Distributed diagnosis of wireless sensor networks

In wireless sensor networks deployed in a harsh environment, sensor nodes are prone to have faults. It is thus desirable to detect and locate faulty sensor nodes to ensure the quality of service of the sensor networks. This paper presents an efficient distributed algorithm for isolating faulty sensor nodes. It is based on local comparisons of sensed data and dissemination of the test results to neighboring nodes. Time redundancy is used to tolerate some transient faults in sensing and communication. Computer simulation shows that faulty sensor nodes are identified with a high accuracy for a wide range of fault rates.