Output-Feedback Sliding-Mode Controller for Blood Glucose Regulation in Critically Ill Patients Affected by Type 1 Diabetes

Roberto Franco, Alejandra Ferreira De Loza, Hector Rios, Louis Cassany, David Gucik-Derigny, Jerome Cieslak, David Henry, Loic Olcomendy

Research output: Contribution to journalArticlepeer-review

11 Scopus citations

Abstract

In this brief, an output feedback continuous twisting algorithm is applied to solve the problem of glucose regulation in critically ill patients affected with type 1 diabetes mellitus. The proposed approach copes with disturbances coming from intrapatient and interpatient variability and does not need meal announcements. The blood glucose measurement and insulin infusion are via intravenous. The proposed algorithm regulates blood glucose and keeps it in the normoglycemia range, i.e., 70-180 mg/dl without nursing intervention. Therefore, the workload at the intensive care unit is alleviated. Besides, we provide a criterion to set the output feedback continuous twisting algorithm gains, which may be used for individualizing the insulin therapy. The approach is validated in the Universities of Virginia and Padua (UVA/Padua) metabolic simulator for ten in silico adult patients with unannounced meal intake. The results show excellent performance as well as minimal risk of hyperglycemic and hypoglycemic events. In addition, an open-loop study case is presented to evidence the advantages of output feedback continuous twisting algorithm over a traditional brief-based protocol in terms of nurse workload mitigation.

Original languageEnglish
Pages (from-to)2704-2711
Number of pages8
JournalIEEE Transactions on Control Systems Technology
Volume29
Issue number6
DOIs
StatePublished - 1 Nov 2021

Bibliographical note

Publisher Copyright:
© 1993-2012 IEEE.

Keywords

  • Diabetes mellitus
  • glucose regulation
  • in silico validation
  • sliding mode control

Fingerprint

Dive into the research topics of 'Output-Feedback Sliding-Mode Controller for Blood Glucose Regulation in Critically Ill Patients Affected by Type 1 Diabetes'. Together they form a unique fingerprint.

Cite this