Supervisory Recurrent Fuzzy Neural Network Guidance Law Design for Autonomous Underwater Vehicle

A guidance law, based on supervisory recurrent fuzzy neural network control (SRFNNC), is proposed for the autonomous underwater vehicle (AUV) guidance systems. This SRFNNC system is comprised of a recurrent fuzzy neural network (RFNN) controller and a supervisory controller. The RFNN controller is used to mimic an ideal controller and the supervisory controller is designed to compensate for the approximation error between the RFNN controller and the ideal controller. The proposed design method is applied to investigate the active acoustic homing guidance of an AUV, which is affected by sonar propagation time-delay and measurement noise. A comparison is made for the proposed SRFNNC, a proportional navigation (PN) and a sliding-mode control (SMC) guidance laws. Simulation results show that the proposed SRFNNC guidance law is more robust and can obtain smaller miss distance than the PN and SMC guidance laws.