ESTIMATORS FOR HYBRID MULTIDIMENSIONAL CONTROL SYSTEM WITH VARIABLE STRUCTURE

TŮMA, L.; HANUŠ, B.

Abstract

Estimator makes it possible to do the control process optimal, giving on the one hand the information about the actual debalance of the controlled system, which is caused by the disturbances input and giving on the other hand the information about the controlled system actual properties. The domain of all possible controlled system properties is supposed to be identified or estimated at least. The controlled systems may be of arbitrary type and hybrid. The properties domain is determined by the net of nominal points with nominal properties of the system. The optimal estimators and optimal controllers are designed according to every nominal point ahead. The most suitable pair – estimator and controller – are switched on into the control operation. The choice is made in accordance to the estimator error. The linear discrete time incremental estimator is used for investigation. The controlled system properties in the nominal points are described by means of the state variables difference equations system in matrix form, by the difference equations system of the output variables or by mixed difference equations system of the output variables. The orders of those descriptions are not equal. The linear controlled system of the fourth order is tested by computer simulation in the distribution.

Coresponding author e-mail: libor[dot]tuma[at]vslib[dot]cz

Session: Robust and Adaptive Control