CAST-WSN: The Presentation of New Clustering Algorithm Based on Steiner Tree and C-Means Algorithm Improvement in Wireless Sensor Networks

Abstract

In recent years, both energy consumption and network lifetime have been among the most important challenges in wireless sensor networks. Due to the limited energy of sensor nodes and the inability sensors recharge, many clustering-based methods have been offered to transmit the received information to the base station by sensors. In this paper, a new method called CAST-WSN has been presented, which is based on three criteria: the node distance from the gravity center of the cluster, the node distance from the gravity center of cluster nodes and the node distance from the energy center of nodes in each cluster. Based on these distances and Steiner tree structure, there exist two types of clustering. In this research, we present a function to evaluate the quality of cluster by means of which the quality of clusters in the two presented states can be examined and the best type of clustering can be selected. CAST-WSN method has been successful in bringing significant improvements to C-Means clustering algorithm. The results from simulation indicate that the clustering based on CAST-WSN method, as compared with the previous similar methods such as LEACH, LEACH-C, LEACH-Mobile, ALEACH, LEACH-SWDN, CBL, PECRP, HUCL, O-LEACH and DCH, has improved the network lifetime. Furthermore, CAST-WSN method has improved a main criterion called cluster density. CAST-WSN simulation results show that the clustering process in this method is better than C-Means algorithm.