Towards fault diagnosis based on agent technology for wireless sensor networks

Fault Diagnosis is an essential management task for any telecommunication network and it is even more crucial for Wireless Sensor Networks due to their dynamic nature. Based on Agent Technology, this paper presents an architecture that combines different network and diagnosis models to carry out a Fault Diagnosis process: a Causal Model to relate fault root causes with their symptoms and a Structural Model to define the network and its properties. The proposed approach has been evaluated in a simulation environment with emulated MICAz devices in a motion detection scenario.