Radial basis function neural network based rudder roll stabilization for ship sailing in waves

This paper presents a rudder-roll stabilization system utilizing Radial Basis Function neural network (RBFNN) for course keeping and roll damping. Roll motion of a vessel sailing under severe weather conditions has adverse effects on crews' health, cargoes and safety, thus it must be damped as much as possible. A new control algorithm for both course keeping and roll damping is proposed based on the RBFNNs. In order to realize the proposed rudder roll stabilization system, a nonlinear mathematical model of a container vessel with effects of wave disturbance is used to simulate the proposed rudder roll stabilization system which consists of two controllers implemented in parallel, one is the autopilot for course keeping and the other is roll damping controller. The performance and robustness of the proposed control system is investigated by taking consideration of the effects of external disturbance. The simulation studies are designed to verify the improved performance of the proposed rudder roll stabilization system and to validate its efficiency of course keeping and roll motion reduction.