Machine Learning Based Effective Clustering Scheme for Wireless Sensor Networks

In Wireless Sensor Networks (WSN) sensor nodes are burdened by the exchange of messages caused by successive and recurring re-clustering, which results in power loss. Presently researchers have been concentrating on enhancing the longevity of these nodes due to non-rechargeable batteries fitted in Sensor Nodes (SN). Clustering mechanism has emerged as a desirable subject because, it is predominantly good at conserving the resources especially energy for network activities. In this work, the problem of load balancing and Cluster Head (CH) selection with minimum energy expenditure is proposed. Unsupervised machine learning based k-means algorithm is employed to form cluster and fuzzy based approach is used to select the CH. Simulation results shows the effectiveness of proposed work in terms of energy usage and CH identification delay.