Clustering Approach In Wireless Sensor Networks Based On K-Means: Limitations And Recommendations

Clustering approach in wireless sensor network is very important, the structure of cluster and how to improve it is a first challenge that faced the developers, because of it represent as a base for design the cluster-based routing protocol. One of most popular cluster algorithms that utilizing into organize sensor nodes is K-means algorithm. This algorithm has beneficial in construct the clusters for various real-world applications of WSN.K-means algorithm suffering from many drawbacks that hampering his work.The lack of adequate studies that investigates in the limitations of this algorithm and seek to propose the solutions motivated us to do this study. In this paper the limitations of K-means and some suggestions are proposed. These suggestions can improve the performance of K-means, which will be reflected on saving the energy forsensor nodes and consequently maximize the lifetime of the wireless sensor networks