Reinforcement Learning-Based Traditional Sports Agents for Personalized Health Training Optimization
International Journal of Healthcare Information Systems and Informatics · 29.07.2026 · 21 (1) · 1-19
DOI: 10.4018/ijhisi.417380 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Healthcare Information Systems and Informatics, erschienen am 29.07.2026. Verfasst von Shaojuan Wang. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 3.0).
Abstract (Originalsprache)
Wushu, a health-promoting exercise and an intangible cultural heritage, relies on teachers to correct movement deviations and provide personalized guidance. Traditional teaching relies on instructors' subjective judgment, which often struggles to adapt to complex movements and individual differences. This paper integrates pose estimation and reinforcement learning to construct a training agent that dynamically generates and optimizes training strategies. Using real-time pose data, action scores, and feedback, the agent forms a closed-loop optimization. Experiments show the agent significantly reduces movement deviations and improves training efficiency, with advantages in adaptability and stability. A personalized state space embedding pose deviation, rhythm, and feedback is validated across different skill levels. The findings provide a new path for digital health intervention of traditional exercises and a practical reference for personalized exercise prescription systems in rehabilitation and fall prevention.
Zitieren
APA
Wang, S. (2026). Reinforcement Learning-Based Traditional Sports Agents for Personalized Health Training Optimization. International Journal of Healthcare Information Systems and Informatics, 21(1), 1-19. https://doi.org/10.4018/ijhisi.417380
Vancouver
Wang S. Reinforcement Learning-Based Traditional Sports Agents for Personalized Health Training Optimization. International Journal of Healthcare Information Systems and Informatics. 2026;21(1):1-19. doi:10.4018/ijhisi.417380
BibTeX
@article{Wang2026,
author = {Wang, Shaojuan},
title = {Reinforcement Learning-Based Traditional Sports Agents for Personalized Health Training Optimization},
journal = {International Journal of Healthcare Information Systems and Informatics},
year = {2026},
volume = {21},
number = {1},
pages = {1--19},
doi = {10.4018/ijhisi.417380}
}
EUVIMED