神经网络自抗扰全垫升气垫船航迹控制

针对全垫升气垫船与常规水面船相比,非线性强,可操纵性较差,设计了神经网络自抗扰航迹引导控制器及航向控制器,以提高其航迹控制效果.利用非线性跟踪微分器(TD)提取微分信号并安排过渡过程.利用扩张状态观测器(ESO)对系统内、外扰动进行观测,并进行扰动补偿.利用非线性状态误差反馈控制律(NLSEF)提供精确控制量.为了提高自抗扰控制器( ADRC)的自适应性,利用递归网模型(RNM)对系统进行辨识,根据辨识信息对ADRC控制参数进行在线整定.仿真结果表明:在强扰动作用下,所设计航迹控制系统能够使全垫升气垫船精确地保持在计划航线上,具有自适应性强、超调量小、鲁棒性强等特点.

Trace-keeping of an air cushion vehicle based on an auto disturbance rejection controller with a recurrent networks model

Compared with conventional surface vessels,the nonlinear characteristics of an air cushion vehicle are obvious,making it difficult to manipulate.In order to improve the control quality of trace keeping,an auto disturbance rejection controller(ADRC)with a recurrent networks model(RNM) was designed for trace-guidance and heading control.A nonlinear tracking differentiator(TD) was used to obtain differential signals and realize transition arrangement.Interior and external disturbance can be estimated by the extended state observer(ESO),and the system can be simultaneously compensated.The nonlinear state error feedback(NLSEF) control law was used to generate control input.System information identified by RNM was used for real-time optimization of the ADRC coefficients.Simulations show that the proposed controller has strong adaptability.A good trace-keeping control effect can be obtained with small overshot and robustness in a harsh environment.