SOFTLY SWITCHED ROBUSTLY FEASIBLE MPC FOR CONSTRAINED LINEAR SYSTEMS UNDER SET BOUNDED UNCERTAINTY - LQ-MPC WITH IC ACTION

摘要

An efficient plant control under state and output constraints under a wide range of operating conditions is not possible by one controller. Hence, the concept of multiple controllers being switched as required by operating conditions is an attractive option. Quite often, hard switching of the controllers may not be desired due to variety of reasons or not possible at all. The paper considers soft switching of the predictive controllers utilizing the invariant sets theory to constrained linear discrete time invariant system operating under bounded additive disturbances. The applied resulting softly switched predictive controller (SS MPC) with the integral action activates smooth and robustly feasible tracking of the plant output reference trajectory under wide range of operating conditions. The controller performance is illustrated by simulations.