Data-based tuning of distinct inverse models used in feedforward and disturbance observer

This work proposes a data-based tuning method which is capable of tuning parameters of inverse models used in the feedforward controller and disturbance observer (DOB). Specifically, it aims to enhance the tracking performance of the positioning system by implementing different inverse models for feedforward controller and DOB, allowing a data-based tuning method to find two sets of optimal parameters for both inverse models simultaneously. Real-time operational data are utilized to achieve the optimization of inverse model’s parameters. Three experiments are required for the completion of single iteration in the presence of disturbances and noise. Compared to the tuning of uniform inverse model used in feedforward controller and DOB, the tuning of distinct inverse models for feedforward controller and DOB results in enhanced tracking performance. The effectiveness of the proposed method is verified through simulations.