| Dynamic Modeling and Model-Based Control of Exercise Disturbances in Type 1 Diabetic Patients | ||
| Anirban Roy and Robert S. Parker, Department of Chemical and Petroleum Engineering, University of Pittsburgh, 3700 O'Hara Street, Room 1249 Benedum Hall, Pittsburgh, PA 15261 Diabetes mellitus is a metabolic disease caused by either the loss of pancreatic insulin secretion (Type-I) or resistance developed by the body towards the glucoregulatory action of insulin (Type-II). In order to prevent major health complications, it is important to maintain plasma glucose concentration within the normoglycemic range (70 - 120 mg/dl) [1,2]. The major long term effects of diabetes are caused due to hyperglycemia, where the plasma glucose concentration exceeds 120 mg/dl due to insufficient endogenous insulin secretion [1,2]. Of more immediate concern is hypoglycemia when the plasma glucose concentration falls below 70 mg/dl. Such conditions can lead to dizziness, coma or even death [1,2]. The most common treatment for insulin-dependent diabetes involves periodic injections of insulin to maintain the blood glucose concentration within the normoglycemic range. Such approaches are adequate, but wide glucose variations still persist. Regulation of plasma glucose concentration for type 1 diabetic patients during exercise can prove to be challenging. Physiological exercise induces several fundamental metabolic changes in the body [3]. An increase in exercise intensity up-regulates glucose uptake by the working muscles. In order to maintain plasma glucose homeostasis, hepatic glucose production also increases [4]. During prolonged exercise, hepatic glycogen stores begin to deplete, which leads to a reduction in hepatic glucose production as the glucose production mechanism shifts from glycogenolysis to gluconeogenesis [5]. Since the energy requirement at a given exercise intensity is approximately constant, the overall plasma glucose concentration tends to fall well below the normoglycemic range for prolonged exercise durations. Elevated exercise intensity also promotes a drop in plasma insulin concentration from its basal level which is essential to maintain glucose homeostasis. Insulin dependent diabetic patients can experience mismatching of endogenous glucose production to glucose uptake during exercise resulting from inappropriate insulin administration by subcutaneous injection which can lead to hypoglycemia. For this reason there is a focus on automated insulin delivery systems, many of which are model-based. In this case, the quality of the model plays a key role as theoretically achievable performance is limited by model quality [6]. The minimal model developed by Bergman et al. [7] was extended to include the major exercise effects on plasma glucose and insulin levels. Additional differential equations were added to the original model structure to capture the glucose uptake, hepatic glucose production and insulin clearance induced by elevated exercise intensity. Parameters representing the exercise effects were estimated from published literature data. Model predictions of glucose and insulin dynamics during and after exercise were consistent with the literature data. Model predictive controllers (MPC) were synthesized based on the extended model in order to maintain normoglycemia for a Type 1 diabetic patient during and after exercise. The results from the closed-loop simulations were also promising; the MPC was able to maintain normoglycemia during mild-to-moderate exercise lasting upto 60 min. Comparisons to control in the absence of exercise information are also discussed.
Extended Abstract Status: File Uploaded | ||