Online optimal control for nonlinear fin stabilizer system of marine vessels via time-based ADP algorithm

摘要

An online optimal control for fin stabilizer system of vessels based on adaptive dynamic programming (ADP) with an unknown system dynamic is developed in this paper. A model network is firstly used to identify the unknown fin stabilizer system dynamic and the optimal control can be obtained by using recorded and current data without knowledge of system dynamic. Then, two neural networks (NN) which are action and critic network are used to approximate the optimal control policy and optimal cost function respectively, two NNs are trained by using current and recorded data of the system and update processing is given for them once at sampling time. The proposed optimal control law can reduce roll motion of vessels at desired value and guarantee all states of the system, control signals and cost function are uniformly ultimately bounded (UUB) with desired bounded errors. Simulation results are given to validate the effectiveness and performance of the proposed optimal control policy.