基于经验模态分解算法的直驱XY平台交叉耦合迭代学习控制  

Cross-couple Iterative Learning Control of Direct Drive XY Table Based on an Empirical Mode Decomposition Algorithm

作　　者：王丽梅[1] 孙璐[1] 

机构地区：[1]沈阳工业大学电气工程学院,辽宁省沈阳市110870

出　　处：《中国电机工程学报》2016年第17期4745-4752,共8页Proceedings of the CSEE

基　　金：国家自然科学基金项目(51175349);辽宁省高等学校优秀人才支持计划资助(LR2013006)~~

摘　　要：在直驱XY平台系统的交叉耦合迭代学习控制(crosscoupled iterative learning control,CCILC)过程中,由于每次运行时轮廓误差的累积,会导致系统出现收敛速度降低甚至发散的现象。针对这一问题,提出一种与经验模态分解(empirical mode decomposition,EMD)算法相结合的CCILC控制方法。首先设计CCILC控制器,直接降低轮廓误差。然后,利用EMD算法分解CCILC过程中的轮廓误差,筛选并剔除其中的发散分量,提高收敛速度和轮廓精度。仿真和实验结果表明,与传统CCILC相比,所提出的控制方法能够使直驱XY平台系统的轮廓跟踪效果更好,并且使输出轨迹在较少的迭代次数下快速且精确地收敛到期望轨迹。In the process of cross-coupled iterative learning control （CCILC） of the direct drive XY table system, the accumulation of contour errors can lead to a low convergence rate or even to system divergence. To solve this problem, a control method combining CCILC with an empirical mode decomposition （EMD） algorithm is proposed. Firstly, a CCILC controller is designed to directly reduce the contour error. Then the EMD algorithm is used to decompose the contour error in the process of CCILC, to screen and to eliminate the divergent components, so as to improve the convergence rate and contour accuracy. Simulation and experiment results show that comparing with the traditional CCILC the presented control method can improve the contour tracking result of the direct drive XY table system, and can make the output trajectory quickly and precisely converge to the desired trajectory in less number of iterations.

关 键 词：直驱XY平台 迭代学习控制 经验模态分解 收敛速度 轮廓误差 

分 类 号：TM359[电气工程—电机] TP232[自动化与计算机技术—检测技术与自动化装置]