Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (14): 54-56.

• 理论研究 • Previous Articles     Next Articles

Research of instrumental variable identification method for general models

LIU Shu-xia,HUANG Min   

  1. School of Communication and Control Engineering,Jiangnan University,Wuxi,Jiangsu 214122,China
  • Received:2007-08-28 Revised:2007-11-26 Online:2008-05-11 Published:2008-05-11
  • Contact: LIU Shu-xia


刘淑霞,黄 敏   

  1. 江南大学 通信与控制工程学院,江苏 无锡 214122
  • 通讯作者: 刘淑霞

Abstract: For General model system of existing correlated interferential noise,this paper researches on a new method of identification,first of all system model is approximated by using to FIR model,and to obtain a Box-Jenkins model which can be identified by the Instrumental Variable method,and finally to determine the parameters of the original systems by means of the model equivalence principle.The simulation results indicate that Recursive Instrumental Variable(RIV)method has better parameter estimation than Recursive Generalized Extended Least Squares(RGELS) in this approximation.

Key words: General model, recursive generalized extended least squares, Instrumental Variable, parameter estimation

摘要: 对于存在相关噪声干扰的General系统,研究了一种新的辨识方法。首先系统模型用一个有限的脉冲响应(FIR)模型逼近,得到一个BoxJenkins模型,再使用辅助变量法辨识系统参数,最后根据模型等价原理确定原系统的参数估计。仿真结果表明:在这种近似下递推辅助变量法(RIV)比递推广义增广最小二乘法(RGELS)可以得到更好的参数估计。

关键词: General模型, 递推广义增广最小二乘, 辅助变量法, 参数估计