EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A safe-enhanced fully closed-loop artificial pancreas controller based on deep reinforcement learning

Yan Feng Zhao, Jun Kit Chaw iD, Mei Choo Ang iD, Yiqi Tew, Xiao Yang Shi, Lin Liu, Xiang Cheng

PLOS ONE · 27.01.2025 · 20 (1) · e0317662

DOI: 10.1371/journal.pone.0317662 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS ONE, online erschienen am 27.01.2025. Verfasst von Yan Feng Zhao und 6 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Visual Informatics, The National University of Malaysia (UKM), Bangi, Malaysia. Hauptthemen laut MEDLINE: Pancreas, Artificial, Diabetes Mellitus, Type 1, Deep Learning. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Universiti Kebangsaan Malaysia.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Patients with type 1 diabetes and their physicians have long desired a fully closed-loop artificial pancreas (AP) system that can alleviate the burden of blood glucose regulation. Although deep reinforcement learning (DRL) methods theoretically enable adaptive insulin dosing control, they face numerous challenges, including safety and training efficiency, which have hindered their clinical application. This paper proposes a safe and efficient adaptive insulin delivery controller based on DRL. It employed ten tricks to enhance the proximal policy optimization (PPO) algorithm, improving training efficiency. Additionally, a dual safety mechanism of ’proactive guidance + reactive correction’ was introduced to reduce the risks of hyperglycemia and hypoglycemia and to prevent emergencies. Performance evaluations in the Simglucose simulator demonstrate that the proposed controller achieved an 87.45% time in range (TIR) median, superior to baseline methods, with a lower incidence of hypoglycemia, notably eliminating severe hypoglycemia and treatment failures. These encouraging results indicate that the DRL-based fully closed-loop AP controller has taken an essential step toward clinical implementation.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Pancreas, ArtificialDiabetes Mellitus, Type 1Deep LearningHumansInsulinBlood GlucoseHypoglycemiaAlgorithmsInsulin Infusion SystemsHypoglycemic AgentsReinforcement Machine Learning

Substanzen

InsulinBlood GlucoseHypoglycemic Agents

Autor:innen und Institutionen

  • Institute of Visual Informatics, The National University of Malaysia (UKM), Bangi, Malaysia.
  • Faculty of Computing and Information Technology, Tunku Abdul Rahman University of Management and Technology, Kuala Lumpur, Malaysia.
  • Department of Endocrinology, Zhengzhou University People's Hospital, Zheng Zhou, China.
  • College of Information Engineering, Henan Vocational University of Science and Technology, Zhou Kou, China.

ORCID-Profile: Jun Kit Chaw, Mei Choo Ang

Förderung

  • Universiti Kebangsaan Malaysia

Zitieren

APA

Zhao, Y. F., Chaw, J. K., Ang, M. C., Tew, Y., Shi, X. Y., Liu, L., & Cheng, X. (2025). A safe-enhanced fully closed-loop artificial pancreas controller based on deep reinforcement learning. PLOS ONE, 20(1), e0317662. https://doi.org/10.1371/journal.pone.0317662

Vancouver

Zhao YF, Chaw JK, Ang MC, Tew Y, Shi XY, Liu L, et al. A safe-enhanced fully closed-loop artificial pancreas controller based on deep reinforcement learning. PLOS ONE. 2025;20(1):e0317662. doi:10.1371/journal.pone.0317662

BibTeX
@article{Zhao2025,
  author = {Zhao, Yan Feng and Chaw, Jun Kit and Ang, Mei Choo and Tew, Yiqi and Shi, Xiao Yang and Liu, Lin and Cheng, Xiang},
  title = {A safe-enhanced fully closed-loop artificial pancreas controller based on deep reinforcement learning},
  journal = {PLOS ONE},
  year = {2025},
  volume = {20},
  number = {1},
  pages = {e0317662},
  doi = {10.1371/journal.pone.0317662}
}
Kostenlos starten →