Szczegóły publikacji
Opis bibliograficzny
Comparative study of reinforcement learning-enhanced control strategies for nonlinear multidimensional systems / Patryk BAŁAZY // W: IEEE MetroXRAINE 2025 [Dokument elektroniczny] : 2025 IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering : 22-24 October 2025, Ancona, Italy : conference proceedings. — Wersja do Windows. — Dane tekstowe. — Piscataway : IEEE, cop. 2025. — Dod. ISBN: 979-8-3315-0280-5 (print on demand). — e-ISBN: 979-8-3315-0279-9. — S. 693–698. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. 698, Abstr. — Publikacja dostępna online od: 2026-01-23
Autor
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 165878 |
|---|---|
| Data dodania do BaDAP | 2026-03-06 |
| Tekst źródłowy | URL |
| DOI | 10.1109/MetroXRAINE66377.2025.11340573 |
| Rok publikacji | 2025 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Institute of Electrical and Electronics Engineers (IEEE) |
Abstract
This paper investigates the application of reinforcement learning (RL) in the control of nonlinear multidimensional systems, specifically focusing on an overhead crane as the test object. Three distinct RL-based control strategies are compared: direct control using RL agent actions, feedback regulator enhancement through RL agent actions, and auxiliary control signals provided by RL agent actions to the feedback regulator. Each control system was initially trained on a mathematical model of the crane and subsequently fine-tuned on the actual physical system. The direct control approach leverages the RL agent's actions as immediate control inputs, aiming for rapid adaptation to dynamic changes. The feedback enhancement strategy integrates RL actions to refine the feedback loop, improving stability and response time. The auxiliary control method uses RL actions to support the primary feedback regulator, enhancing overall system robustness. Experimental results demonstrate the effectiveness of each approach, with detailed performance metrics highlighting their respective advantages and limitations. This study provides valuable insights into the practical implementation of RL-based control systems for complex, real-world applications.
