Stress and fatigue evaluated with help of textile sensors embedded in smart clothes and artificial intelligence methods in human daily life activity

The purpose of the study was to evaluate physiological stress and fatigue based on cardiac activity in healthy subjects with help of textile sensors embedded in smart clothes. 18 practically healthy subjects aged 19-55 years participated in the study. The cardio-respiratory activity was collected with help of Hexoskin smart garments (Hexoskin Shirt, Carré Technologies Inc., Canada) in daily life (rest condition, physical and mental professional and leisure activity). Heart rate, respiratory rate and motion were monitored. From the heart rate, variety linear and nonlinear parameters of heart rate variability (HRV) were computed. In addition, subjects evaluated their level of stress with help of analogous visual scale. The data were processed with help of Machine Learning Algorithms (Random Forest, CatBoost, XGB, LGBM). All algorithms allowed predicting the level of both strain and fatigue (ranged from 1 to 10) with probability 83%. The Random Forest Classifier proved the best in assessing the level of both stress and fatigue.