Title :
Control period selection: Verification on coupled tanks
Author :
Berec, Ludek ; Rojicek, Jiri
Author_Institution :
Inst. of Inf. Theor. & Autom., Prague, Czech Republic
Abstract :
Model structure identification of a particular `generalized´ ARX model form, leading to the systematic data-based procedure of control period selection, is performed, using the Bayesian identification paradigm. The usual high dimensionality of the task is counteracted by an efficient algorithm. The functionality of both the procedure and the algorithm is tested on real data generated by the physical system of coupled tanks. The results are verified by-proper control experiments.
Keywords :
Bayes methods; autoregressive processes; optimal control; tanks (containers); Bayesian identification paradigm; control period selection; coupled tanks; generalized ARX model form; model structure identification; optimal control period; physical system; systematic data-based procedure; Computational modeling; Data models; Indexes; Optimal control; Periodic structures; Predictive models; ´generalized´ ARX model; Bayesian model structure identification; control period selection; coupled tanks physical system;
Conference_Titel :
Control Conference (ECC), 1997 European
Conference_Location :
Brussels
Print_ISBN :
978-3-9524269-0-6
