Traffic safety analysis on mixed traffic flows at signalized intersection based on Haar-Adaboost algorithm and machine learning

摘要

Mixed traffic is the typical characteristic of urban traffic in China. Especially on the crosswalks at signalized intersection, mixed traffic flow is large, traffic conflicts happen frequently. So, the traffic safety analysis is one of the research directions at signalized intersection. The main conflicts on the pedestrian crosswalk include the conflict among pedestrians and non-motorized vehicles as well as the conflict between the right-turning motor vehicles and pedestrians. Several impacts are imposed on the order of mixed traffic flow on the crosswalk by the above conflicts. In this paper, on the basis of on-the-spot filming and investigation, 3 regions are categorized based on Harr-Adaboost algorithm according to traffic characteristics of mixed traffic flows. Typical parameters such as speed and trajectory of pedestrian, non-motorized vehicle and right-turning motor vehicle as well as the variation of each characteristic results from the traffic conflicts in 3 regions including the pedestrian crosswalk and its vicinity are analyzed. The concluded characters could be used to optimize the intersection design and traffic control and improve the safety of mixed traffic individuals.