Bihormonal Sliding Mode Controller Applied to Blood Glucose Regulation in Patients with Type 1 Diabetes

Abstract

Control algorithms for a bihormonal artificial pan-Creas are proposed in such a way that insulin and glucagon actions are incorporated, aiming to avoid hypoglycemic and hyperglycemic episodes. Such control algorithms are utilized in order to ensure the blood glucose regulation in type 1 diabetic patients. The mathematical model utilized has experimental validation and represents the glucose-insulin-glucagon dynamics. First-Order Sliding Mode Controllers with and without Boundary Layer were utilized. One desires to reach, in finite time, an equilibrium condition of the closed-loop control system, even in the presence of disturbances related to feeding. The stability proof is presented assuming that the state is known. In the simulation results, exact differentiators are used in order to recover the information associated with the unmeasured state variables of the system. At last, the performance of two control strategy is evaluated by means of numerical examples.