Nonlinear robust fin roll stabilization of surface ships using neural networks

A nonlinear robust control method using neural networks is developed to stabilize ship roll. Fins driven by hydraulic systems are used to provide the active moment against that caused by waves. The useful nonlinearities of restoring forces and damping are taken into account to reduce conservation of the control law. Neural networks are utilized to approximate the rest of unknown nonlinearities in ship roll dynamics. The roll angle and velocity are guaranteed to converge to arbitrary small values by adjusting gain coefficients. Since the stability analysis is based on the Barbashin and Krasovkii theorem, mean valued theorem and Lyapunov stability theory, the off-line training phase is removed