Study On Clustering Algorithm And Backup Mechanism In Wireless Sensor Networks

Clustering technology is an important technology of topology control in wireless sensor networks (WSN). Through clustering technology, not only the networks' expansibility can be advanced and resource can be managed effectively, but also energy can be saved. How to effectively implement clustering in large-scale WSN is the desiderate problem need to settle. In this paper, the clustering algorithm and cluster-head back-up mechanism for WSN are researched. The research work is as follows:First, a clustering algorithm based on K-means clustering for WSN is proposed. Since the questions of the determining number of the best clusters, the choice of initial cluster centers and objective function are settled, then the sub-clusters of WSN can be achieved with the K-means clustering algorithm. Finally, evaluating the sub-cluster algorithm performance and simulating it which shows that the WSN clustering algorithm based on the K-means clustering can not only save energy and prolong the life of the network, but also improve the energy consumption of a balanced network and ensure uniform distribution of the clustering heads.Second, a cluster-head back-up mechanism based on the integrated index for WSN is proposed. With node's residual energy, node degree and communicate price, an effective integrated index is constructed. It selects node with the best integrated index as the back-up cluster-head and maintain the continuous rotation between the cluster-head and backed-up cluster-head. The proposed method is implemented and the simulating results show that the clustering WSN with this kind of mechanism can reduce the losses caused by the malfunction of cluster-heads effectively, enhance the stability of the sub-clusters, extend the life cycle of the network and improve the overall performance of the network.