CA-MWSN: Clustering Algorithm for Mobile Wireless Senor Network

International Journal on Applications of Graph Theory in Wireless Ad hoc Networks and Sensor
Networks(GRAPH-HOC) Vol.6, No.4, December 2014
DOI:10.5121/jgraphoc.2014.6402 15
CA-MWSN: CLUSTERING ALGORITHM FOR
MOBILE WIRELESS SENOR NETWORK
Priti Kumari1
1
Department of Computer Science Engineering, Netaji Subash Institute of Technology,
Bihta, Patna, Bihar, India
ABSTRACT
This paper proposes a centralized algorithm for cluster-head-selection in a mobile wireless sensor network.
Before execution of algorithm in each round, Base station runs centralized localization algorithm whereby
sensors update their locations to base station and accordingly Base station performs dynamic clustering.
Afterwards Base station runs CA-MWSN for cluster-head-selection. The proposed algorithm uses three
fuzzy inputs Residual energy, Expected Residual Energy and Mobility to find Chance of nodes to be elected
as Cluster-head. The node with highest Chance is declared as a Cluster-head for that particular cluster.
Dynamic clustering provides uniform and significant distribution of energy in a non-uniform distribution of
sensors. CA-MWSN guarantees completion of the round.
KEYWORDS
Mobile wireless sensor network, fuzzy logic, cluster-head, CA-MWSN, centralized algorithm, dynamic
clustering
1. INTRODUCTION
Sensor networks have a deep impact on our day-to-day life. These are quite promising for the
convenience of human and risk management. Another significant factor for their omnipresence in
our life is ‘Safety’; Safety of human beings or their belongings or animals or climate. Sensors
have been mounted everywhere for the significant task of monitoring. And in such scenario
Mobile wireless sensor network has fascinated and eased our life. In Consequence, Mobile
wireless sensor network is getting distinct hot research areas these days like node localization and
discovery algorithms [1] techniques for energy utilization and conservation, resource allocation
and optimization, distributed deployment strategies, cognitive radio techniques ,dynamic
adjusting strategies, network coding and its applications, network capacity, outage probability,
performance analysis and optimization, MAC/network layer cooperative communications,
security and privacy [2]. Among these research areas, energy conservation is one of the
prominent challenges as sensor lacks the battery recharging or changing options, simultaneously
they are supposed to operate for long network lifetime [3] like habitat monitoring requires months
of monitoring. For supervising long term degradation of civil engineering structures, wireless
sensors are supposed to communicate over several years [4]. In order to maximize the total
battery life of a wireless network, we must minimize the energy consumption of the entire
network. Clustering of sensor nodes, data-driven approaches [5], scheduling sleep intervals for
extraneous nodes [6] are efficient ways of saving energy in sensor network. Earlier experiments
show that Clustering of sensor nodes saves energy by 8 times [7]. Furthermore, cluster-based

16
sensor network is considered here among various wireless sensor network topology like cluster-
based sensor network, sensor network with a fusion center, and concatenated sensor network.
This paper considers hypothesis and the design of a Clustering Algorithm in Mobile Wireless
Sensor Network (CA-MWSN). Base station executes centralized localization algorithm [8] for
performing dynamic clustering then CA-MWSN is run. After successful completion of each
round of cluster-head localization algorithm is run and subsequently CA-MWSN. With the help
of input variables and fuzzy inference engine, a node within a cluster is decided as the cluster-
head. Before designing any clustering algorithm, numerous issues need to be pondered and these
issues significantly differ from the underlying network. Like in homogenous network, the role of
cluster-head certainly is transferred from one sensor node to another sensor node to avoid energy
shortage of an individual node but in case of heterogeneous network one sensor node can perform
the duty of cluster-head more rounds than others. In case of mobile wireless sensor networks,
topology of network is supposed to change, as in our case all the sensor nodes including cluster-
heads are movable. In such scene, cluster-radius and location of cluster-head are thought to be
decided dynamically looking over current positions of sensor nodes. And this gives uniform
distribution of cluster-heads and better coverage either locally and globally. In case of Leach [7],
if randomly generated probability did not lie below threshold, no cluster-head is being chosen.
This vagueness is overcome in CA-MWSN by definite selection of cluster-head in each round
until expected residual energy of all nodes in network go below to ‘low’.
The rest of this paper is organized as follows. Section 2 gives the idea of working of algorithms
previously designed for mobile wireless sensor network. Section 3 tells about pre-conditions
required before executing CA-MWSN algorithm. Section 4 describes the premise of algorithm
CA-MWSN. Section 5 concludes the paper and gives some prospects for future work.
2. RELATED WORK
Evolving needs of human being has speeded up researches in the area of mobile wireless sensor
network. Though techniques and yields grew up but challenges also got multiplied as mobility of
nodes came into existence. In the context of clustering algorithms for mobile wireless sensor
network, some algorithms are advancement of low-energy adaptive clustering hierarchy LEACH
[7], like LEACH-M (Leach-Mobile) [9] and LEACH-ME (Leach-Mobile- Enhanced) [10].
LEACH-M [9] is similar to LEACH [7] except the fact that sensor nodes are mobile and due to
mobility sensors change their position and sometimes their clusters also. Hence before
transmitting data, each sensor node confirms about its cluster-head firstly. And cluster heads
removes mobile nodes from their node list if no data is received in continuous two time slots. In
Leach-ME [Kumar et al. 2008], before election of cluster-head mobility factor of node is seen.
Mobility factor is calculated in terms of remoteness where remoteness is distance between two
nodes. Mobility factor, energy level and transition count are decision factors for selection of
cluster-head. Apart from Leach-based protocols, there is LUDC (Location- Unaware distributed
clustering algorithm for mobile wireless sensor networks using fuzzy logic) [11] in which cluster-
head of a node is selected on the basis of four fuzzy descriptors: energy level, mobility, quality of
link and received signal strength. There is another algorithm ACE-L (Algorithm for Cluster-head
Election with Location) [12] which are suitable for pre-defined mobile wireless sensor network.
For the selection of cluster-head, location of node is considered as it can be anticipated by name.
Cluster-heads are chosen corresponding to some fixed reference points. But in case of gathering
of sensor nodes at one place which is possible as nodes are movable, one cluster-head
corresponding to one reference point may give lower performance.

17
Hitherto, algorithms are not discerning the real problem or property of mobile wireless sensor
network i.e; movement of sensor nodes which should be considered firmly during design of an
algorithm because logistics changes holistically in case of random motion probability. Scenario
change in following respects:
1. If sensors become dense in particular area, more number of cluster-heads is needed than
earlier.
2. If sensors become sparse in particular area, one cluster-head is able to cover larger area
than earlier.
This paper tries to resolve issues created with non-uniform distribution of sensor nodes with
proper energy usage. The idea is same as dynamic cell sectorization [13] in cellular
communication. Most of the calculations are performed at the Base Station which is considered as
source of infinite energy.
3. PRE-CONDITIONS
1. All sensors in the network are homogeneous and energy constrained.
2. Sensors and cluster-heads are movable.
3. Base station is static and placed far from the intended area.
4. Each sensor is equipped with location finding system [14] or network contains anchor nodes.
5. After selection of cluster-head, only cluster-heads communicate with base station.
Before execution of algorithm, it is assumed that all sensors must unicast their location
information to Base Station and on judging the location of all nodes in network; Base Station
divides the network into certain clusters dynamically. For each cluster, a cluster-head is decided.
Dynamic clustering is inevitable due to following causes:
1. Ad-hoc deployment of sensor nodes- Sensors are deployed randomly, so clusters should be
made after knowing the locations of sensor nodes.
2. Mobility of sensors- Sensors are free to move anywhere, as if there is accumulation of nodes
cluster-radius is made sort.
3. Better coverage and energy balance because base station keeps global knowledge of network.
Once network gets divided into clusters, sensors belonging to a particular cluster send
information of their Residual Energy (RE) and Mobility as shown in Figure 1.

18
Fig 1. Round phases of CA-MWSN
On the basis of residual energy, Expected Residual energy (ERE) after next round is calculated. If
ERE falls below Threshold energy, node is discarded from being competitor for cluster-head.
ERE is a parameter for judging if a node can sustain for a round as a c luster-head. The Threshold
Energy is the minimum energy required by a node to run a cluster-head round. Although,
Threshold Energy is variable because it depends on number of nodes forming a cluster and
distance of node from base station, however in the interest of maintaining simplicity of algorithm
it can be considered as fixed amount of energy. Base station calculates Expected Residual Energy
by following formula for round ‘r’:
Expected Residual Energy (r) = Residual Energy (r) – Energy Consumption (r)
......… (1)
Where Energy Consumption is calculated on the basis that if node acts as cluster-head. Energy
Consumption in round ‘r’ depends on the location of sensor and the number of nodes in cluster.
The distance between sensor and base station is followed by free space model [15]. The sensor
node far from base station consumes more energy a s well as node with more number of nodes in
cluster consumes more energy. I n this way, the node with more ERE is stronger candidate for the
selection of cluster head.

19
4. THE PROPOSED ALGORITHM CA-MWSN
CA-MWSN is an algorithm designed for mobile wireless sensor network which uses fuzzy logic
to find th e cluster-head. There are three fuzzy descriptor Residual energy, Expected Residua l
Energy and Mobility with which fuzzy inferen ce engine deals to decide ‘Chance’ of a node for
being cluster-head. The node with highest ‘Chance’ is selected as c luster-head. Cluster-head
works for a round and after completion of round becomes a general node and again participate in
the election for cluster-head. The CA-MWSN algorithm is depicted below:
A-Total area
Ai- Sub-area
N- Total number of nodes
P- Total number of nodes in sub-area
RE- Residual Energy
ERE- Expected Residual Energy
EC- Energy Consumption
TH- Threshold Energy
M- Mobility
BS- Base station
X- x-coordinate of node
Y- y-coordinate of node
FIE- Fuzzy Inference Engine
CH- Cluster head
THE CALCULATIONS ARE PERFORMED AT THE BASE STATION WHICH IS
CONSIDERED AS SOURCE OF INFINITE ENERGY.
ALGORITHM 1. CA-MWSN (A, N)
1 for (i: =1 to N)
{
2 Send i (X, Y) -> BS
3 i++
4 BS divides A into Aj…………Ak
5 for Aj (j: =1 to k)
{
6 for Ajk (k: =1 to P)
{
7 Send A1k (RE, M) -> BS
8 ERE (A1k):= RE (A1k) – EC
9 If (ERE (A1k)) < TH
10 k++
11 else
12 FIE (ERE (A1k), RE (A1k), M (A1k)):= x[k]
13 k++
14 Max x[k]
15 k -> CH of A1
16 j++
}

20
}
}
5. CONCLUSIONS AND FUTURE WORK
This algorithm is able to satisfy the rational of proper coverage which is quite difficult in case of
mobile sensor nodes. Location sensor nodes are considered before selection of cluster-heads
whereby energy balance of all nodes can be maintained. However, execution of this algorithm can
result in selection of two cluster-heads which may reside quite nearer to each –other, though
residing in different sub-areas. Ergo, removal of this drawback is subsequent prospective of this
work. In future, this very algorithm can be augmented with the selection of optimal fuzzy set and
back-up cluster head. Choosing speed of nodes with random probability can be replaced with
different walk models. Multi-hop architecture can also be supported with CA-MWSN in coming
time.
ACKNOWLEDGEMENTS
The authors would like to thanks Ravi Kumar Gupta, Sachin, Shashi Ranjan and Shashi Shekhar
for their valuable assistance.
REFERENCES
[1] Akcan, Hüseyin, et al. 2006. GPS-Free node localization in mobile wireless sensor networks. In
Proceedings of the 5th ACM international workshop on Data engineering for wireless and mobile
access. ACM Press, New York, NY, 35-42.
[2] Ren Y. Oleshchuk V. Li F. and Ge X. 2011. Security in mobile wireless sensor networks–a survey.
Journal of Communications 6, 2 (Apr. 2011), 128-142.
[3] Giuseppe Anastasi, Marco Conti, Mario Di Francesco, Andrea Passarella, “Energy Conservation in
Wireless Sensor Networks: a Survey”, Elsevier Ad Hoc Networks Journal, Vol. 7, No. 3, pp. 537-568,
May 2009.
[4] Spencer BF, Ruiz-Sandoval M, Kurata N. “Smart Sensing Technology for Structural Health
monitoring.” Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver,B.C.,
Canada, Paper No. 1791. 2004.
[5] Gizem, Aksahya & Ayese, Ozcan (2009) Coomunications & Networks, Network Books, ABC
Publishers.
[6] Xing, G., Wang, X., Zhang, Y., Lu, C., Pless, R., and Gill,C. Integrated coverage and connectivity
configuration and energy conservation in sensor networks. ACM Trans. on Sensor Networks, 1, 1
(Aug. 2005), 36-72.
[7] Heinzelman, W.R.; Chandrakasan, A.; Balakrishnan, H., "Energy-efficient communication protocol
for wireless microsensor networks," System Sciences, 2000. Proceedings of the 33rd Annual Hawaii
International Conference on , vol., no., pp.10 pp. vol.2,, 4-7 Jan. 2000.
[8] Patwari, N.; Ash, J.N.; Kyperountas, S.; Hero, AO.; Moses, R.L.; Correal, N.S., "Locating the nodes:
cooperative localization in wireless sensor networks," Signal Processing Magazine, IEEE , vol.22,
no.4, pp.54,69, July 2005.
[9] Do-Seong Kim, Yeong-Jee Chung, “Self-Organization Routing Protocol Supporting Mobile Nodes
for Wireless Sensor Network”, in Proc. Computational Sciences IMSCCS, 2006, pp.622-626.
[10] G. Santhosh Kumar, Vinu Paul M V, K. Poulose Jacob, Mobility Metric based LEACH-Mobile
Protocol, 16th International Conference on Advanced Computing and Communications, 2008 IEEE.
[11] C.-M. Liu, C.-H. Lee, L.-C. Wang, Power-efficient communication algorithms for wireless mobile
sensor networks, in: Proceedings of the 1st ACM International Workshop on Performance Evaluation
of Wireless Ad hoc, Sensor, and Ubiquitous Networks, 2004, pp. 121–122.

21
[12] Chuan-Ming Liu, Chuan-Hsiu Lee, Li-Chun Wang, Distributed clustering algorithms for data-
gathering in wireless mobile sensor networks, J. Parallel Distrib. Comput. 67 (2007) 1187 – 1200,
Elsevier.
[13] Savazzi, P.; Favalli, L., "Dynamic Cell Sectorization Using Clustering Algorithms," Vehicular
Technology Conference, 2007. VTC2007-Spring. IEEE 65th, vol., no., pp.604, 608, 22-25 April
2007.
[14] Nekooei, S.M.; Manzuri-Shalmani, M. -T, "Location Finding in Wireless Sensor Network Based on
Soft Computing Methods," Control, Automation and Systems Engineering (CASE), 2011
International Conference, vol., no., pp.1, 5, 30-31July 2011.
[15] T. S. Rappaport, Wireless Communications: Principles and Practice, 2nd ed. Singapore: Pearson
Education, Inc., 2002.
Authors
Ms. Priti Kumari research focuses on wireless sensor network particularly in area of energy
conservation. Currently, she is working in designing localization algorithm for mobile wireless sensor
network. Security aspects are also being considered.

CA-MWSN: Clustering Algorithm for Mobile Wireless Senor Network

More Related Content

Similar to CA-MWSN: Clustering Algorithm for Mobile Wireless Senor Network (20)

More from GiselleginaGloria (20)

Recently uploaded (20)

CA-MWSN: Clustering Algorithm for Mobile Wireless Senor Network