FEASIBLE MODEL PREDICTIVE CONTROL WITH BOUNDED DISTURBANCES

This paper shows how to calculate feasible regions, parameterized in terms for the present state xk, for MPC controllers for constrained linear systems. The dependence of the feasible region on the prediction horizon is also made clear. It is also shown how the procedure may be modified to find guaranteed feasible regions in the presence of unknown, bounded disturbances. These ‘robust’ feasible regions are used to propose a very simple MPC controller which achieves robust feasibility.