论文信息 - Comparison of Three Classes of Identification Methods

Comparison of Three Classes of Identification Methods

Abstract Three linear discrete-time identification methods are compared in this work. First, comparing criteria are proposed from an application viewpoint. Then the methods, namely, a subspace method, the prediction error method family and an asymptotic method are briefly introduced. All the methods are used to identify two multi variable industrial processes. The comparison is made using the criteria and based on the performances of the methods for the real life data. The purpose of this work is to make some contribution in unifying the field of identification and in bridging the gap between theory and application.

[1] Lennart Ljung,et al. Unprejudiced optimal open loop input design for identification of transfer functions , 1985, Autom..

[2] L. Ljung,et al. Asymptotic Properties of the Least Squares Method for Estimating Transfer Functions and Disturbance Spectra , 1992 .

[3] Bart De Moor,et al. N4SID: Subspace algorithms for the identification of combined deterministic-stochastic systems , 1994, Autom..

[4] Lennart Ljung,et al. System Identification: Theory for the User , 1987 .

[5] B. Wahlberg. Model reductions of high-order estimated models : the asymptotic ML approach , 1989 .

[6] Yucai Zhu. Black-box identification of mimo transfer functions: Asymptotic properties of prediction error models , 1989 .

[7] A.C.P.M. Backx,et al. Multivariable process identification based on frequency domain measures , 1991, [1991] Proceedings of the 30th IEEE Conference on Decision and Control.

[8] Lennart Ljung,et al. Optimal experiment designs with respect to the intended model application , 1986, Autom..

[9] Yucai Zhu,et al. Identification of Multivariable Industrial Processes: for Simulation, Diagnosis and Control , 1993 .

[10] L. Ljung. Asymptotic variance expressions for identified black-box transfer function models , 1984, The 23rd IEEE Conference on Decision and Control.