Identifying Pedestrian Crash Contributing Factors using Association Analysis and Their Implications for Development of Active Pedestrian Safety System

Abstract

An active pedestrian safety system (APSS) would be more effective by considering the implications of crash contributing factors. In addition, the APSS needs to be tested and evaluated in the field; therefore, a comprehensive scenario library is necessary. In this study, 135 pedestrian crash reports were investigated. The driving reliability and error analysis method was first applied to identify the contributing factors that can be potentially solved by the APSS function; then, the association rules method was adopted to analyze the joint effect of contributing factors and roadway facility features on injury/fatal pedestrian crashes. The results showed that “inattention,”“failure intention prediction,”“reduced visibility,” and “temporary/permanent obstruction of view” were the first four most frequent contributing factors. Moreover, injury/fatal pedestrian crashes resulting from “failure intention prediction” and “temporary/permanent obstruction of view” were more likely to occur at a location with more than three lanes, a curb shoulder, and a posted speed limit of 40–45 mph. Further, based on the crash contributing factors, the APSS’s functional design is suggested to provide conflict-time-based warning information, pedestrian movement prediction, and detection and tracking of moving objects behind the obstruction. The APSS’s sensing ability is required to detect the vehicle’s nearby area and to be adaptive to poor lighting conditions. Finally, a scenario library was proposed for field testing/evaluation of the APSS. The scenario library has 10 sub-scenarios with detailed object configurations as well as required testing/evaluation items for the APSS. This study’s findings would be helpful for automobile manufacturers to improve the APSS.