Uniting local and global controllers

Abstract

A controller structure is proposed that dynamically combines predesigned local (e.g., linear) and global set-point stabilizing controllers. The local controller is only modified when the system to be controlled enters a region where the local controller loses its effectiveness. Thus, local performance and global stability are combined. Control objectives other than set-point tracking and the issue of robustness are discussed. The methodology is applied to achieve local performance and global stability for a class of fully-actuated Euler-Lagrange systems with input saturation.