Adaptive Neural-net Control System for Ship Roll Stabilization

In this paper, an adaptive neural-net control system, in which learning is performed in a loop totally independent from the control loop, is proposed for the problem of ship roll stabilization. The modeling of the ship and the controller are adjusted continuously in order to deal with changes of dynamic properties caused by disturbances. Based the experimental data in tank, disturbance model caused by sea wave is presented. A recurrent neural network is used to approaching the dynamics of the ship, and the real time recurrent learning algorithm is described to train the forward model. This paper proposes the adaptation process of control system and applies it to the ship HD702. The control effect of roll stabilization and the approaching accuracy of forward model network are investigated.