Desarrollo de Potenciostato para el análisis en muestras de orina en pacientes patológicos y procesamiento de datos.

Abstract

This work presents the development of a potentiostat device controlled by a Raspberry Pi embedded system for rapid in-situ analysis of urine samples in diabetic and non-diabetic patients using screen-printed gold, platinum and carbon sensors. To this end, the development of this project was structured in three moments that were; the construction of the data acquisition system of the urine sample, conditioning of the embedded system for data collection and storage through the design of a graphical interface designed in Python software and executed within the Raspberry Pi system. Finally, the construction of the voltammetry curves resulting from the current signals recorded by the system as a function of the potential applied to the urine sample. A data set was created with the results obtained by applying statistical algorithms such as average, standard deviation, energy, entropy, maximum and minimum peaks to these signals in order to determine which technique is the most optimal to perform a classification of patients with diabetes and patients without diabetes, a correlation matrix was analyzed for the discrimination of patons and the implementation of a learning machine based on least squares Vector Support Machines. For the development of these tests, urine samples were taken from 20 diabetic patients and 20 control patients, at the moment of applying the learning machine, it was possible to discriminate between diabetic and non-diabetic patients.