Distributed Fault Detection for Wireless Sensor Based on Weighted Average

Abstract
This paper presents a distributed fault detection algorithm for wireless sensor networks (WSNs) by exploring the weighted average value scheme. Considering the spatial correlations in WSNs, a faulty sensor can diagnose itself through comparing its own sensed data with the average of neighbors' data. Simulation results show that sensor nodes with permanent faults are identified with high accuracy for a wide range of fault rate, and keep false alarm rate for different levels of sensor fault model.

This publication has 10 references indexed in Scilit: