Novel Approach of Fault Diagnosis in Wireless Sensor Networks Node Based On Rough Set and Neural Network Model

Nodes of wireless sensor network (WSN) will appear various faults, because the influence of many unavoidable factors and environment is very complex and harsh. Rough set can deal with incomplete information, especially in the data reduction, and it is easy to realize low energy consumption problem of on-line fault diagnosis based on WSN node energy Co. This paper adopts attribute reduction algorithm by integrate rough set with neural network model to eliminate WSN node failure, so as to achieve data reduction and to improve the accuracy and efficiency of fault diagnosis purpose. The paper makes use of rough set and neural network to the failure phenomenon of WSN node by using knowledge reduction of discernibility matrix and logic operation, eliminating the redundant attribute WSN node fault. Then, fault decision complex table is built by he classified fault, and finally determine the fault location corresponding to fault phenomenon and repair of the final decision table. The experimental results show that this method improves the robustness of the fault diagnosis, and enhances the practicability of WSN limited energy.