Constrained predictive control with guaranteed stability and convex optimization

Although predictive control has shown to be very successful, both in theory and practice, general stability can not be guaranteed. In this paper a new algorithm is proposed, which yields guaranteed stability. This is achieved by minimizing the used criterion function with respect to all stabilizing controllers, instead of the unstructured future controller outputs. The algorithm only needs restrictions which are already necessary in the classical approach to predictive controller design without constraints. Furthermore, if the classical approach yields a stable linear closed-loop system, the here presented method will give the same results.