Vector ARMA estimation: a reliable subspace approach

A parameter estimation method for finite-dimensional multivariate linear stochastic systems, which is guaranteed to produce valid models approximating the true underlying system in a computational time of a polynomial order in the system dimension, is presented. This is achieved by combining the main features of certain stochastic subspace identification techniques with sound matrix Schur restabilizing procedures and multivariate covariance fitting, both of which are formulated as linear matrix inequality problems. All aspects of the identification method are discussed, with an emphasis on the two issues mentioned above, and examples of the overall performance are provided for two different systems.

[1]  J. Fuchs,et al.  ARMA-Order estimation via matrix perturbation theory , 1986, 1986 25th IEEE Conference on Decision and Control.

[2]  J. Maciejowski.  Guaranteed stability with subspace methods , 1995 .

[3]  J. Fuchs.  Structure and order estimation of multivariable stochastic processes , 1990 .

[4]  Anders Lindquist,et al.  Canonical correlation analysis, approximate covariance extension, and identification of stationary time series , 1996, Autom..

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

[6]  Stephen P. Boyd,et al.  Semidefinite Programming , 1996, SIAM Rev..

[7]  Petre Stoica,et al.  MA estimation in polynomial time , 2000, IEEE Trans. Signal Process..

[8]  Randolph L. Moses,et al.  On the unit circle problem: The Schur-Cohn procedure revisited , 1992, Signal Process..

[9]  Petre Stoica,et al.  Introduction to spectral analysis , 1997 .

[10]  P. Caines.  Linear Stochastic Systems , 1988 .

[11]  Masanao Aoki,et al.  State Space Modeling of Time Series , 1987 .

[12]  W. Larimore.  System Identification, Reduced-Order Filtering and Modeling via Canonical Variate Analysis , 1983, 1983 American Control Conference.

[13]  P. Dooren.  Polynomial matrix factorizations via recursive pole-zero cancellation , 1990 .

[14]  R. Kálmán,et al.  On Invariants, Canonical Forms and Moduli for Linear, Constant, Finite Dimensional, Dynamical Systems , 1976 .

[15]  Petre Stoica,et al.  Decentralized Control , 2018, The Control Systems Handbook.

[16]  Henk Nijmeijer,et al.  System identification in communication with chaotic systems , 2000 .

[17]  B. Anderson.  The inverse problem of stationary covariance generation , 1969 .

[18]  P. Dooren.  A Generalized Eigenvalue Approach for Solving Riccati Equations , 1980 .

[19]  Stephen P. Boyd,et al.  Linear Matrix Inequalities in Systems and Control Theory , 1994 .

[20]  Anders Lindquist,et al.  Experimental evidence showing that stochastic subspace identification methods may fail 1 1 This rese , 1998 .

[21]  R. V. Patel,et al.  Computation of matrix fraction descriptions of linear time-invariant systems , 1981 .

[22]  Petre Stoica,et al.  Comparative Study of Rank Test Methods for ARMA Order Estimation , 2018, Statistical Methods in Control & Signal Processing.

[23]  M. Deistler,et al.  The Analysis of the Asymptotic Variance of Subspace Algorithms , 1997 .

[24]  J. Doob.  Stochastic processes , 1953 .

[25]  J. Willems,et al.  Stability theory of dynamical systems , 1970 .

[26]  J. Cadzow,et al.  High performance spectral estimation--A new ARMA method , 1980 .

[27]  Jan M. Maciejowski,et al.  Realization of stable models with subspace methods , 1996, Autom..

[28]  Gene H. Golub,et al.  Matrix computations , 1983 .

[29]  Thomas Kailath,et al.  Linear Systems , 1980 .

[30]  David Q. Mayne,et al.  Linear identification of ARMA processes , 1982, Autom..

[31]  C. Loan.  How Near is a Stable Matrix to an Unstable Matrix , 1984 .

[32]  W. E. Roth.  On direct product matrices , 1934 .

[33]  A. Laub,et al.  On the numerical solution of the discrete-time algebraic Riccati equation , 1980 .

[34]  R. Ober.  Balanced parametrization of classes of linear systems , 1991 .

[35]  Jerry M. Mendel,et al.  ARMA parameter estimation using only output cumulants , 1990, IEEE Trans. Acoust. Speech Signal Process..

[36]  Petre Stoica,et al.  Detection tests for array processing in unknown correlated noise fields , 1997, IEEE Trans. Signal Process..

[37]  Petre Stoica,et al.  A Semidefinite Programming Approach to ARMA Estimation , 2000 .

[38]  H. Akaike.  Canonical Correlation Analysis of Time Series and the Use of an Information Criterion , 1976 .

[39]  Jorge Mari,et al.  Modifications of rational transfer matrices to achieve positive realness , 2000, Signal Process..

[40]  P. Faurre,et al.  Identification par Minimisation d'une Representation Markovienne de Processus Aleatoire , 1970 .

[41]  Arkadi Nemirovski,et al.  The projective method for solving linear matrix inequalities , 1997, Math. Program..

[42]  B. Moore,et al.  Singular value analysis of linear systems , 1978, 1978 IEEE Conference on Decision and Control including the 17th Symposium on Adaptive Processes.

[43]  James Durbin,et al.  The fitting of time series models , 1960 .

[44]  E. Hannan,et al.  The statistical theory of linear systems , 1989 .

[45]  Tomas McKelvey,et al.  System identification using an over-parametrized model class-improving the optimization algorithm , 1997, Proceedings of the 36th IEEE Conference on Decision and Control.

[46]  Β. L. HO,et al.  Editorial: Effective construction of linear state-variable models from input/output functions , 1966 .

[47]  P. Van Overschee,et al.  Subspace algorithms for the stochastic identification problem , 1991 .

[48]  C. L. Nikias,et al.  Higher-order spectra analysis : a nonlinear signal processing framework , 1993 .

[49]  J. Willems,et al.  Parametrizations of linear dynamical systems: Canonical forms and identifiability , 1974 .