INTERNAL MODEL CONTROL DESIGN FOR BLOOD GLUCOSE REGULATION IN A TYPE 1 DIABETIC PATIENT

Abstract

Internal Model Control (IMC) applied to the blood glucose regulation process in body of a type 1 diabetic patient in this study. Minimal Model (BMM) adopted as the type 1 diabetic patient was linearized using Taylor Series expansion to get appropriate transfer function.  Insulin infusion rate was used as the manipulated variable to control the blood glucose concentration while glucose meal was used as disturbance variable. IMC algorithm was developed to control the process model using the IMC structure along with the Single Input Single Output (SISO) tool in MATLAB. Single parameter associated with IMC was tuned by using percentage of settling time (% OLST) for open the loop model representing the patient. The parameter equivalent to 5% OLST gave the best setting for full-order IMC with closed-loop settling time of 55.913 min, without violating input constraint. The parameter equivalent to 10% OLST gave the best setting for reduced-order IMC with closed-loop settling time of 93.788 min. The controllers gave good performance in term of set point tracking and disturbance attenuation. Comparison of IMC with PID controllers based on Zeigler-Nichols and Cohen-Coon indicated that IMC outperformed PID controllers.