Fault Tolerant and Energy Aware Clustering for improving the Lifetime of Wireless Sensor Networks with Isolated Nodes

Clustering techniques are used in to improve the lifetime of Wireless Sensor networks. The cluster heads receive the data from the cluster members, perform aggregation of the data and send the data to the sink where control decisions are done. Existing clustering protocols like LEACH uses probability based cluster head selection mechanism which may results in isolated sensor nodes and association of more cluster members to cluster heads. Due to this, the heavily loaded cluster heads may exhaust their energy and die soon. The isolated nodes will send the data directly to the sink and may also die soon. Since the sensor nodes are subjected to failure, if the selected cluster head is found to be a faulty node, then it affects the reliability of data transfer. Many of the existing algorithms does not use fault tolerance as a metric in the cluster head selection. This paper proposes a novel Fault Tolerant and Energy Aware clustering for improving the lifetime of the WSNs with Isolated Nodes. To select non faulty nodes as cluster heads and to reduce the number of isolated nodes, remaining energy, distance, measurement correlation and number of isolated nodes are used as parameters in the clustering algorithm. To save the energy consumption, isolated nodes will send the data to the nearest fault free cluster member. If all the nearest cluster members are found to be faulty, then the isolated node sends the data directly to the sink node. The simulations are done in Matlab and the results prove that the proposed work improves the network lifetime.