Blood Glucose Regulation by Switching Between Adaptive Predictive Control and Robust Filter

Abstract

This work presents a set of proposed innovations to modify the original control algorithm composed by a set of rules to choose between two alternatives: Adaptive Predictive part and a Robust Filter (APCwRF). The main objective is to provide improvements to the Artificial Pancreas (AP) technology for patients with Type 1 Diabetes Mellitus (T1DM). The new version, called enhanced (E) APCwRF, has two alternative models the Finite Impulse Response (FIR) and the nonlinear Wiener-type block-oriented. In both cases the model parameters are updated by performing recursive identification using a Kalman Filter (KF). Hence, the control algorithm checks the quality of the prediction of the two options (FIR and Wiener) and decides which of them will be adopted. The robust filter part is able to choose between the two models to be implemented. To test the EAPCwRF we use a previously developed global data-driven model (GM) for defining the final tuning (commissioning). The GM uses the KF to improve the predictions for the next hour only by estimating two weight parameters. This technique allows achieving a good representation of the true glycemia dynamic behavior. The closed loop glycemic responses of the original and EAPCwRF together with real data of the patient without control are compared. The Control Variability Grid Analysis (CVGA) is usefull to warn about which techniques drive to maximum or minimum values outside the recommended ranges. Meanwhile, the percentage of time with dangerous excursions of the glycemic concentrations inside the risky zones are considered too as a tool to ponderate the control performances. Finally, we present the conclusions and discuss some future works.