Direct Data-Driven Feedforward Controller Design Based on Sparse Modelling Approach for Two-Degree-of-Freedom Control System

Estimated Response Iterative Tuning (ERIT) is a data-driven tuning method for two-degree-of-freedom control systems. ERIT can update the feedforward controller using the initial output signal and the initial controllers. This feature offers a significant advantage, as it can greatly reduce the cost and time required for the design or tuning of the controller. When the desired orders of the numerator and denominator of the feedforward controller are unknown, the controller should be configured as FIR filters. Consequently, this study introduces a sparse modeling approach based on ERIT and Lasso. The proposed method can eliminate unnecessary terms in the feedforward controller using Lasso. In this study, we evaluate the control performance of the proposed methods through experimental verification.