含死区输入不确定系统的鲁棒学习控制

Robust Learning Control for Nonlinear Systems With Dead-zone Input

针对一类在有限时间区间上重复运行的含死区输入非线性不确定系统,提出一套鲁棒迭代学习控制算法,用于实现系统状态在整个作业区间对参考信号的精确跟踪。该算法采用Lyapunov综合方法设计迭代学习控制器,并利用鲁棒学习方法处理死区特性及非参数不确定性。理论分析与仿真结果表明,所提方法可以实现系统状态在整个作业区间对参考信号的精确跟踪,并保证闭环系统所有变量有界。

In this paper,we present a robust iterative leaning control scheme for a class of non-parametric systems with deadzone input,aiming to realize the system state perfectly tracks its reference signal over the full interval.The controller is designed by Lyapunov synthesis approach.Robust strategy and learning strategy are unitedly used to deal with non-parametric uncertainties and dead-zone nonlinearity.It shows that the closed-loop system can get good performance both in the theoretical analysis and in the numerical simulation.And all signals in the closed-loop system are guaranteed to be bounded.