引用本文:黄祖毅,陈建清.基于脉冲响应的输出误差模型的辨识[J].控制理论与应用,2003,20(5):793~796.[点击复制]
HUANG Zu-yi,CHEN Jian-qing.Identification of output error models based on impulse response[J].Control Theory and Technology,2003,20(5):793~796.[点击复制]
基于脉冲响应的输出误差模型的辨识
Identification of output error models based on impulse response
摘要点击 1661  全文点击 2143  投稿时间:2002-06-15  修订日期:2003-03-12
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/j.issn.1000-8152.2003.5.030
  2003,20(5):793-796
中文关键词  辨识  参数估计  最小二乘
英文关键词  identification  parameter estimation  least squares
基金项目  国家自然科学基金(60074029); 国家自然科学基金重点项目(69934010); 清华大学信息学院创新基金(985信息-50-基金-50)资助项目
作者单位E-mail
黄祖毅 清华大学 热能工程系, 北京 100084 huangzy01@mails.tsinghua.edu.cn 
陈建清 清华大学 自动化系, 北京 100084  
中文摘要
      基于系统脉冲响应参数, 利用相关分析方法, 提出了一种辨识输出误差模型参数的方法. 该方法是利用有限脉冲响应模型逼近输出误差模型, 通过依次递增脉冲响应参数的数目N来提高逼近精度. 理论分析表明, 只要N足够大, 模型的辨识精度可以满足实际要求. 提出的辨识方法可以在假设阶次N =1的条件下, 依次递增计算N较大时的脉冲响应参数和目标函数值, 从而根据脉冲响应确定系统的参数. 仿真试验说明提出的方法估计输出误差模型的参数是有效的.
英文摘要
      Based on system impulse response, identification problem of the systems described by output error model was studied using correlation analysis, and its basic idea was to use a finite impulse response parameter model to approach the system model, and the model approximation precision was enhanced with the number N of the number of the impulse response parameters increasing. The theory analysis indicated that the model approximation could meet the requirements in Engineering as long as N is large enough. The proposed method could recursively compute the impulse response parameters and objective function by increasing N in turn from N=1, and estimated the system model parameters by the impulse response parameters. The simulation example shows that the proposed algorithm is effective for estimating the parameters of output error models.