Robust Controller Design for T1DM Individualized Model:Gain-Scheduling Approach

Abstract

This paper deals with the robust gain-scheduled controller design for individualized type 1 diabetes mellitus (T1DM) subject model. The controller is designed using LPV model created from T1DM minimal model with two additional subsystems - absorption of digested carbohydrates and subcutaneous insulin absorption. Data collected from continuous glucose monitoring with the help of pharmacodynamics and pharmacokinetics characteristics were used for model identification. The closed-loop stability and cost for all scheduled parameters is guaranteed by the controller design approach. The benefits of the presented approach are shown in the simulation results.