Developing Construction Workers' Mental Vigilance Indicators through Wavelet Packet Decomposition on EEG Signals

Many studies suggest accidents in construction projects mainly relate to unsafe behaviors of construction workers. When construction workers perceive environment hazards wrongly, they take aggressive shortcuts and opportunistic activities to save effort and time. Therefore, their unsafe behavior highly determined by their perceived risks. In this paper, we propose to develop and utilize some vigilance indicators to quantitatively reflect the perceived risks of construction workers. The indicators are formulated based on electroencephalography (EEG) signals, which associated with human brain activities. However, the detected EEG signal is a hybrid signal that caused by brain activities and muscles movements, which yield to unique data patterns. Therefore, we proposed to implement wavelet packet decomposition to investigate and validate the proposed indicators. We conducted an on-site experiment to analyze the EEG signal patterns and benchmarked the results of the proposed indicators with other existing vigilance assessment approaches.