K-Means Like Minimum Mean Distance Algorithm for wireless sensor networks

Abstract

A wireless sensor network consists of hundreds or thousands of low cost, small wireless sensor nodes which allow users to monitor a variety of remote environments. In this paper, a novel cluster-head selection algorithm is presented and analyzed which uses the minimum mean distance between sensor nodes as a selection parameter. The proposed algorithm has clear advantages and takes 1.2 times longer to reach the point where 50% sensor nodes remain alive than the Low Energy Adaptive Clustering Hierarchy algorithm (LEACH) while maintaining information throughput at a high level. Furthermore, the new algorithm takes 1.2 times longer than the LEACH to reach the point where the number of cluster-heads and the throughput from the cluster-heads to the base station are reduced by 50%. In the example 100 sensor network presented, the new algorithm ensures that there are always five cluster-heads during each round throughout the network lifetime, which is longer than the LEACH protocol and delivers 2.1 times more data to the base station compared to the LEACH protocol. All simulation results show that this new algorithm can significantly improve the overall wireless sensor network performance compared with the LEACH.