Eye-Tracking Experimental Study Investigating the Influence Factors of Construction Safety Hazard Recognition

This material may be downloaded for personal use only. Any other use requires prior permission of the American Society of Civil Engineers. This material may be found at https://ascelibrary.org/doi/abs/10.1061/%28ASCE%29CO.1943-7862.00018...

Construction site accidents could be reduced if hazards leading to accidents are correctly and promptly detected by employees. The proactive safety measures such as safety perceptions and safety detection capability of employees play an important role in improving the safety performance. This study was initiated by three research questions related to: (1) the measurement indicators of employees’ cognitive load in recognizing safety hazards; (2) site condition factors (e.g., brightness) that could affect subjects’ cognitive load; and (3) the quantification of the effects of these site factors on cognitive load. An eye-tracking experimental approach was adopted by recruiting a total of 55 students from construction management or other civil engineering disciplines to visually search hazards in 20 given site scenes. These site scenes were defined by a combination of three different categories, namely distinctiveness of hazards, site brightness, and tidiness. Quantitative measurements of experimental participants’ visual search patterns were obtained from data captured by the eye-tracking apparatus. Based on metrics related to experimental participants’ fixation, visual search track, and attention map, these measurements were computed to evaluate participants’ cognitive load in detecting hazards. Descriptive statistical comparisons were performed to analyze these metrics under pre-defined categories of site conditions, i.e., distinctness versus obscurity/blur, brightness versus darkness, and tidiness versus mess. The findings revealed that: distinct site conditions reduced participants’ time in saccades to search hazards but did not improve the accuracy rate of first fixation; messy sites with dis-organized items increased participants’ cognitive load in detecting hazards in terms of all five measurement items (i.e., accuracy rate of first fixation, fixation count, intersection coefficient, fixation duration, and fixation count in the attention center); the effect of increased brightness on-site was a double-edged sword which needed further studies on the optimal balance of brightness level and allocation. Recommendations based on the findings were provided to enhance safety education in terms of site hazard distinctiveness, brightness, and housekeeping best practice. This study extended a few prior studies in adopting the eye-tracking technology for safety monitoring by evaluating the impacts of site conditions on participants’ cognitive load which was linked to their hazard detection performance. The current study provided insights for evaluating construction employees’ hazard detection capabilities to enhance safety education. Future work was proposed in evaluating employees’ safety hazard detection pattern under dynamic construction scenarios.