Regresar

A Multivariate Signal Analysis of a Sensing Platform Prototype for Stress Detection

Abstract:

This article presents a multivariate signal analysis of a sensing platform prototype specialized in stress analysis. Five physiological signals from 50 subjects were simultaneously recorded by the platform, as each one watched four audiovisual stimuli (three minutes and twenty seconds long) performed on two different sessions, to stimulate a state of relaxation or stress. The objective of this work is to use the raw physiological data and extract the time and frequency characteristics of each signal, to generate a classification model and use it to differentiate between a baseline and stressed estate (psychological or arithmetic) of subjects. The results suggested that some physiological signal as respiratory flow and galvanic skin response are better to be used in combination with a classification model to identify if it is psychological or arithmetic stress.