부산대학교 도서관

We investigate the problem of energy conservation in ClusterHead rotation in wireless sensor networks. Nodes are battery powered without being replenished, thus maximizing lifetime of network by minimizing energy consumption poses challenge in design of protocols. We argue, though clustering addresses lifetime and scalability goals, it results to an expensive load-balancing scheme based on ClusterHead rotation (i.e. re-clustering).The load balancing technique of existing clustering schemes uses global rotation of ClusterHead roles in order to prevent any single node from complete energy exhaustion. Theoretically, the problem of rotating the responsibility of being a ClusterHead has been abstracted as domatic partitioning problem for maximum cluster-lifetime problem. Instead, this work presents an analysis of its design and implementation aspects. We propose a domatic partitioning based scheme for ClusterHead rotation in clustering protocol. Our self-organizing protocol achieves energy-conservation in achieving local load balancing of nodes. The simulation results demonstrate that our approach outperforms re-clustering in terms of energy consumption, and lifetime parameters