Multiprogram control for dynamic systems: a point of view

Abstract

In this paper, we analyze potentials and describe methods of synthesis of so-called multiprogram controls in different classes of dynamic systems. The main idea is to construct the control that provides predesigned finite set of asymptotically stable program motions for closed-loop system. Special attention is paid to modification of algorithms to extend possibilities of its application. For this aim, firstly, special classes of multiprogram controls are involved. Hybrid controls have continuous elements to realize program motions, and discrete stabilizers. Secondly, we analyze a case of incomplete feedback when some components of the phase state vector cannot be measured. In this case, the multiprogram control system is supplemented with asymptotic observer (state estimator), which also could be hybrid. Practically, the system closed with a multiprogram control is a nonlinear program automate that can realize any program motion of predesigned class depending on initial values. Developed theoretical base allows constructing of multiprogram control systems in cases of both complete and incomplete feedback when information in feedback channels is discrete. Research results are illustrated with an example of real technical application. The multiprogram control problem for providing of several modes of flywheel work in a car engine is considered.