An Improved K-Means Algorithm Based on Density in the Selfish Behavior Detection

Abstract

Aiming at selfish nodes in wireless sensor network (WSN) competition Shared channel unfair use of channel resources, for the purpose of selfish node Detection, puts forward an Improved Clustering Method (Improved Density-based Clustering Detection Method, ID-BCDM).With an average transmission delay, average throughput of node link sequence and collision probability of network performance characteristics, add the density function, improved K-means algorithm was carried out on the characteristics of sequence analysis and clustering. The simulation test results show that the ID-BCDM on three dimensional feature space has good clustering effect, also able to complete detection of single and multiple selfish node.