Design of Optimal Generalized Predictive Control System

Abstract

We propose a new design method for the generalized predictive control (GPC) system which realizes the optimal tracking performance. The objective is to design the control system which has the similar feedback properties to the two-degree-of-freedom optimal type-1 servo system. The design procedure consists of two steps. First, we design the optimal GPC system without integrator which minimizes the cost index. Next, we incorporate the integrator into the control system which acts only the case of the existences of a disturbance or a plant perturbation. A numerical example and an application are given to illustrate the effectiveness of the proposed method by comparing with the conventional method.