Search

The purpose of this study is to investigate the unsafe conditions of pedestrian crossings at signalized intersections under the mixed traffic environment in Changsha, China. For this purpose, based really pedestrian-related crash data of Changsha city and visualized crash geographic information, six typical urban signalized intersections were selected. The peak and off-peak surveillance videos of six signalized intersections were collected via field surveys, 1,070 conflicts were extracted through using professional trajectory tracking software Traker to track pedestrian and vehicle trajectories. PSMT (Pedestrian Safety Margin Time) was used as an indicator to identify the severity of conflicts, and manually recorded the characteristics of pedestrian crossing behavior, vehicles and roads during the conflict. An Ordered Probit model was established to analyze the risk factors that led to different severities of the conflict between pedestrians and vehicles. The model outcome indicates a significant relationship between the severity level and characteristics of pedestrian behavior, vehicles and conflict. Moreover, it was identified that, in addition to the above characteristics, roadway characteristics significantly influenced likelihood of severe pedestrian-vehicle conflict. Finally, these factors are discussed, and suggestions for improving the pedestrian traffic environment are proposed from different perspectives. [ABSTRACT FROM AUTHOR]

• A combination of GIS methods and binary logistic regression and tree-based models was applied. • Several clusters of pedestrian crashes were identified and the most significant accident-prone area was discussed. • The frequency of pedestrian crashes is correlated with urban functional zoning, population and road network distribution. • Significant risk factors- darkness, lighting conditions, road isolation facilities and pedestrian age and behavior. Improving the pedestrian traffic environment is important for the sustainable development of the city. This study goes beyond the traditional analysis of traditional traffic accident police reports by integrating them with rich geographic information system (GIS) resources to analyze a sample of 791 pedestrian crashes that occurred in the main urban area of Changsha city, China, from 2014 to 2016. The use of Baidu or Google Street View maps improves the accident reports and facilitates the detailed description of the characteristics that are related to pedestrian crashes. The focus of the analysis is to use GIS to visualize the distribution of pedestrian crashes in cities to explore the relationships between pedestrian crashes and the population, road network, land use and social services and activities and to analyze the impacts of the building environment and road characteristics on the severity of pedestrian crashes by combining the binary logistic regression and tree-based models. The results demonstrate that there are several clusters of pedestrian crashes in urban areas, which are related to the population, road network, regional functional zoning and social and economic characteristics. However, the severity of pedestrian casualties has strong relationships with darkness, lighting conditions, road isolation facilities and pedestrian age and behavior. Casualties are more severe at night than during the day, and school-age children and elderly pedestrians tend to suffer more. This study identifies the important influencing factors of urban pedestrian safety and of the risk awareness of pedestrians and other road users and proposes solutions from the perspectives of behavior and infrastructure. [ABSTRACT FROM AUTHOR]