Distributed soft fault detection algorithm in wireless sensor networks using statistical test

Abstract

Wireless sensor networks (WSN) continue to get tremendous popularity because of the increasing number of applications. Some times few sensor nodes are unable to send the correct data to the fusion center or to the neighbour node. Since the network is unaware of the faulty status of the node, so the performance degrades more. The data received from the faulty node can treat as outliers. So that a statistical test can be used to identify the outliers data send by the faulty sensor node. In this paper a neighbouring coordination based self detectable distributed fault detection algorithm and statistical Z-test is proposed. The proposed distributed algorithm is implemented in NS3 and the performance is evaluated in terms of false alarm rate (FAR) and detection accuracy (DA). The results are compared with existing algorithm and shows that the proposed approach gives better performance than the existing algorithms.