多变量非线性系统的变阶采样迭代学习控制  

Varying-Order Sampled-Data Iterative Learning Control for MIMO Nonlinear Systems

作　　者：孙明轩[1] 李芝乐[1] 朱胜[2] 

机构地区：[1]浙江工业大学信息工程学院,杭州310023 [2]浙江大学城市学院,杭州310015

出　　处：《自动化学报》2013年第7期1027-1036,共10页Acta Automatica Sinica

基　　金：国家自然科学基金(61174034;60874041);浙江省自然科学基金(LQ12F03005)资助~~

摘　　要：针对存在初态误差的情形,提出多变量非线性系统的变阶采样迭代学习控制方法.相对固定阶迭代学习算法,变阶算法可有效降低跟踪误差.对变阶采样迭代学习算法进行了收敛性分析,推导出收敛充分条件.给出了变阶学习的两种实现策略-DD(Direct division)和DIP(Division in phases)策略.数值仿真表明,基于DIP策略的变阶采样迭代学习算法在获得较高的控制精度的同时,具有较快的收敛速度.In this paper,the problem of sampled-data iterative learning control is addressed for a class of nonlinear MIMO systems in the presence of perturbed initial conditions.In contrast to the fixed-order learning algorithms,a varying-order learning algorithm is proposed,for enhancing tracking performance against repositioning errors.Sufficient convergence conditions of the proposed varying-order learning algorithm are given,by which the learning gain can be chosen.The proposed learning algorithm is shown to be a unified one,because it is applicable to the systems with arbitrary but well-defined relative degree.Two implementation schemes,sampled-data direct division（DD） and division in phases（DIP） schemes,are presented,and numerical results are given to demonstrate effectiveness of the proposed schemes.

关 键 词：多变量系统 采样系统 初始修正作用 变阶迭代学习控制 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]