The fault diagnosis research for the underwater vehicle system based on SOFCMAC

For the fault diagnosis problems of the underwater vehicle sensor systems, the solution is combined by the Principal Component Analysis (PCA) and Self-Organizing Fuzzy Cerebellar Model Articulation Controller (SOFCMAC). The signal prediction model approach based on PCA and SOFCMAC is proposed in this paper. According to the history data, it can predict the signal data in the future time using the SOFCMAC method. And the statistic, Squared Prediction Error (SPE), is introduced into the approach. According to the change of the SPE value, this model can judge whether the underwater system fault occurs. Then the linear variable reconstruction method is used to isolate the fault. The water tank experimental results show that the proposed approach is capable of detecting and isolating the fault in the vehicle sensor systems efficiently and accurately.