Fault diagnosis for AUVs using support vector machines

In this paper an observer-based fault diagnosis (FD) approach for autonomous underwater vehicles (AUVs), subject to actuator faults (i.e., faults affecting the propulsion system and/or the control surfaces), is proposed. A diagnostic observer is developed based on the available dynamic model of the AUV. Compensation of unknown dynamics, uncertainties and disturbances is achieved through the adoption of a class of neural interpolators (support vector machines, SVMs) trained off line. On the other hand, interpolation of unknown actuator faults is performed by adopting a radial basis function (RBF) network, whose weights are adaptively tuned on line. The effectiveness of the approach is tested in a simulation case study developed for the NPS AUV II (PHOENIX) vehicle.