离散时间系统自适应迭代学习控制的研究进展

Research Advances on Discrete-time Adaptive Iterative Learning Control

离散自适应迭代学习控制是针对在有限时间区间上运行的不确定非线性离散时间系统提出的一类方法,可有效抵抗系统的不确定性,并放宽了传统迭代学习控制中要求相同初始条件和参考轨迹这两个关键假设。即可在随机初始条件下实现对迭代变化参考轨迹的几乎完全跟踪性能。本研究给出了迭代学习控制方法的分类,对其中的自适应迭代学习控制方法的设计思路和适用背景进行了阐述。重点综述了离散时间系统自适应迭代学习控制方法的发展过程,讨论了所提出离散时间自适应控制方法的特点和适用范围,提出了基于数据驱动的自适应迭代学习控制发展的必然趋势和有待于进一步研究的问题。

Discrete adaptive iterative learning control（DAILC）is proposed for a class of nonlinear discrete-time systems,which are operating on a finite time interval repetitively.By applying DAILC,the system uncertainties can be repelled effectively.Moreover,the match conditions of identical initial state and identical reference trajectory,which are the two key assumptions in traditional ILC,are relaxed by the DAILC.In other words,the DAILC can achieve an almost perfect tracking performance under random initial state and iteration-varying reference trajectory.The classification is given for the existing iterative learning control（ILC）methods,where the adaptive iterative learning control（AILC）is elaborated with its design method and its possible applicability.On the basis of a brief review on AILC of continuous-time systems,a survey on discrete-time adaptive iterative learning control（DAILC）is presented prominently with its increasing developments,and the advantages and applicability of different DAILCs are also discussed in this paper.It is emphasized in this work that data-driven adaptive ILC is an inevitable trend of AILC.Some prospective research topics are also listed in the conclusion section.