Estimativa de Pressão Arterial Utilizando Dados de Fotopletismografia (PPG) de Curta Duração

Abstract

The use of artificial neural networks (ANN) to estimate systolic and diastolic blood pressure using photoplethysmography (PPG) signal characteristics. Models were developed based on a database containing physiological information and short-term PPG signals. Analysis of the PPG signals and their derivatives, VPG and APG, were carried out, seeking to find ways to obtain characteristics of the signals to feed the ANN, during such evaluations it was possible to perceive that there are certain groups of individuals who present APG with specific behavior. Some characteristics of the signals, especially the APG, could be used to classify individuals into groups, such classification could be applied to improve the estimation or for other purposes. The best results had a mean absolute error of 7.16 ± 0.50 mmHg for diastolic blood pressure and 12.75 ± 1.50 mmHg for systolic blood pressure. The results show that the model is capable of estimating blood pressure through short-term PPG signals.