CLARET: A Cooperative cLuster-heAd failuRE deTection mechanism for wireless sensor networks

This paper considers the cluster-head failure detection problem in WSNs and proposes a Cooperative cluster-head failure detection (CLARET) mechanism to accurately detect the failure status of a cluster head. The proposed CLARET mechanism allows the cluster head of a cluster to periodically broadcast a “heartbeat” message to all its members and each cluster member to independently detect the status of the cluster head based on the “heartbeat” messages it has received. Furthermore, it introduces a cooperative detection mechanism for a cluster member detecting a potential failure of the cluster head to confirm its detected failure by checking the status information piggybacked in the data packets sent by its neighbors. An analytical model is developed to analyze the false detection probability of CLARET. Simulation results show that the analytical model is basically effective and accurate in evaluating the false detection probability, and CLARET outperforms a traditional cluster-head failure detection mechanism in terms of both the failure detection time and the false detection probability.