S.-H. Huang

Pages: 21-38

Abstract
Perception reaction time (PRT) is an important parameter that incorporates drivers’ reaction capability in roadway design. However, the current 2.5 s PRT has been challenged as too low because drivers would have fleeting visions away from the frontal side from time to time. Part of the 2.5 s is usually spent before drivers can identify the existence of an obstacle. Complex driving conditions and unexpected events can also impact drivers’ reactions. As a consequence, an incident can happen when the required PRT is longer than the one used for car-following or in roadway design. Therefore, from the attention allocation perspective, this study aims to investigate the duration that drivers allocate their attention to forward and other non-forward focal points under varying conditions. Moreover, the robustness of the 2.5 s PRT rule is examined from the perspective of driver attention allocation. This study applied the naturalistic glance data from the 100-car event database. The results of glance duration analyses were promising and clearly indicated that drivers glancing consecutively at more than one non-forward focal point are more likely to encounter insufficient time for responding to harmful changes in front of them. Furthermore, drivers are found to alter the duration of glancing at forward and non-forward focal points for tasks with different levels of distractions, maneuver intentions, and traffic density. As for the 2.5 s PRT rule, results showed that the current standard is robust for most driving conditions.
Keywords: road safety; attention allocation; perception reaction time; naturalistic driving