东北大学学报(自然科学版) ›› 2010, Vol. 31 ›› Issue (4): 469-472+476.DOI: -

• 论著 • 上一篇    下一篇

含有过程噪声的Hammerstein-Wiener模型辨识算法及其收敛性分析

李妍;毛志忠;王琰;袁平;   

  1. 东北大学信息科学与工程学院;
  • 收稿日期:2013-06-20 修回日期:2013-06-20 出版日期:2010-04-15 发布日期:2013-06-20
  • 通讯作者: -
  • 作者简介:-
  • 基金资助:
    国家高技术研究发展计划项目(2007AA04Z194,2007AA041401)

Identification algorithm of Hammerstein-Wiener model with process noise and its convergence analysis

Li, Yan (1); Mao, Zhi-Zhong (1); Wang, Yan (1); Yuan, Ping (1)   

  1. (1) School of Information Science and Engineering, Northeastern University, Shenyang 110004, China
  • Received:2013-06-20 Revised:2013-06-20 Online:2010-04-15 Published:2013-06-20
  • Contact: Li, Y.
  • About author:-
  • Supported by:
    -

摘要: 针对含有过程噪声的Hammerstein-Wiener模型,提出一种偏差补偿递推最小二乘辨识方法.通过将偏差补偿引入到递推最小二乘算法中,在线辨识包含原系统参数乘积项的参数向量.并用鞅收敛定理证明偏差补偿递推最小二乘辨识算法的收敛性,分析表明在持续激励的条件下参数估计偏差一致收敛于零.仿真结果表明该方法优于递推最小二乘辨识方法.

关键词: Hammerstein-Wiener模型, 偏差补偿递推最小二乘法, 鞅收敛定理, 收敛性, 参数辨识

Abstract: A bias-compensated recursive least squares (BCRLS) algorithm was proposed to identify the Hammerstein-Wiener model with process noise, i.e., the bias compensation was introduced into the recursive least squares algorithm. The parameter vector containing the product of the original system parameters was thus identified on-line. Then, the convergence of the algorithm was proved by the martingale convergence theorem. It was found that the estimation error of parameter product uniformly converges to zero during persistent excitation. Simulation results showed that the proposed method is better than the identification method by recursive least squares.

中图分类号: