Hybrid MPC for Minimum Phase Nonlinear Plants

The optimal constrained problem resulting from the MPC formulationfor nonlinear plants is commonly solved either by some gain scheduling linear approach, which require rather short sample times, or by computationally intersive numerical methods. In this paper the use of a hybrid structure is proposed, consisting of a nonlinear controller and of a standard linear predictive controller. The predictive controller acts on the system resulting from the original plant and the nonlinear controller. Conditions for this structure to yield better performance than the standard gain scheduling approach are derived.