The Theory of Trajectory Tubes for the Problem of Output Feedback Tracking Control

Abstract

This paper presents a solution scheme for the problem of tracking control under on-line limited information and possible state constraints for a partially observed motion of a linear system. The control strategy is based on available observations corrupted by unknown but hard-bounded noise. The given solutions are based on the widely known set-membership ("bounding") approach to such models. They exploit the techniques of the developed theory of trajectory tubes (see Kurzhanski and Varaiya (2014)) and include appropriate computational procedures for set-valued functions. This enables to produce a complete solution scheme"from theory to computation".