基于奇异值分解的广义系统迭代学习控制算法  

Iterative Learning Control Algorithm for Generalized Systems Based on Singular Value Decomposition

作　　者：周秀锦[1] 田森平[1] 

机构地区：[1]华南理工大学自动化科学与工程学院,广州510640

出　　处：《电光与控制》2012年第5期33-35,共3页Electronics Optics & Control

基　　金：国家自然科学基金(60674034)

摘　　要：在对广义系统进行奇异值分解的基础上,研究了一类广义系统的迭代学习控制问题。针对快子系统和慢子系统的特点,分别利用状态误差代入输出误差,得到了一类新的广义系统迭代学习控制算法结构,这一算法是全新的。然后从理论上对所提出的算法进行完整的收敛性分析。分析结果表明,满足给定的收敛条件,系统输出可以渐近地跟踪给定的期望轨迹。The problem of iterative learning control for a class of generalized systems was investigated based on singular value decomposition of matrix.According to the different characteristics of fast subsystem and slow subsystem,the state errors were introduced into the output errors respectively,and a class of new iterative learning control algorithm for generalized systems was obtained,which was different from the present algorithm.Then,a complete analysis was made to the convergence of the proposed algorithm theoretically.The results show that the system output can track the desired trajectory asymptotically when the given convergence conditions were satisfied.

关 键 词：广义系统 迭代学习控制 矩阵奇异值分解 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TP237[自动化与计算机技术—检测技术与自动化装置]