Region of convergence expansion for model predictive control of non-linear discrete time systems by linear matrix inequalities
In this thesis a gain scheduling method is proposed for robust model predictive control of nonlinear discrete-time systems. The system is composed of a linear model perturbed by an additive state-dependent nonlinear term. A robust model predictive controller is designed in the literature to compensate for the uncertainty of the system. In order to enlarge the region of convergence it is proposed that several equilibrium points are considered, and several robust controllers are designed. By switching between the controllers it is verified that the region of convergence may be enlarged, while the overall stability of the system is preserved. The stability analysis is based on Lyapunov functions for each of level sets, while state feedback control law are designed by minimization of a desired cost function formed in linear matrix inequalities.
|2014||M.Sc.||Dynamical Systems Analysis and Control|