一种新的时变大纯滞后对象在线辨识方法  

A New Method for Online Identification of Time-Varying System with Dominant Time Delay

作　　者：李鑫[1] 党选举[1] 薛万伟[1] 

机构地区：[1]桂林电子工业学院智能系统与工业控制研究室,广西桂林541004

出　　处：《计算机仿真》2005年第5期95-97,共3页Computer Simulation

基　　金：广西区教育厅科研项目(桂教科研2003[22号])

摘　　要：对具有时变特性的大纯滞后对象进行参数估计是系统辨识的一项重要课题,该文结合递归遗忘最小二乘法和互相关函数,提出了一种联合辨识算法,该方法扩展了递归遗忘最小二乘算法的应用范围。其中,递归遗忘最小二乘法用于快速地辨识对象的动态参数,而对输入输出信号进行滤波后,对象的滞后时间参数的辨识由互相关函数实现。仿真结果表明,此种方法对于时变大纯滞后对象的在线辨识是有效的,并且优于间接估计法;该文的方法具有算法简便,实用性强的特点。<Abstrcat>The parameter estimation for time-varying system with dominant time delay is an important issue for system identification. A mixed identification method, which extends the range of application of RFF algorithm, combining RFF algorithm with mutual correlation technique is presented for time-varying systems with dominant time delay. The dynamic parameters of the system are identified quickly via RFF algorithm, and the time delay parameter of the one is estimated by mutual correlation technique after input signal and output signal being filtered .The effectiveness of the proposed method is finally demonstrated through simulation on the time-varying system with dominant time delay, and it is better than the indirect estimation algorithm. The proposed parameter estimation algorithm is simple and easy to be applied。

关 键 词：递归遗忘最小二乘法 互相关函数 在线辨识 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]