Approximability models and optimal system identification

This article considers the problem of optimally recovering stable linear time-invariant systems observed via linear measurements made on their transfer functions. A common modeling assumption is replaced here by the related assumption that the transfer functions belong to a model set described by approximation capabilities. Capitalizing on recent optimal recovery results relative to such approximability models, we construct some optimal algorithms and characterize the optimal performance for the identification and evaluation of transfer functions in the framework of the Hardy Hilbert space and of the disk algebra. In particular, we determine explicitly the optimal recovery performance for frequency measurements taken at equispaced points on an inner circle or on the torus.

[1]  Jie Chen,et al.  Worst-case system identification in H∞: validation of apriori information, essentially optimal algorithms, and error bounds , 1992, 1992 American Control Conference.

[2]  R. DeVore,et al.  Data assimilation and sampling in Banach spaces , 2016, 1602.06342.

[3]  P. Erdös,et al.  Interpolation , 1953, An Introduction to Scientific, Symbolic, and Graphical Computation.

[4]  J. Doyle,et al.  Essentials of Robust Control , 1997 .

[5]  Erik Weyer,et al.  Finite sample properties of system identification methods , 2002, IEEE Trans. Autom. Control..

[6]  B. Pasik-Duncan.  Control-oriented system identification: An H∞ approach , 2002 .

[7]  Wolfgang Dahmen,et al.  Data Assimilation in Reduced Modeling , 2015, SIAM/ASA J. Uncertain. Quantification.

[8]  Benjamin Recht,et al.  Non-Asymptotic Analysis of Robust Control from Coarse-Grained Identification , 2017, ArXiv.

[9]  Parikshit Shah,et al.  Linear system identification via atomic norm regularization , 2012, 2012 IEEE 51st IEEE Conference on Decision and Control (CDC).

[10]  Guoxiang Gu,et al.  Identification in H∞ with nonuniformly spaced frequency response measurements , 1992, 1992 American Control Conference.

[11]  T. J. Rivlin,et al.  Optimal Estimation in Approximation Theory , 1977 .

[12]  József Bokor,et al.  Approximate H/sub /spl infin// identification using partial sum operators in disc algebra basis , 1995, Proceedings of 1995 American Control Conference - ACC'95.

[13]  Jie Chen,et al.  Worst case system identification in H∞: validation of a priori information, essentially optimal algorithms, and error bounds , 1995, IEEE Trans. Autom. Control..

[14]  Charles A. Micchelli,et al.  A Survey of Optimal Recovery , 1977 .

[15]  Mathukumalli Vidyasagar,et al.  A learning theory approach to system identification and stochastic adaptive control , 2008 .

[16]  C. Nett,et al.  The Caratheodory-Fejer problem and H/sub /spl infin// identification: a time domain approach , 1993, Proceedings of 32nd IEEE Conference on Decision and Control.

[17]  Carl N. Nett,et al.  Control oriented system identification: a worst-case/deterministic approach in H/sub infinity / , 1991 .

[18]  Boris Shekhtman,et al.  On discrete norms of polynomials , 2006, J. Approx. Theory.

[19]  Guoxiang Gu,et al.  Identification in ℋ∞ with nonuniformly spaced frequency response measurements , 1994 .

[20]  Ronald A. DeVore,et al.  Computing a Quantity of Interest from Observational Data , 2018, Constructive Approximation.

[21]  Jie Chen,et al.  Optimal Non-Parametric System Identification From Arbitrary Corrupt Finite Time Series: A Control-Oriented Approach , 1992, 1992 American Control Conference.

[22]  Jonathan R. Partington,et al.  Interpolation, identification, and sampling , 1997 .