带扰动观测器的船舶航向自适应神经网络控制

Adaptive neural network control for ship course with disturbance observer

针对遭受未知外界扰动且存在不确定项的船舶航向非线性操纵系统,提出一种基于扰动观测器的自适应神经网络控制策略.构造神经网络逼近船舶模型中的不确定项,利用扰动观测器估计未知的外界环境干扰,并对控制量进行前馈补偿,提高控制策略的鲁棒性.应用Lyapunov函数证明了船舶航向跟踪闭环系统的误差信号一致最终有界.仿真结果表明,所设计的控制策略可以迫使船舶跟踪期望的航向.相对于传统的PID控制器,所设计的控制策略可以有效克服未知外界扰动,具有较强的鲁棒性.

Aiming at the nonlinear steering system of ship course subjected to unknown external disturbances and uncertain items,an adaptive neural network control strategy based on a disturbance observer was proposed.The neural network was constructed to approximate the uncertain items in ship model.A disturbance observer was developed to estimate the unknown external disturbances and make feedforward compensation for control amounts so as to improve the robustness of control strategy.By using a Lyapunov function,the ultimately uniform boundedness of error signals in the closed-loop system for ship course tracking was proved.The simulation results show that the as-proposed control strategy can force the ship to follow desired course.Compared with the traditional PID controller,the as-proposed control strategy can effectively overcome unknown external disturbances and has strong robustness.