Time-varying parametric system multiresolution identification by wavelets

In this paper, the problem of time-varying parametric system identification by wavelets is discussed. Employing wavelet operator matrix representation, we propose a new multiresolution least squares (MLS) algorithm for time-varying AR (ARX) system identification and a multiresolution least mean squares (MLMS) algorithm for the refinement of parameter estimation. These techniques can achieve the optimal tradeoff between the over-fitted solution and the poorly represented identification. The main features of time-varying model parameters are extracted in a multiresolution way, which can be used to represent the smooth trends as well as track the rapidly changing components of time-varying parameters simultaneously and adaptively. Further, a noisy time-varying AR (ARX) model can also be identified by combining the total least squares algorithm with the MLS algorithm. Based on the proposed AR (ARX) model parameter estimation algorithm, a novel identification scheme for time-varying ARMA (ARMAX) system is presented. A higher-order time-varying AR (ARX) model is used to approximate the time-varying ARMA (ARMAX) system and thus obtain an initial parameter estimation. Then an iterative algorithm is applied to obtain the consistent and efficient estimates of the ARMA (ARMAX) system parameters. This ARMA (ARMAX) identification algorithm requires linear operations only and thus greatly saves the computational load. In order to determine the time-varying model order, some modified AIC and MDL criterions are developed based on the proposed wavelet identification schemes. Simulation results verify that our methods can track the rapidly changing of time-varying system parameters and attain the best balance between parsimonious modelling and accurate identification.

[1]  Peter C. Young,et al.  Recursive Estimation and Time-Series Analysis: An Introduction , 1984 .

[2]  Peter J. W. Rayner,et al.  Generalized feature extraction for time-varying autoregressive models , 1996, IEEE Trans. Signal Process..

[3]  D. G. Lainiotis,et al.  Identification of Autoregressive Moving-Average Parameters of Time Series , 1975 .

[4]  Spilios D. Fassois,et al.  A polynomial-algebraic method for non-stationary TARMA signal analysis - Part I: The method , 1998, Signal Process..

[5]  Alan Oppenheim,et al.  Time-varying parametric modeling of speech , 1977, 1977 IEEE Conference on Decision and Control including the 16th Symposium on Adaptive Processes and A Special Symposium on Fuzzy Set Theory and Applications.

[6]  Bo Wahlberg,et al.  A Wavelet Shrinkage Approach for Frequency Response Estimation , 1994 .

[7]  Yves Grenier,et al.  Time-dependent ARMA modeling of nonstationary signals , 1983 .

[8]  Ingrid Daubechies,et al.  Ten Lectures on Wavelets , 1992 .

[9]  G. Kitagawa,et al.  A smoothness priors time-varying AR coefficient modeling of nonstationary covariance time series , 1985, IEEE Transactions on Automatic Control.

[10]  Georgios B. Giannakis,et al.  Time-varying system identification and model validation using wavelets , 1993, IEEE Trans. Signal Process..

[11]  Steven Kay,et al.  Modern Spectral Estimation: Theory and Application , 1988 .

[12]  J. Rissanen,et al.  Modeling By Shortest Data Description* , 1978, Autom..

[13]  Pablo A. Iglesias,et al.  On the poles and zeros of linear, time-varying systems , 2001 .

[14]  E. J. Hannan,et al.  Time series in the time domain , 1986 .

[15]  G. Kitagawa,et al.  A Smoothness Priors–State Space Modeling of Time Series with Trend and Seasonality , 1984 .

[16]  E. J. Hannan,et al.  Linear estimation of ARMA processes , 1983, Autom..

[17]  Jacques Kiseta Sabiti.  A fast estimation method for ARMA processes , 1996, Autom..

[18]  H. Akaike.  A new look at the statistical model identification , 1974 .

[19]  Stéphane Mallat,et al.  A Theory for Multiresolution Signal Decomposition: The Wavelet Representation , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[20]  Michel Barlaud,et al.  On AR Modeling of Nonstationary Signals , 1988 .

[21]  H. Morikawa,et al.  Adaptive estimation of time-varying model order in the ARMA speech analysis , 1990, IEEE Trans. Acoust. Speech Signal Process..

[22]  J. Cadzow,et al.  Spectral estimation: An overdetermined rational model equation approach , 1982, Proceedings of the IEEE.

[23]  E. Parzen.  Some recent advances in time series modeling , 1974 .