Optimal Control Of Ship Course Based On Adaptive Dynamic Programming

With the continuous development of social economy and control theory,the requirements for ship course controllers are constantly increasing.This paper is mainly aimed at the ship course control system,designing an economical and energy-saving controller with better control performance.Therefore,the main research contents of this article are as follows:First,an adaptive dynamic programming algorithm is designed based on the BP neural network,and an optimized control strategy is obtained through the design of the critic-action network structure,and finally the ship’s optimal controller is realized.In order to solve the problem of unknown model,BP neural network is applied to identify unknown system dynamics.This method solves the problem of ship model parameter information that cannot be accurately obtained.The BP neural network is trained using ship input-output data and verified by simulation.On this basis,the neural network is used to design the critic-action network structure to approximate its optimal cost function and optimal control strategy,respectively,so as to obtain the optimal control of the ship’s course system.Then the stability analysis proves the states of the closed-loop control system and the weights approximation errors are UUB by the Lyapunov theorem.Secondly,the broad learning system is introduced into the adaptive dynamic programming algorithm to effectively improve the accuracy and speed of the optimized control algorithm.The broad learning system greatly improves the learning efficiency.It directly obtains the weights based on the ridge regression algorithm,thus eliminating the iterative process of traditional methods and reducing training time.At the same time,applying it to the identification of unknown systems can effectively improve the accuracy of identification.It is also applied to the approximation of the optimal cost function and the optimal control strategy in the critic-action network,so as to obtain the optimal control law.Then the stability analysis proves the states of the closed-loop control system and the weights approximation errors are UUB by the Lyapunov theorem.Finally,an optimal control algorithm based on broad learning system and adaptive dynamic programming considering input saturation is proposed,which effectively solves the input saturation problem in the actual ship course control system.The algorithm uses broad learning system to identify nonlinear systems.A non-quadratic functional is introduced into the cost function to deal with the ship rudder angle limitation,and the optimal cost function and optimal optimal control strategy are approximated by the broad learning system,then the optimal control law of the ship heading considering the input saturation is obtained.Then the stability analysis proves the states of the closed-loop control system and the weights approximation errors are UUB by the Lyapunov theorem.It provides an effective theoretical basis for the actual ship control system.The optimal algorithms proposed in this paper are simulated in MATLAB,which proves the effectiveness of the algorithms.