A Study of Collision Risk Estimation and Users Evaluation at Merging Section of Urban Expressway in Japan

Abstract

For usability of urban expressway in Japan, there are some problems due to the limitation of spatial con-straint. As the result of the limitation, there are merging sections that do not have enough length and set a right-side entrance ramp which is connected to the passing lane, directly. Therefore, unfamiliar drivers such as female and/or elderly drivers are subjected to stress and aware of danger when they merge onto the expressway and it is necessary to measure these situations quantitatively from the viewpoints of both traffic conflict risk and users evaluation. The purpose of this study is to estimate collision risk by using Surrogate Safety Measure and evaluate Quality of Service for drivers at merging sections of urban expressway in Japan. We focused on each ve-hicle merging movements and traffic conflicts at three on-ramps, which has different geometries, by im-age processing technology. First, in order to elucidate the occurrences of collision risk in merging section, we analyzed the collision risk between a merging vehicle and a main lane vehicle by using the Possibility Index for Collision with Urgent Deceleration (PICUD) index. It is shown that the position of conflicts is affected by the amount of traffic volume. Second, we conducted discriminant analyses for the results of the PICUD estimation in order to reveal the influence factors of collision risk at each merging section. As the result of discriminant analyses, it is found that the percentage of merging position and traffic volume of main lane have a huge effect on collision risk for merging movement. In addition, both the percentage of heavy vehicles and the existence of other merging vehicle at a time trigger rear-end collision risk. Fur-thermore, we examined some countermeasures to reduce collision risk in objective merging sections by sensitivity analysis and proposed a Level of Service index which is based on the rank of rear-end collision risk considering users’ subjective evaluation.