An Automatic Insulin Infusion System Based on Kalman Filtering Model Predictive Control Technique

During the past few decades, optimal control of blood glucose (BG) concentration with adequate feedback loop has been of ample importance for Type-I diabetes mellitus (TIDM) patients as far as an artificial pancreas realization is concerned. Now-a-days, in addition to the BG control, the design of the micro-insulin dispenser (MID) with a robust control algorithm to regulate the other chronic clinical disorders based on prolonged medications is also quite indispensable. A novel Kalman filtering model predictive controller (KFMPC) has been proposed in this work to solve the aforementioned problem. For designing of the KFMPC, a ninth-order state-space model of the TIDM patient with MID is considered. In this control strategy, the conventional model predictive controller is re-formulated with a state estimator based on the Kalman filtering methodology to improve the control execution. The justification of enhanced control performance of KFMPC is demonstrated by comparative result analysis with other published control techniques. The simulations are carried out through matlab/simulink environment, and the results indicate comparatively better control ability of the suggested algorithm to control the BG level within the normoglycemic range (70–120 mg/dl) as far as accuracy, stability, quick damping, and robustness.