Synthesizing on-line constrained robust model predictive control based-on nominal performance cost
Baocang Ding,Peng Yang,Hexu Sun,Li, Shaoyuan
DOI: https://doi.org/10.1109/WCICA.2004.1340650
2004-01-01
Abstract:A new method to address constrained robust model predictive control for systems with polytopic description is proposed. Inheriting from an earlier method, we split the infinite-horizon control moves into a set of free control moves over a fixed horizon and a state feedback law in the terminal region, all on-line optimized. Then, the basic idea is to utilize, at each sampling instant, a fixed number of free control moves to drive (implicitly) the terminal state into a terminal constraint set changing with time. However, a nominal, instead of worst-case, infinite horizon performance cost is adopted. This can reduce the computational burden and improve feasibility and optimality. Simulation result illustrates the advantages of the new technique.