Traffic safety has become a more and more severe problem for urban freeways in China, thus, it is important to establish a real-time crash risk model to identify traffic condition causing crash, which can be used in active traffic management control and accident prevention. In this study, we aimed to explore the real-time crash risk for urban freeways in China and obtain dynamic crash risk level. About three years (January 2013 to January 2016) crash data and their matching traffic sensor data from the Beijing section of Jingha expressway were utilized for this research. The traffic data in eight 5-minute intervals between 0 to 40 minutes prior to crash occurrence was extracted respectively. To obtain the appropriate data training period, the data during eight different periods was collected as training data respectively, and the crash risk value under different data conditions was defined. Then we proposed a novel real-time crash risk assessment model using decision tree method and adaptive neural network fuzzy inference system (ANFIS). By comparing several real-time crash risk assessment methods, such as logistic regression, decision tree and supported vector machine (SVM), it was found our proposed method had higher precision than others. The prediction accuracy of the crash occurrence could reach 69.8% when 0.60 was considered as the crash prediction threshold. This study can be applied to monitor real-time traffic risk on urban freeways in China, forecast the crash occurrence promptly and assist traffic control decisions to reduce and avoid traffic accident.

CICTP