Path following of ship based on sliding mode control with improved RBF neural network and virtual circle

结合改进RBF与虚拟圆弧的船舶路径滑模控制

Received: 6 June 2020

Abstract

To address the unmeasured velocity, external disturbance and internal model uncertainty for following the path of an under-actuated ship, the paper presents a sliding mode control method based on the radial basis function(RBF) neural network and the velocity observer. To enhance the RBF performance of approximating the unknown, an arc tangent function was exploited in the RBF neural network to update its weight values. Then, the nonlinear observer was built via the hyperbolic tangent function to deal with the unmeasured velocity of the ship. Furthermore, in order to avoid overshoots when the ship is moving to its way points, the virtual paths of a variable circle based on the turning angle were designed at the joints of the path of the ship to enhance its path following capability. Finally, the simulation results show that the sliding mode controller designed in the paper can force the ship to follow accurately the reference path in case of time-varying disturbances without measured velocity and enhance the path following performance of the ship and the accuracy of the RBF neural network, thus demonstrating its effectiveness.

摘要

为解决欠驱动船舶路径跟踪中存在速度状态不易获取﹑外界环境干扰及内部模型不确定等问题，提出结合速度观测的径向基函数(RBF)神经网络滑模控制算法。并为改进RBF对未知项的逼近能力，引入反正切函数对RBF权值进行更新。为处理船舶速度不可测问题，基于双曲正切函数建立了非线性观测器。此外，为避免船在转向点处容易产生超调的情况，提出在路径衔接处根据转向角大小而设计可变圆弧的虚拟路径，以提高路径跟踪性能。最终对比仿真表明，在不需获取速度值的情况下，控制器仍能使船在时变干扰下准确地跟踪上参考路径，并提高了路径跟踪性能和RBF逼近性能，验证了所提方法的有效性。