Long horizon input parameterisations to enlarge the region of attraction of MPC

In this paper, the efficacy of structured and unstructured parameterisations of the degree of freedom within a predictive control algorithm is investigated. While several earlier papers investigated the enlargement of the region of attraction using structured prediction dynamics, little consideration has been given to the potential of unstructured parameterisations to handle the trade-off between the region of attraction, performance and computational burden. This paper demonstrates how unstructured dynamics can be both selected and used effectively and furthermore gives a comparison with structured methods.