The subject is aimed to explanation of basic identification methods to obtain mathematical description of deterministic and stochastic systems. Experimental identification methods are explained for linear stochastic and deterministic dynamic systems in greater detail. Analytic identification is applied for several examples and compared to experimental identification. During this course the mathematical background of system identification in both time domain and frequency domain is given. Lectures are concentrated to the most frequent methods which are applied in practice. It is expected prior experience with Matlab.

1. Fundamental terms of identification;

2. Analytic and experimental identification;

3. Model classification, parametric and nonparametric models;

4. Experimental identification with deterministic signals;

5. Parameterization of step, impulse and frequency characteristics;

6. Fundamental terms from probability and stochastic processes;

7. System identification in frequency domain and time domain;

8. Least squares methods;

9. Models ARMA, AR, MA,ARX, OE, ARMAX, BJ.

10. Prediction Error Method

11. Maximum Likelihood Method

12. Identification in closed loop

Solution of 3 projects which are aimed at
1) analytical identification

2) structure and parameter estimation of linear process through the use of deterministic input signal

3) experimental identification of a stochastic system

L.Ljung: System Identification - Theory for User, Prentice Hall PTR, 1999

D.E.Sebotg, T.F.Edgar, D.A.Mellichamp: Process Dynamics and Control, John Wiley & Sons,1989

Zhu Y.C.: Multivariable System Identification for Process Control, Pergamon, Oxford, 2001.

Söderström T., Stoica P.: System Identification. Prentice Hall International, London, 2001

Hsu H.P.: Probability, Random Variables nad Random Processes, McGraw-Hill, New York, 1996

Hofreiter M.: Identifikace systémů I, ČVUT v Praze, Praha, 2009

It is expected prior knowledge of basic control engineering and experience with Matlab.

System identification, linear nonparametric and parametric models, estimators, correlation analysis