Soft Threshold Based Cluster-Head Selection Algorithm for Wireless Sensor Networks

In recent years, many clustering algorithms have been proposed. Among them, LEACH is the most famous one. However, in LEACH, within each 1/p rounds, once a node has been selected as a cluster-head (CH), its threshold will be set to 0, and thus it will lose the chance to participate cluster-head selection, even if it still has enough energy. In this paper, we present a novel cluster-head selection algorithm. Instead of changing the threshold to 0 directly, the proposed algorithm adjusts the threshold of each node gradually according to the roles they have played in the last round, so more nodes could have the opportunity to be CHs. Simulation results show that the proposed algorithm outperforms LEACH in network lifetime by an average of 30% approximately.