Utilizing Decision Tree Method and ANFIS to Explore Real-Time Crash Risk for Urban Freeways

Traffic safety has become a severe problem on freeways in China, thus, it is important to establish a real-time crash risk model to identify traffic conditions causing crashes. In this study, we explore the real-time crash risk for urban freeways in China and obtain dynamic crash risk level. Crash and their matching traffic sensor data from a Beijing section of Jingha expressway in eight 5-min intervals between 0 to 40 min prior to crash occurrence was extracted during eight different periods. The crash risk value under different data conditions was defined. Then, a real-time crash risk assessment model using decision tree method and adaptive neural network fuzzy inference system (ANFIS) was proposed. Comparing several real-time crash risk assessment methods, such as logistic regression, decision tree and supported vector machine (SVM), it was found the proposed method had higher precision than others. This study can be applied to monitor real-time traffic risk on urban freeways, to assist traffic control decisions and reduce traffic accidents.