Tuning of an artificial pancreas controller: an in silico methodology based on clinically-relevant criteria - CNRS - Centre national de la recherche scientifique Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Tuning of an artificial pancreas controller: an in silico methodology based on clinically-relevant criteria

Résumé

This paper presents a methodology to tune an artificial pancreas controller by minimizing the time spent in endangering glycaemic ranges (hypo- and hyperglycaemia). The risk associated to the patient’s glycaemia is evaluated with an objective metric (the blood glucose risk index), which has an established clinical relevance. The tuned controller is validated in the UVA/Padova environment where the resulting artificial pancreas achieves minimal glucose risk index in realistic 24-hour long scenarios with unannounced glucose intake.
Fichier non déposé

Dates et versions

hal-02908200 , version 1 (28-07-2020)

Identifiants

  • HAL Id : hal-02908200 , version 1

Citer

Loïc Olçomendy, Antoine Pirog, Yannick B Bornat, Jérôme Cieslak, David Gucik-Derigny, et al.. Tuning of an artificial pancreas controller: an in silico methodology based on clinically-relevant criteria. 42nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Jul 2020, Montréal, Canada. ⟨hal-02908200⟩
55 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More