Backstepping Model Predictive Controller for Blood Glucose Regulation in Type-I Diabetes Patient

Abstract

A robust method for optimal insulin injection as an accurate remedial measure has been focused to design since early three decades. This paper presents a novel approach to regulate the blood glucose (BG) based on backstepping model predictive controller (BMPC) for type-I diabetes mellitus (TIDM) patients. To implement the proposed approach BMPC, initially a linearized state-space model is derived from the non-linear TIDM patient with micro-insulin dispenser (MID). Through necessary simulation in MATLAB/SIMULINK platform, the proposed approach is verified and the results are presented to demonstrate its accuracy, stability, and robustness. The comparative results are presented with other suggested methods by various researchers in recent times. The results obtained under various physiological conditions with regards to meal, exercise, and uncertainties due to noise reveal the enhanced performance of the approach to control the BG level within the range of normoglycaemia (70–120 mg/dl).