The Frisch scheme in algebraic and dynamic identification problems

Guidorzi, Roberto   ; Diversi, Roberto   ; Soverini, Umberto

Kybernetika, Tome 44 (2008) no. 5, pp. 585-616

Abstract (VO)

This paper considers the problem of determining linear relations from data affected by additive noise in the context of the Frisch scheme. The loci of solutions of the Frisch scheme and their properties are first described in the algebraic case. In this context two main problems are analyzed: the evaluation of the maximal number of linear relations compatible with data affected by errors and the determination of the linear relation actually linking the noiseless data. Subsequently the extension of the Frisch scheme to the identification of dynamical systems is considered for both SISO and MIMO cases and the problem of its application to real processes is investigated. For this purpose suitable identification criteria and model parametrizations are described. Finally two classical identification problems are mapped into the Frisch scheme, the blind identification of FIR channels and the identification of AR + noise models. This allows some theoretical and practical extensions.

Détail

Classification : 93C05, 93E12
Keywords: system identification; errors-in-variables models; Frisch scheme; linear systems

Bibliographie

[1] Abed-Meraim K., Qiu, W., Hua Y.: Blind system identification. Proc. IEEE 85 (1997), 1310–1322

[2] Anderson B. D. O., Deistler M.: Identifiability of dynamic errors-in-variables models. J. Time Ser. Anal. 5 (1984), 1–13 | MR

[3] Anderson B. D. O., Deistler, M., Scherrer W.: Solution set properties for static errors-in-variables problems. Automatica 32 (1996), 1031–1035 | MR | Zbl

[4] Beghelli S., Castaldi P., Guidorzi, R., Soverini U.: A robust criterion for model selection in identification from noisy data. In: Proc. 9th International Conference on Systems Engineering, Las Vegas 1993, pp. 480–484

[5] Beghelli S., Guidorzi, R., Soverini U.: The Frisch scheme in dynamic system identification. Automatica 26 (1990), 171–176 | MR | Zbl

[6] Bobillet W., Grivel E., Guidorzi, R., Najim M.: Noisy speech de-reverberation as a SIMO system identification issue. In: Proc. IEEE Workshop on Statistical Signal Processing, Bordeaux 2005

[7] Bobillet W., Diversi R., Grivel E., Guidorzi R., Najim, M., Soverini U.: Speech enhancement combining optimal smoothing and errors-in-variables identification of noisy AR processes. IEEE Trans. Signal Process. 55 (2007), 5564–5578 | MR

[8] Deistler M.: Linear errors-in-variables models. In: Time Series and Linear Systems (Lecture Notes in Control and Information Sciences; S. Bittanti, ed.). Springer–Verlag, Berlin 1986, pp. 37–67 | MR

[9] Diversi R., Guidorzi, R., Soverini U.: A new criterion in EIV identification and filtering applications. In: Preprints 13th IFAC Symposium on System Identification, Rotterdam 2003, pp. 1993–1998

[10] Diversi R., Guidorzi, R., Soverini U.: Frisch scheme-based algorithms for EIV identification. In: Proc. 12th IEEE Mediterranean Conference on Control and Automation, Kusadasi 2004

[11] Diversi R., Guidorzi, R., Soverini U.: Blind identification and equalization of two-channel FIR systems in unbalanced noise environments. Signal Process. 85 (2005), 215–225 | Zbl

[12] Diversi R., Guidorzi, R., Soverini U.: A noise-compensated estimation scheme for AR processes. In: Proc. 44th IEEE Conference on Decision and Control and 8th European Control Conference, Seville 2005, pp. 4146–4151

[13] Diversi R., Guidorzi, R., Soverini U.: Yule–Walker equations in the Frisch scheme solution of errors-in-variables identification problems. In: Proc. 17th International Symposium on Mathematical Theory of Networks and Systems, Kyoto 2006, pp. 391–395

[14] Diversi R., Guidorzi, R., Soverini U.: Identification of autoregressive models in the presence of additive noise. International J. Adaptive Control and Signal Process. 22 (2008), 465–481 | MR

[15] Diversi R., Soverini, U., Guidorzi R.: A new estimation approach for AR models in presence of noise. In: Preprints 16th IFAC World Congress, Prague 2005 | MR

[16] Fernando K. V., Nicholson H.: Identification of linear systems with input and output noise: the Koopmans–Levin method. IEE Proc. 132 (1985), 30–36 | Zbl

[17] Frisch R.: Statistical Confluence Analysis by Means of Complete Regression Systems. Economic Institute, Pub. No. 5, Oslo University 1934 | Zbl

[18] Guidorzi R.: Equivalence, invariance and dynamical system canonical modelling. Part I, Kybernetika 25 (1989), 233–257, Part II, Kybernetika 25 (1989), 386–407 | Zbl

[19] Guidorzi R.: Certain models from uncertain data: the algebraic case. Systems Control Lett. 17 (1991), 415–424 | MR | Zbl

[20] Guidorzi R.: Errors-in-variables identification and model uniqueness. In: Statistical Modelling and Latent Variables (K. Haagen, D. J. Bartholomew, and M. Deistler, eds.), North Holland, Amsterdam 1993, pp. 127–150 | MR

[21] Guidorzi R.: Identification of the maximal number of linear relations from noisy data. Systems Control Lett. 24 (1995), 159–166 | MR | Zbl

[22] Guidorzi R.: Identification of multivariable processes in the Frisch scheme context. MTNS’96, St. Louis 1996

[23] Guidorzi R., Diversi R.: Determination of linear relations from real data in the Frisch scheme context. In: Proc. 17th International Symposium on Mathematical Theory of Networks and Systems, Kyoto 2006, pp. 530–535

[24] Guidorzi R., Diversi, R., Soverini U.: Blind identification and equalization of multichannel FIR systems in unbalanced noise environments. Signal Process. 87 (2007), 654–664 | Zbl

[25] Guidorzi R., Diversi R., Soverini, U., Valentini A.: A noise signature approach to fault detection and isolation. In: Proc. 16th International Symposium on Mathematical Theory of Networks and Systems, Leuven 2004

[26] Guidorzi R., Pierantoni M.: A new parametrization of Frisch scheme solutions. In: Proc. XII International Conference on Systems Science, Wroclaw 1995, pp. 114–120

[27] Guidorzi R., Soverini, U., Diversi R.: Multivariable EIV identification. In: Proc. 10th IEEE Mediterranean Conference on Control and Automation, Lisboa 2002

[28] Guidorzi R., Stoian A.: On the computation of the maximal corank of a covariance matrix under the Frisch scheme. In: Proc. 10th IFAC Symposium on System Identification, Copenhagen 1994, pp. 171–173

[29] Kalman R. E.: Identification from real data. In: Current Developments in the Interface: Economics, Econometrics, Mathematics (M. Hazewinkel, H. G. Rinnooy Kan, and D. Reidel, eds.), Dordrecht 1982, pp. 161–196

[30] Kalman R. E.: Nine Lectures on Identification. (Lecture Notes on Economics and Mathematical Systems.) Springer–Verlag, Berlin (to appear)

[31] Kalman R. E.: System identification from noisy data. In: Dynamical Systems II (A. R. Bednarek and L. Cesari, eds.), Academic Press 1982, pp. 135–164 | MR

[32] Kay S. M.: The effects of noise on the autoregressive spectral estimator. IEEE Trans. Acoustics, Speech and Signal Process. 27 (1979), 478–485 | Zbl

[33] Kay S. M.: Noise compensation for autoregressive spectral estimates. IEEE Trans. Acoustics, Speech and Signal Process. 28 (1980), 292–303 | Zbl

[34] Levin M. J.: Estimation of a system pulse transfer function in the presence of noise. IEEE Trans. Automat. Control 9 (1964), 229–235

[35] Malinvaud E.: Méthodes statistiques de l’économétrie. Third edition. Dunod, Paris 1980 | Zbl

[36] Schachermayer W., Deistler M.: The set of observationally equivalent errors-in-variables models. Systems Control Lett. 34 (1998), 101–104 | MR | Zbl

[37] Söderström T.: Errors-in-variables methods in system identification. Automatica 43 (2007), 939–958 | MR | Zbl

[38] Söderström T.: Accuracy analysis of the Frisch estimates for identifying errors-in-variables systems. IEEE Trans. Automat. Control 52 (2007), 985–997 | MR

[39] Stoica P., Nehorai A.: On the uniqueness of prediction error models for systems with noisy input-output data. Automatica 23 (1987), 541–543 | Zbl

[40] Tong L., Perreau S.: Multichannel blind identification: from subspace to maximum likelihood methods. Proc. IEEE 86 (1998), 1951–1968

[41] (ed.) S. Van Huffel: Recent Advances in Total Least Squares Techniques and Errors-in-Variables Modelling. SIAM, Philadelphia 1997 | MR

[42] Huffel S. Van, (eds.) P. Lemmerling: Total Least Squares and Errors-in-Variables Modelling: Analysis, Algorithms and Applications. Kluwer Academic Publishers, Dordrecht 2002 | MR

[43] Woodgate K. G.: An upper bound on the number of linear relations identified from noisy data by the Frisch scheme. Systems Control Lett. 24 (1995), 153–158 | MR | Zbl