An indirect prediction error method for system identification

Abstract A new form of prediction error method (PEM) is developed. It is applicable to the case where the model structure of interest can be imbedded in a larger model structure whose estimation is relatively easy. An optimal way of reducing the larger model to the smaller model structure is presented and various interpretations of this reduction are given. The proposed method will have the same asymptotic statistical properties as the standard PEM but it can be implemented by a more efficient algorithm. The properties of the method are illustrated by the means of some simulated examples.

[1]  Benjamin Friedlander,et al.  Performance analysis of MA parameter estimation algorithms based on high-order moments , 1988, ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing.

[2]  Anthony J. Jakeman,et al.  On the decoupling of system and noise model parameter estimation in time-series analysis , 1981 .

[3]  T. Söderström.  Brief paper: Test of pole-zero cancellation in estimated models , 1975 .

[4]  T. Söderström.  On model structure testing in system identification , 1977 .

[5]  T. Söderström,et al.  On the parsimony principle , 1982 .

[6]  M. Gevers,et al.  Identification of linearly overparametrized nonlinear systems , 1992 .

[7]  B. Wahlberg.  On the Identification and Approximation of Linear Systems , 1987 .

[8]  Soederstroem.  ON THE CONVERGENCE PROPERTIES OF THE GENERALIZED LEAST SQUARES IDENTIFICATION METHOD. , 1972 .

[9]  Bo Wahlberg,et al.  Estimation Of Autoregressive Moving‐Average Models Via High‐Order Autoregressive Approximations , 1989 .

[10]  T. Söderström,et al.  On reparametrization of loss functions used in estimation and the invariance principle , 1989 .

[11]  Lennart Ljung,et al.  On The Consistency of Prediction Error Identification Methods , 1976 .

[12]  Bo Wahlberg,et al.  On Model Reduction in System Identification , 1986, 1986 American Control Conference.

[13]  B. Porat,et al.  ARMA parameter estimation based on sample covariances, for missing data , 1986, ICASSP '86. IEEE International Conference on Acoustics, Speech, and Signal Processing.

[14]  Petre Stoica,et al.  The parsimony principle for a class of model structures , 1985 .

[15]  Petre Stoica,et al.  Large-sample estimation of the AR parameters of an ARMA model , 1985 .

[16]  Soura Dasgupta,et al.  Adaptive identification of systems with polynomial parameterizations , 1988 .

[17]  T. Söderström.  Convergence properties of the generalised least squares identitication method , 1974, Autom..

[18]  L. Ljung,et al.  Identification of linear, multivariable systems operating under linear feedback control , 1974 .

[19]  Brian D. O. Anderson,et al.  Identification of physical parameters in structured systems , 1988, Autom..

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

[21]  Anthony J. Jakeman,et al.  Advanced methods of recursive time-series analysis , 1983 .

[22]  B. Friedlander,et al.  An approximate maximum likelihood approach to ARMA spectral estimation , 1985, 1985 24th IEEE Conference on Decision and Control.

[23]  Benjamin Friedlander,et al.  On the limiting behavior of estimates based on sample covariances , 1986, ICASSP '86. IEEE International Conference on Acoustics, Speech, and Signal Processing.