基于径向基函数神经网络的船舶航迹自抗扰控制

Active disturbance rejection control for ship trajectory based on radial basis function neural network

针对常规非线性自抗扰控制(nonlinear active disturbance rejection control,NLADRC)技术在船舶航迹控制中存在的参数整定难、抗干扰能力差的问题,搭建船舶三自由度MMG数学模型,设计船舶航迹NLADRC系统。利用径向基函数(radial basis function,RBF)神经网络对系统进行辨识,使网络输出逼近系统输出。根据辨识信息自适应整定对系统整体控制效果影响较大的两个参数,提出基于RBF神经网络的船舶航迹NLADRC系统。仿真结果表明,参数的自适应整定能加快系统收敛速度,大幅减小超调量,对外界环境具有更强的鲁棒性。神经网络对NLADRC的优化,使其控制性能得到提升。

In view of the problem of difficult parameter tuning and poor anti-interference ability of the conventional nonlinear active disturbance rejection control(NLADRC),a ship three-degree-of-freedom MMG mathematical model is established,and a ship trajectory NLADRC system is designed.The radial basis function(RBF)neural network is used to identify the system so that the network output approximates the system output.According to the identified information,the two parameters that have great influence on the overall control effect of the system are adjusted adaptively,and a ship trajectory NLADRC system based on RBF neural network is proposed.The simulation results show that the adaptive tuning of parameters can accelerate the convergence speed of the system,greatly reduce the overshoot,and have stronger robustness to the external environment.The NLADRC optimization by the neural network improves the control performance.