An Audio-Based Emotion Monitoring System for Enhancing Construction Worker Safety and Mental Health

The mental well-being of construction workers significantly influences their performance and safety. The persistently high suicide rate among construction workers in the US is a significant concern. For every 100,000 construction workers, 53.3 commit suicide, in contrast to the national average of 12.93. Real-time monitoring of the emotional state of construction workers is very important. This research introduces an in situ audio-based human emotion monitoring system to enhance worker safety and proactively address potential concerns. In this study, a unique stimulus system encourages construction workers to engage with the AI, recording their voices. An emotion classification model is then utilized to assess the worker's emotional state. Recognizing the challenges posed by the noisy construction environment, the human voices are mixed with the construction noises recorded on site, enhancing the system's reliability in real-world conditions. This approach allows for real-time emotional monitoring without the need for disturbing equipment like EEGs. Should anomalies be detected in the worker's emotional trajectory, recommended preventive actions can be initiated. This study offers a novel means of monitoring the mental health of construction workers on-site, equipping management with crucial data to make informed intervention decisions.