At construction workplaces, workers should be consistently attentive to approaching and nearby safety hazards. However, workers tend to allocate most of their attentional resources to a work task and often exhibit inattentive behaviors to hazards, which may lead to serious injuries and fatalities. Predicting construction workers’ inattentiveness is thus critical to preventing accidents in construction workplaces. With the advent of biosensing technologies, the potential of using biosignals to predict human behaviors has been proven in various fields of study. However, to date there has been little discussion about utilizing biosignals to predict construction workers’ inattentive behaviors. To this end, this study examines whether construction workers’ inattentive behaviors can be predicted by assessing biosignal reactivity. A virtual road construction environment was created and used for an experiment to expose participants to a repeated struck-by hazard without risking actual injury. Participants’ biosignals (i.e., electrodermal activity, pupil dilation, and saccadic eye movement) and physical engagement in inattentive behaviors were collected and analyzed. The results of statistical analyses revealed significant differences in biosignal reactivities between participants’ attentive behaviors (i.e., paying attention to the hazard) and inattentive behaviors (i.e., ignoring the hazard). The outcomes of the machine learning-based behavior classification also indicate the usefulness of predicting inattentive behaviors by monitoring workers’ biosignals during a construction task and provide a foundation for the utilization of biosignals in safety management to prevent accidents resulting from inattentive behaviors.

在建筑工作场所，工人应始终注意接近和附近的安全隐患。然而，工人倾向于将大部分注意力资源分配给工作任务，并且经常表现出对危险的疏忽行为，这可能导致严重的伤害和死亡。因此，预测建筑工人的注意力不集中对于预防建筑工作场所的事故至关重要。随着生物传感技术的出现，使用生物信号预测人类行为的潜力已在各个研究领域得到证实。然而，迄今为止，关于利用生物信号来预测建筑工人的注意力不集中行为的讨论很少。为此，本研究检查是否可以通过评估生物信号反应来预测建筑工人的注意力不集中行为。创建了一个虚拟道路施工环境并用于一项实验，让参与者暴露于反复被撞的危险中，而不会冒实际受伤的风险。参与者的生物信号（即皮肤电活动、瞳孔扩张和扫视眼球运动）和注意力不集中行为的身体参与被收集和分析。统计分析结果表明，参与者的注意力行为（即注意危险）和注意力不集中行为（即忽视危险）之间的生物信号反应性存在显着差异。