Hirokazu Akahane

Assessing Safety of Signalized Intersections

Traffic accidents constitute a serious social problem. Thus, it is essential to identify and remove the risk factors affecting traffic accidents to enhance traffic safety at intersections. This study identified and quantified the main factors influencing traffic accident risk at signalized intersections to propose effective countermeasures. Google Earth was used to collect numerical data related to the geometric attributes of intersections in three regions in Japan: Kagawa, Shiga, and Aichi. A lognormal hurdle model that considered regionality and geometric attributes was then used to quantify factors influencing the risk of traffic accidents involving various types of collisions. The results indicate the existence of significant regional differences in the geometric attributes of intersections in those regions. The findings reveal that intersection size (indicated by the distance between stop lines), length of crosswalks, and setback distance of crosswalks generally and significantly influence all collision risk types; in addition, a compact intersection lowered the risk of all collision types. And in most cases, regional dummy variables were statistically significant. This finding suggests that in addition to intersection geometric attributes, regional differences in some factors influenced collision risk. Thus, it could be reasonably assumed that regionality stemmed from differences in driving characteristics.

Saga of Traffic Simulation Models in Japan

This paper narrates a history of developments in the traffic simulation models in Japan. Starting from 1971, basically two kinds of logic existed for reproducing dynamic traffic flow. These were the Block Density Method and the Input-Output Method. These methods were compared to the calculation engine of the network traffic simulation model. Subsequently, these methods evolved, were modified, and became more advanced as found in the AVENUE and SOUND simulation models. This development catered better to the changes in needs required by newer traffic simulation models and resulted in part from developments in computer technology. The first models do not include drivers’ route choice behaviour, but the later do. Initially these methods were applied to urban expressways; however, they later expanded to include surface street networks.