SENSOR FAULT DIAGNOSIS OF AUTONOMOUS UNDERWATER VEHICLE BASED ON NONLINEAR PRINCIPALCOMPONENT ANALYSIS

An approach of sensor fault diagnosis based on a NLPCA (Nonlinear Principal Component Analysis) model is described in this paper. Firstly, in order to perform NLPCA, a double RBF(Radial Basis Function) neural network is introduced; Secondly, NLPCA fault detection and isolation algorithm is presented; Finally, A simulation example of sensor fault diagnosis of autonomous underwater vehicle is given to show the performances of the proposed approach.