Localized detection of faults in wireless sensor networks

Abstract

In this paper, we present a localized fault detection algorithm for wireless sensor networks. It makes local decision based on comparisons between neighbors and disseminates the decision to neighbors. Time redundancy is used to enhance the accuracy of diagnosis and tolerate transient faults in sensing and communication. A sliding window is employed to eliminate delays involved in the time redundancy scheme. Simulation results show that sensor nodes with permanent faults are identified with high accuracy for a wide range of fault rates. Moreover, transient faults can be tolerated with negligible performance degradation.