任意初值非线性不确定系统的迭代学习控制  

Iterative Learning Control for Nonlinear Uncertain Systems with Arbitrary Initial State

作　　者：严求真[1,2] 孙明轩[1] 李鹤[1] 

机构地区：[1]浙江工业大学信息工程学院,杭州310023 [2]浙江水利水电学院信息工程学院,杭州310018

出　　处：《自动化学报》2016年第4期545-555,共11页Acta Automatica Sinica

基　　金：国家自然科学基金(61174034;61374103;61573320);浙江省高等学校访问学者专业发展项目(FX2013206)资助~~

摘　　要：为解决任意初态下的轨迹跟踪问题,针对一类含参数和非参数不确定性的非线性系统,提出基于滤波误差初始修正的自适应迭代学习控制方法.利用修正滤波误差信号设计学习控制器,并以Lyapunov方法进行收敛性能分析.依据类Lipschitz条件处理非参数不确定性,对于处理过程中出现的未知时变参数向量,利用自适应迭代学习机制进行估计.经过足够多次迭代后,藉由修正滤波误差在整个作业区间收敛于零,实现滤波误差本身在预设的作业区间也收敛于零.仿真结果表明了本文所提控制方法的有效性.This paper presents a filtering-error rectified adaptive iterative learning control method to tackle the trajectory-tracking problem for a class of both parametric and nonparametric uncertain systems in the presence of arbitrary initial states. A novel rectification is made to modify the filtering-error error signal such that the learning control design and performance analysis could be simplified and easy for implementation. The proposed learning control design is a Lyapunov synthesis-based adaptive iterative learning control scheme. The Lipschitz-like assumption is used for handling nonparametric uncertainties, where the estimation for unknown time-varying parameters is given by learning mechanisms.As iteration increases, the rectified filtering-error converges to zero over the entire time interval, and the filtering-error itself converges to zero on the specified interval. Numerical results are presented to demonstrate effectiveness of the proposed learning control scheme.

关 键 词：迭代学习控制 初值问题 参数不确定性 非参数不确定性 LYAPUNOV方法 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]