A Hidden Markov Model based Prediction Mechanism for Cluster Head Selection in WSN

Abstract

Energy is an important resource and requires efficient management and utilization. Energy conservation is fundamental concern for designing in WSN. Clustering is an effective technique for energy conservation in WSN. With the increase in large scale networks, the need of energy aware and efficient clustering algorithms is rising. The efficacy of clustering depends on appropriate selection of ClusterHead(CH). The rotation of CH involves gathering energy levels of the nodes in every round; this augments the cost of cluster Head rotation. In this paper, a Prediction Based CH Selection (PBCS) algorithm is proposed which uses combined Particle Swarm optimization to form clusters and Hidden Markov Model based technique to choose the (CH) based on predicted energy level of nodes. The proposed approach is implemented and result reveals that network lifetime is improved and messages required for CH selections are reduced.