Fault Diagnosis Method of WSN Nodes Based on Wavelet Packet and Belief Rule Base

Abstract

In order to diagnose node faults in Wireless Sensor Network (WSN), a new fault diagnosis method based on wavelet packet and belief rule base (BRB) is proposed. Firstly, based on the data set of wireless sensor, the experimental samples of wireless sensor nodes under different typical fault conditions are established through fault simulation method; Secondly, three-layer wavelet packet decomposition method is used to calculate the energy of each frequency band of the samples, and the feature vectors of WSN nodes faults are according to the energy ratio of every frequency band in normal operation; Finally, the fault diagnosis method of WSN nodes is proposed by using the BRB model and the corresponding expert knowledge. The experimental results show that the proposed method can effectively use qualitative knowledge and quantitative data to establish a non-linear model between input and output with small samples, which can achieve satisfactory results for WSN nodes faults diagnosis.