System identification. An introduction. (Q533003): Difference between revisions

The book presents a systematic overview of the fundamental problems and methods in the modern system identification theory. The material is divided into four parts covering data based non-parametric identification methods, time-invariant system identification, time-varying system identification and model validation problems. In Part I (Chapters 2-4), an introduction to the time-domain and frequency-domain identification methods is presented and system response (Chapter 2), frequency response (Chapter 3) and correlation (Chapter 4) methods are reported. Part II (Chapters 5-6) deals with time-invariant system identification techniques. In Chapter 5 linear and nonlinear regression methods are discussed for static systems, including identifiability conditions, error-in-variables and bounded-noise problem characterizations. In Chapter 6, the identification of time-invariant, linear and nonlinear, dynamic systems is considered. Equation error, output error, and prediction error identification methods are presented for the linear dynamic systems, and completed with subspace identification and closed-loop identification tasks. Part III (Chapters 7-8) is devoted to the identification of time-varying systems. Identification of time-varying, linear and nonlinear, static systems is considered in Chapter 7, and recursive estimation, time-varying parameters and multioutput problems are discussed in the context of linear regression models. In turn, state-space representation end extended Kalman filter approach are considered for nonlinear systems. In Chapter 8 time-varying dynamic system identification is reported. Recursive least-squares and recursive prediction error estimation are presented for linear systems, and extended Kalman filtering and observer-based methods are described for nonlinear systems there. Part IV (Chapter 9) is concerned with model validation and model reduction problems. Each chapter of the book is finished with references, historical notes and exercises to be solved by the reader. Appendices A-F contain basic facts from matrix algebra, mathematical statistics, Laplace, Fourier and \(z\)-transforms, Bode diagrams, shift operator calculus and recursive least squares derivation. Numerous examples, also derived from real physical, chemical and biological systems, demonstrate the practical applicability of the presented methods. The book can be recommended for students and practitioners for self-study.