A Survey on Cluster Head Selection Techniques

Abstract

Wireless Sensor Networks are becoming very popular now days as of low cost and easy to deploy and maintain. The network consists of collection of sensor nodes which are capable of computing, sensing and communicating. Sensor nodes are equipped with limited energy and are deployed in inaccessible areas so it is hard to replace the batteries. Therefore to increase the lifetime of the network proper clustering and cluster head selection methods should be adopted. In this paper we investigate fuzzy, genetic and neural network based cluster head selection methods with their working techniques. Motivation behind genetic algorithm is Darwin’s theory of evolution. Darwin suggested that an individual who is fittest will survive in the competition of the existence. Genetic algorithm selects a node as a cluster head depending upon its fitness i.e. node which has higher fitness will be a candidate for cluster head selection. Fuzzy logic can be used to work on partial data. Fuzzy logic variable can have a partial truth value. In neural network, three layers are used. Nodes in the input layer will match the input pattern and node in the output layer is a cluster head.