In-silico works on control of blood glucose level for Type 1 Diabetes (T1D) using improved Hovorka equations and enhanced Model Predictive Control (eMPC) / Amar Mohd Maarof

Artificial pancreas technology has been continuously developed over the past few years. However, there are still flaws found in recent technology in relation to injection of insulin subcutaneously into type 1 diabetes patient. The injection of insulin into the patient boy must be specific, exact and precise to ensure that the blood glucose level is between the normoglycaemic ranges, 4.5 mmol/L to 6.0 mmol/L. If the blood glucose level (BGL) were below or over normoglycaemic range, patients will experience effects caused by hyperglycemia or hypoglycemia. Therefore, the research seeks to find optimum insulin infusion rate into the patient for the blood glucose level to be at safe glycemic range. The research on development of artificial pancreas is mainly focusing on the algorithm that will be programmed into controller system. This research will use enhanced Model Predictive Controller (eMPC) and improved Hovorka equations for insilico works for controlling blood glucose level for Type 1 Diabetes (T1D). The simulations were ran on MATLAB software. Only meal disturbance factor is include and varies in CHO intake during breakfast, lunch and dinner. Reference data to be substituted into related parameter value in the equation for meal disturbance are taken from real life patient data. The simulation was successfully carried out and the result was observed, evaluated and discussed. The value of insulin administered is at 0.0529 U/min, 0.001 U/min and 0.000001 U/min for breakfast, lunch and dinner respectively. Blood Glucose Level (BGL) observed at each meal time is either at between normoglycaemic ranges or at slight deviation.