控制方向未知的无人帆船自适应动态面航向控制

Adaptive dynamic surface course control for an unmanned sailboat with unknown control direction

针对四自由度无人帆船存在模型不确定、控制方向和外部环境扰动均未知的情况,本文提出一种神经网络自适应动态面航向控制方法。该方法采用RBF神经网络逼近无人帆船模型不确定部分,利用Nussbaum函数处理系统的未知控制增益函数,并设计σ-修正泄露项的参数自适应律对神经网络逼近误差与外界环境扰动总和的界进行估计,同时引入动态面方法,消除反演法中的"计算膨胀"问题,降低控制器的复杂性。Lyapunov函数稳定性分析证明所设计控制器能够保证航向保持闭环系统内所有信号的一致最终有界性,并通过一艘12 m型无人帆船模型进行仿真验证。结果表明:无人帆船航向保持响应速度快,所设计的控制器能有效地处理模型不确定项和风浪等外界扰动,具有较强的鲁棒性。

A neural network-based adaptive dynamic surface course control method is proposed for cases of 4 degrees of freedom(DOF)unmanned sailboat model uncertainty,whereby both the control direction and the external environmental disturbances are unknown.In this strategy,the neural network was used to approximate the model′s uncertainty.The problem of unknown control gain was properly solved by using the Nussbaum gain function.The adaptive laws based on the leakage term,σ-modification was used to estimate the bounds of neural network errors and unknown external environmental disturbances.Additionally,dynamic surface control technique was introduced to eliminate the“computational expansion”problem of the backstepping method.The stability analysis of the Lyapunov function proved that all signals of the resulting closed-loop system can be guaranteed with the uniformly ultimate boundedness of the proposed controller.Simulation results based on a 12 m unmanned sailboat model showed that the unmanned sailboat course-keeping response speed was fast,and the design controller had strong robustness against the system model′s uncertainty,wind,flow,and other external disturbances.