EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Energy-Based Exploration for Reinforcement Learning of Underactuated Mechanical Systems

Jeroen Taets, Tom Lefebvre, Frederik Ostyn, Guillaume Crevecoeur

IEEE Access · 2025 · 13 · 98847-98859

DOI: 10.1109/access.2025.3576394 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2025. Verfasst von Jeroen Taets und 3 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Strategisch Basis Onderzoek (SBO) and CTRLxAI Project from Fonds Wetenschappelijk Onderzoek (FWO) Research Foundations Flanders.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Reinforcement Learning (RL) for underactuated mechanical systems presents unique challenges due to the limited control inputs and complex dynamics of the system. Efficient exploration poses a significant problem for two reasons: First, these systems must start each episode of learning from an equilibrium position instead of a random position, severely limiting exploration. Second, these systems are often chaotic, which limits model-based exploration. To address these challenges, we introduce Energy-based Exploration for Reinforcement Learning (EBERL), which exploits a model of the system’s energy represented by a Deep Lagrangian Network (DeLaN), coupled to an energy-based controller to ensure goal-directed exploration on top of the exploration principles that are already present in RL. First, we validated our method in simulation on a Cartpole, Furuta pendulum and Pendubot with challenging torque constraints that cannot be solved by traditional methods. Additionally, we experimentally validated our method on the swing-up and stabilization of a Quanser QUBE2 Furuta pendulum.

Förderung

  • Strategisch Basis Onderzoek (SBO) and CTRLxAI Project from Fonds Wetenschappelijk Onderzoek (FWO) Research Foundations Flanders

Zitieren

APA

Taets, J., Lefebvre, T., Ostyn, F., & Crevecoeur, G. (2025). Energy-Based Exploration for Reinforcement Learning of Underactuated Mechanical Systems. IEEE Access, 13, 98847-98859. https://doi.org/10.1109/access.2025.3576394

Vancouver

Taets J, Lefebvre T, Ostyn F, Crevecoeur G. Energy-Based Exploration for Reinforcement Learning of Underactuated Mechanical Systems. IEEE Access. 2025;13:98847-98859. doi:10.1109/access.2025.3576394

BibTeX
@article{Taets2025,
  author = {Taets, Jeroen and Lefebvre, Tom and Ostyn, Frederik and Crevecoeur, Guillaume},
  title = {Energy-Based Exploration for Reinforcement Learning of Underactuated Mechanical Systems},
  journal = {IEEE Access},
  year = {2025},
  volume = {13},
  pages = {98847--98859},
  doi = {10.1109/access.2025.3576394}
}
Kostenlos starten →