Evaluación de controladores PID para el control de glucemia de pacientes con diabetes tipo 1 a partir de sincronizaciones mediante el método LAMBDA

Abstract

In this degree work, an adjustment method called Lambda Tuning, widely used in PID control algorithms at an industrial level, was studied for the control of glycemia in patients with type 1 diabetes. Two controllers were implemented, a PID and a PD- basal, which were used for glycemic control of the average adult patient with type 1 diabetes that is virtually included in the Uva-Padova simulator, approved by the Food and Drug Administration (FDA) as a substitute for conducting preclinical trials in animals. For the application of the Lambda method, first order models plus dead time were determined for the average adult patient of the Uva-Padova simulator in different operating conditions (operating points). These models were used to adjust the PID and PD-basal controllers. Finally, an evaluation of the performance of the controllers and of each control configuration was carried out under different test conditions (scenario of one and three meals). The performance obtained was analyzed according to metrics such as time in normoglycemia and percentage of time in hypoglycemia. It is expected that, thanks to the completion of this project, new adjustment alternatives for PID controllers can be implemented in Artificial Pancreas systems.