Control of a T1DM Model Using Robust Fixed-Point Transformations Based Control With Disturbance Rejection

Control of blood glucose levels for type 1 diabetes (T1DM) patients using artificial pancreas is a challenging task due to disturbances, and inter- and intrapatient variability. The solution to these problems requires a special control algorithm. A Robust Fixed-Point Transformation (RFPT) based controller was developed to control a T1DM diabetes model, which uses a linear state feedback controller in the kinematic prescription. The controller was implemented in discrete-time using computer algebra software to compute the equations for input-output linearization. Parametric disturbances between the model and the controller were also introduced which showed that the RFPT approach can provide beneficial results in the presence of these circumstances. The performance of the controller was shown through a simulation from which both the advantages and limitations of the scheme were analyzed.