The effects of the estmated plant models accuracy on the control system signals quality after generalized predictive controllers design are studied in the paper. Two identification approaches are used for different in structures discrete-time models – by optimization procedure in Optimization Toolbox based on the plant step response as a standard deterministic plant characteristic and by functions in System Identification Toolbox after experiments with random signals on the plant. The generalized predictive controllers are design according to the estimated models. The processes in the simulated control systems are analyzed concerning the effects of different kind plant models on the designed controllers.
1. Clarke, D. W., C. Mohtadi, and P. S. Tuffs. Generalized Predictive Control. – Automatica, 23, 1987, 137-160.
2. Garipov, E. Identification Systems – Part 1 and 2 (Third Edition). TU Sofia, 2007 (in Bulgarian).
3. MathWorks. MATLAB Optimization Toolbox, 2011.
4. MathWorks. MATLAB System Identification T., 2011.
5. Bao-Cang, Ding. Modern Predictive Control. Taylor and Francis Group, 2010.
6. Normey-Rico, J. E., E. F. Camacho. Control of Dead-time Process, Springer, 2006.
7. Garipov, E., V. Ruykova, T. Peltekov. Generalized Predictive Control with Varying Degrees of Uncertainty in the Plant Model. FA’2014, Sozopol, 2014 (in Bulgarian).
8. Garipov, E., V. Ruykova, T. Peltekov. Appropriate Values Selection for Adjustable Factors in the Generalized Predictive Controller Design. SAI’2014, Sofia, 2014 (in Bulgarian).