A Model for Predicting Construction Worker Fatigue

Fatigue impairs workers’ judgment, reduces their productivity, and jeopardizes their safety. The paper presents a tool to predict workers’ fatiguebased on their vital signs. An experimental study was conducted in whichthe heart rate and sleep quality for three individuals were monitored usingfitness trackers (wearable sensors). The data collected were used to developtwo models based on regression analysis and Artificial Neural Networks(ANN), to predict their fatigue level. A Borg’s scale was used to estimatethe Rating of Perceived Exertion (RPE) of the participants. The two modelswere able to satisfactorily predict the RPE (workers fatigue level) with anaverage validity of 75% and 80% for the regression ANN models, respectively. The developed models can provide project managers and superintendents with early warning to avoid potential worker overexertion, injuries,and fatalities.