Learning classifiers for analysis of Blood Volume Pulse signals in IoT-enabled systems
Gloria Cosoli, Grazia Iadarola, Angelica Poli, Susanna Spinsante
Abstract
Physical exertion undoubtedly influences physiological parameters. The aim of this paper is to propose a Machine Learning classifier able to evaluate the physical state of subjects monitored through a wearable device, by simply analysing their Blood Volume Pulse signals. Moreover, a Fatigue-Related Index is presented to quantify the physical well-being status. Results show that the Support Vector Machine classifier provides the best performance for detecting fatigue-induced stress, since it shows an accuracy of 97.50%. The obtained results prove that the proposed approach allows to support the assessment of the worker's well-being status, with the aim of improving the workload management in the context of Industry 4.0.