsensor networks," in Proceedings of HICSS-33, Jan
2000, pp. 3005-3014.
[2] W.R.Heinzelman, A.Chandrakasan, and H.Balakrishnan,
"An application-specific protocol architecture for wireless
microsensor networks," IEEE Transactions on Wireless
Communications, Oct 2002, pp. 660.670.
[3] O.Younis and S.Fahmy, "Distributed clustering in adhoc
sensor networks: A hybrid, energy-efficient approach," in
Proceedings of IEEE INFOCOM, Mar 2004.
[4] http://nile.wpi.edu/NS/
[5] Jamal N. Al-Karaki, Ahmed E. Kamal, "Routing
Techniques in Wireless Sensor Networks : A survey," IEEE
Wireless Commun. Vol. 11, Issue 6, Dec 2004, pp. 6-28.
[6]http://www.isi.edu/nsnam/ns/
[7]http://www-nrg.ee.lbl.gov/ns/
[8] Srinivasan Keshav, REAL : A NetworkSimulator, Dec
1988.
[9] Clare, Pottie, and Agre. "Self-Organizing Distributed
Sensor Networks," In SPIE Conference on Unattended
Ground Sensor Technologies and Applications, Apr 1999, pp.
229–237.
[10] K. Scott and N. Bambos, "Routing and Channel
Assignment for Low Power Transmission in PCS," In 5th
IEEE Int. Conf. on Universal Personal Communications,
volume 2, Sep 1996, pp. 498–502.
[11] S. Basagni, “Distributed Clustering Algorithm for Ad-
hoc Networks,” in International Symposium on Parallel
Architectures, Algorithms, and Networks (I-SPAN), 1999.
[12] W. Ye, J. Heidenmann, and D. Estrin, “An Energy-
Efficient MAC Protocol for Wireless Sensor Networks,” in
Proceedings of IEEE INFOCOM, New York, NY, Jun 2002.
in base and cluster
184184