Fault Self-Diagnosis Based on Mutual Information of Wireless Sensor Network's Node

The fault self-diagnosis and fault monitoring can improve the maintainability and reliability of wireless sensor network (WSN) and prolong the lifetime of WSN. In this paper, a novel method of fault self-diagnosis for the nodes of WSN is brought forward. The attribute reduction for decision-making of fault diagnosis is obtained based on mutual information algorithm for reduction of knowledge in rough set theory. A set of node's fault self-diagnosis techniques based on attribute matching in WSN are built. Finally, the faults for each module of nodes in WSN are diagnosed and located in detail. The simulation shows that this method only needs a little expenditure of communication, very low energy of nodes in WSN and low misdiagnosis rate when fault self-diagnosis is implemented in WSN.