A New Fault Detection Approach in Wireless Sensor Networks

Abstract

Fault detection in Wireless Sensor Networks (WSNs) is of great importance to maintain the precision of data obtained from the network. Since, sensor nodes are deployed in unpleasantly rough and unfriendly environment for long duration; they are subjected to faults and failure. Once the sensor node is faulty, it generates erroneous and incorrect data which result in wrong interpretation and false alarms. Therefore, there is need for fault detection in Wireless Sensor Networks. In this paper, a new fault detection approach is proposed, which is based on Spearman's correlation coefficient and K-nearest neighbor classification algorithm. The Correlation coefficient is used for revealing the internal status of sensor nodes and K-nearest neighbor algorithm is used for classifying the abnormal nodes from the normal nodes. We simulated our proposed algorithm and MCDFD and on comparison found that the proposed algorithm outperforms the existing MCDFD algorithm in terms of detection accuracy and false positive rate.