Enhancement of Tracking and Disturbance Suppression Performance via Adaptive Parameter Tuning of Two-degree-of-freedom Control System Using On-line FRIT

Abstract

In general, the characteristics of the control object changes over time due to aging, breakdowns, and changes in operating conditions and environment. In such cases, tracking and disturbance suppression performance of the entire closed-loop system may deteriorate. Based on a data-driven approach (online FRIT), we propose a method for simultaneously and sequentially tuning the control parameters of the feedforward and feedback controller of a two-degree-of-freedom control system in real time. We set two reference models as indicators of control performance: a target response model and a sensitivity characteristic model. The adaptive computation mechanism of the proposed method is based on the RLS (recursive least-squares) algorithm with a forgetting factor which can appropriately reflect changes in the characteristics of the controlled object. The proposed method enhances target tracking and disturbance suppression performance for time-varying systems without control system shutdown in real time. Finally, the effectiveness of the proposed method is demonstrated through simulation on benchmark process system.