Dynamic Models and Open-Loop Control of Blood-Glucose for Type 1 Diabetes Mellitus

People with type 1 diabetes must rely on exogenous insulin for survival. A good control of blood glucose will substantially reduce the possibility of complication. In this research, modeling the blood glucose using the clinical data of a type-1 patient from a continuous glucose monitoring system (CGMS) is studied. Two models, DDE model of Jiaxu Li (2006) and physiological model of Hovorka (2006) are considered for this modeling of blood glucose. Appraisal of the results from the two models is made. The appraisal includes its ability to estimate the amounts of insulin needed for basal and bolus injections, its ability to predict the future blood glucose, and its sensitivity to detect hypoglycemia and hyperglycemia. The basal and bolus injections mentioned are aimed to maintaining a fasting blood glucose and a postprandial blood glucose within a reasonable range, respectively. Control strategies for patients that use continuous subcutaneous infusions via an insulin pump or discontinuous subcutaneous basal and bolus injections are presented. Comparisons between these two strategies in terms of blood glucose and insulin dosage are also made.