A Fuzzy based Approach to Cluster Formation in Wireless Sensor Networks to Enhance the Lifetime of the Network

Energy efficient routing of information from sensing node to sink is achieved using clustering of nodes, thereby reducing the number of hops and rotating the role of the cluster head among the sensor nodes periodically. The life time of the network is further enhanced by reducing the number of control packets (overhead) in the network. The process of cluster formation is as follows. (1) The fuzzy c-means (FCM) algorithm is used by the sink to find cluster centers and their associated member nodes with a high membership grade. At the beginning of the network operation, the sink selects a node randomly from the associated members of each cluster center and broadcasts them as cluster heads. (2) Cluster heads broadcast (with highest transmission power) the advertisement of their selection as cluster head for the next round. The nodes in the network choose the nearest cluster head to form the clusters. Thus the cluster formation is adaptive and distributed. (3) At the end of each round, the nodes transmit their fitness to become the next cluster head to their respective current cluster head that in turn is forwarded to the sink. We assume that each node in the network compute its fitness to become the next cluster head using some network parameters. (4) The sink selects a node with highest fitness value from each cluster center associated member nodes as cluster head for the next round and broadcasts. Hence the selection of the cluster heads by the sink is centralized using FCM which helps to disperse the cluster heads throughout the network. (5) Above steps are repeated from step 2 till the death of the last sensor node in the network. The lifetime of the network is measured in terms of number of rounds, first node death (FND), half node alive (HNA) and last node death (LND) in the network.