基于神经网络滑模的欠驱动船舶路径跟踪与避障协同控制

Collaborative Control of Path Following and Obstacle Avoidance for Underactuated Ships Based on Neural Network Sliding Mode

【目的】针对存在模型不确定性和外界环境干扰的欠驱动船舶路径跟踪与避障问题,结合反演法与径向基函数(RBF)神经网络技术,提出一种神经网络滑模自适应控制律和改进的人工势场。【方法】首先根据误差方程设计辅助纵荡速度和艏摇角,然后分别对控制输入设计滑模面,并利用RBF神经网络逼近总未知项,设计控制律和自适应律。【结果与结论】Lyapunov稳定性分析证明闭环系统误差是一致最终有界的。对静态、动态障碍物分别改进人工势场,克服局部极小值问题以及未考虑船舶和障碍物的位置、相对速度关系问题。仿真对比结果表明,在海浪干扰下船舶路径跟踪误差收敛精度更高,且避障更安全。所提控制方法可改善路径跟踪与避障控制效果,验证了所提控制算法的有效性和鲁棒性。

【Objective】To solve the path following and obstacle avoidance problems of underactuated ships with model uncertainty and external environmental disturbance,a neural network sliding mode adaptive control law is proposed by combining the backstepping method and radial basis function(RBF)neural network technology,and an improved artificial potential field is also proposed.【Method】Firstly,based on the error equation,the auxiliary surge velocity and yaw angle were designed.Then,sliding surfaces were designed for the control inputs separately.The RBF neural network was used to approximate the total unknown terms,and control laws and adaptive laws were designed.【Result and Conclusion】Lyapunov stability analysis proved that the closed-loop system error was ultimately uniformly bounded.Design improved artificial potential field for static and dynamic obstacles respectively to overcome the problems of local minimum and not considering the relationship between the position and relative velocity of ships and obstacles.The comparative results of simulation show that under the disturbance of waves,the convergence accuracy of the ship's path following error is higher and the obstacle avoidance is safer.The proposed control method can improve the path following and obstacle avoidance control effect.The effectiveness and robustness of the proposed algorithm have been demonstrated.