A design method for a robust LQ optimal model-following controller: with application to MuPAL-α and verification by flight experiments

摘要

This paper proposes a design method for a model-following controller that is globally optimized for robust LQ performance; that is, model-following errors between plant and model outputs against uncertainties and perturbations of the plant are globally optimized. First, the robust LQ performance for an LPV (Linear Parameter-Varying) system is defined, and a design method for a robust LQ optimal regulator (hereinafter referred to as the "robust LQR problem") is presented. Next, the robust LQR problem is applied to a model-following controller design. Model-following errors with uncertainties and perturbations of a plant are defined as robust LQ performance, and a design method for a robust LQ optimal model-following controller is presented using the robust LQR problem. Finally, to validate the proposed method, two model-following controllers for the lateral/directional motions of an airplane are designed and the results of flight experiments are presented.