Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : May 29, 2024 - June 01, 2024
In actual controlled objects, it may not be possible to obtain rapidly changing responses, open loop responses, or conduct long-term experiments. In such cases, it is difficult to obtain an accurate model through system identification. Therefore, Data driven control, which adjusts the controller directly from experimental data without requiring a model, has been studied in recent years. Among them, V-Tiger predicts overshoot and settling time based on frequency components of experimental data, and repeatedly evaluates them and redesigns control gains. However, V-Tiger has the problem that response prediction performance deteriorates when experimental data contains noise. Therefore, in this study, we propose a noise countermeasure that combines Disturbance observer that can predict disturbances, and Sparse modeling that can reduce non-zero parameters.
