Adaptive clustering approach over unreliable links for WSN

Clustering techniques for WSN have been proved to be an effective approach to improve network lifetime and scalability. As for link failure inherent over wireless communication links, traditional WSN routing protocol suggested using multi-path routing techniques for reliable packet delivery. However, establishment and maintenance of multiple data forwarding paths to a sink node are cost-expensive for energy-constrained sensor nodes. In this paper, we first design an energy-efficient cluster forming protocol for WSN based on relative difference between energy level of an ordinary node and its cluster head node in the previous round. Then we present a cluster head selection process for member nodes based on the estimated communication reliability of a neighboring head node. To estimate reliability of a link to a neighboring node, we introduce a sliding window to record communication status during previous round via acknowledgement message. Simulation results show that the presented clustering protocol over unreliable links can prolong network lifetime when compared with other energy-efficient clustering protocol for WSN.