REGULATION OF BLOOD GLUCOSE IN TYPE I DIABETIC PATIENTS

Abstract

There has been a significant amount of research in developing closed loop artificial pancreas to mimic the normal pancreatic function in Type 1 Diabetic (T1D) patients. Such a system involves glucose sensor, continuous insulin infusion pump and a robust control algorithm to infuse the right amount of insulin in a timely manner to maintain blood glucose (BG) in normoglycemic region. In an era of P4 (Personalized-Predictive-Preventive-Participatory) medicine, it would be worthwhile to develop controllers that: (i) cope well with inter- and intra-patient variability, and (ii) are simple, practical to implement but yet effective. Towards this end, this doctoral study proposes, implements and compares the performance of rule-based controllers, robust controllers and personalized controllers based on mathematical models. The proposed algorithms are validated on a wide range of virtual patient cohorts via simulations and their efficacies are compared.