High-risk working conditions in construction, such as working at height, may elicit elevated mental stress in workers and pose significant safety challenges. This study aims to physiologically assess construction workers’ mental stress under high-risk working conditions using heart rate variability (HRV) features derived from electrocardiograph (ECG) signals. An experimental study in the field was conducted, where inexperienced scaffolding workers’ (n = 20) ECG signals were collected when working at three different heights corresponding to low, medium, and high levels of mental stress. Supervised machine learning algorithms, including Support Vector Machine (SVM), KNearest Neighbor (KNN), Linear Discriminant Analysis (LDA), and Random Forest (RF), were applied for model development. The results show that the HRV features obtained good prediction accuracy. The classification accuracy is up to 85.00% between low and medium stress levels, 92.50% for differentiating low and high stress levels, and 87.50% for classifying medium and high stress levels. These findings demonstrate the potential of ECG-derived HRV features for differentiating the mental stress responses of construction workers under high-risk working conditions and provide empirical evidence supporting the feasibility of physiological monitoring of workers’ mental stress in real construction environments.
Lei et al. (2026) studied this question.