Energy-Efficient Fault-Tolerant Mechanism for Clustered Wireless Sensor Networks

Abstract

Clustering is an effective topology control and communication protocol in wireless sensor networks ("sensornets"). However, the harsh deployed environments, the serious resource limitation of nodes, and the unbalanced workload among nodes make the clustered sensornets vulnerable to communication faults and errors, which undermine the usability of the network. So mechanisms to improve the robustness and fault-tolerance are highly required in real applications of sensornets. In this paper, a distributed fault-tolerant mechanism called CMATO (Cluster-Member-based fAult-TOlerant mechanism) for sensornets is proposed. It views the cluster as an individual whole and utilizes the monitoring of each other within the cluster to detect and recover from the faults in a quick and energy-efficient way. CMATO only needs the local knowledge of the network, relaxing the pre-deployment of the cluster heads and a k-dominating set (k>1) coverage assumptions. This advantage makes our mechanism flexible to be incorporated into various existing clustering schemes in sensornets. Furthermore, CMATO is able to deal with failures of multiple cluster heads, so it effectively recovers the nodes from the failures of multiple cluster heads and the failures of links within the cluster, gaining a much more robust and fault-tolerant sensornets. The simulation results show that our mechanism outperforms the existing cluster-head based fault-tolerant mechanism in both fault coverage and energy consumption.