Command governors for constrained linear systems under partial state information

A method is presented for designing command governors for discrete-time linear systems under partial state information and in the presence of unknown but bounded input and measurement disturbances. The solution is based on the set-membership state estimation paradigm and hinges upon the state uncertainty tube, viz. the set of all state trajectories compatible with the given measurements and initial state uncertainty. A convergence analysis is presented for both the case wherein the exact state uncertainty tube is used, and the computationally less demanding case wherein suitable outer approximations of the state uncertainty tube are exploited.