Hardware Implementation of a Non-Linear Observer for a Diabetes Mellitus Type 1 Mathematical Model

Abstract

This research paper studies a type 1 Diabetes Mellitus (DM1) model in which the dynamics of insulin, glucose and beta cells variables in a non-linear system given by ordinary differential equations are described. In order to estimate glucose and insulin from beta cells, a reduced order observer is designed, using Lyapunov's theory of stability and Lasalle's theorem. Both the DM1 mathematical model and the observer are implemented in hardware using the Altera Cyclone V FPGA development card, where the efficiency of the designed observer is checked, obtaining the dynamics of the error. In this way, the benefits of reconfiguration of an FPGA are used and it is demonstrated that the evolution of glucose in a patient can be predicted by means of a low power embedded device, which aims to improve the quality of life of people suffering from DM1 that require a constant invasive sampling of the level of beta cells in the blood, looking for treatments to be determined based on this information.