تقرير
A Fast Convergence Algorithm for Iterative Adaptation of Feedforward Controller Parameters
| العنوان: | A Fast Convergence Algorithm for Iterative Adaptation of Feedforward Controller Parameters |
|---|---|
| المؤلفون: | Serrano-Seco, Eloy, Moya-Lasheras, Eduardo, Ramirez-Laboreo, Edgar |
| سنة النشر: | 2024 |
| المجموعة: | Computer Science |
| مصطلحات موضوعية: | Electrical Engineering and Systems Science - Systems and Control |
| الوصف: | Feedforward control is a viable option for enhancing the response time and control accuracy of a wide variety of systems. Nevertheless, it is not able to compensate for the effects produced by modeling errors or disturbances. A solution to improve the feedforward performance is the use of an adaptation law that modifies the parameters of the feedforward control. In the case where real-time feedback is not possible, a solution is a run-to-run numerical optimization method that is fed with a cost based on a measured signal. Although the effectiveness of this approach has been demonstrated, its performance is hindered by slow convergence. In this paper, we present an algorithm based on Pattern Search and Adaptive Coordinate Descent methods that makes use of the sensitivity of the feedforward controller to its parameters so that the convergence speed improves significantly. Like many algorithms, this is a local strategy so the algorithm might converge to a local minimum. Therefore, we present two versions, one without a learning rate and one with it. To compare them and to demonstrate the effectiveness of the algorithm, simulated results are shown on a well-known control problem in electromechanics: the soft-landing control of electromechanical switching devices. Comment: 6 pages, 7 figures. Version submitted to the 63rd IEEE Conference on Decision and Control (CDC) |
| نوع الوثيقة: | Working Paper |
| URL الوصول: | http://arxiv.org/abs/2404.00036 |
| رقم الأكسشن: | edsarx.2404.00036 |
| قاعدة البيانات: | arXiv |
| الوصف غير متاح. |