An extended MPC convergence condition

Abstract

Nominal convergence of Constrained Model Predictive Control has been extensively analyzed in the last fifteen years. The inclusion of a terminal constraint into the optimization problem and the expansion of the prediction horizon up to infinity are the main strategies already proposed in order to achieve the desired stability. However, when a model is used in which the inputs are in the incremental form, these strategies tend to be infeasible. This paper extends the contracting constraint idea by including a simple-to-apply and less restrictive new set of constraints into the optimization problem, to allow nominal convergence.