Journal of East China Normal University(Natural Sc ›› 2003, Vol. 2003 ›› Issue (1): 25-30.

• Article • Previous Articles     Next Articles

Identification of Laguerre Model and Design Variable for Time-varying Systems(2) ------Kalman filter algorithms

DING Zhao-hong1;SHA Quan1; YUAN Zhen-dong2   

  1. 1. Department of Automation, Shanghai College of Applied Technology, Shanghai 200233, China; 2. Department of Mathematics, East China Normal University, Shanghai 200062. China
  • Received:2001-06-12 Revised:2002-07-03 Online:2003-03-25 Published:2003-03-25

Abstract: It is supposed that the time-varying parameters included in the system are stationary AR(1) variable. The estimate of the mean square error (MSE) of transfer function for time-varying Laguerre model is discussed. The approximate expression of MSE for Kalman filter algorithms can be derived under following assumptions:the dynamic of the system is slowly changing, the adaptation is also quite slow and the order of model system is high enough. Using Laguerre model instead of FIR model,the MSE will be reduced and the order of Laguerre model is reduced as well. The optimization problems for design variables of time-varying system identification algorithms are discussed.

Key words: MSE, laguerre model, kalman filter algorithms, design variable, time-varying system, MSE, laguerre model, kalman filter algorithms, design variable