| Signalized intersections are key traffic terminals connected to the city road and the concentrations at which accidents frequently happen. There are a large number of crossing pedestrian flow and right-turning vehicle flow at the intersections, furthermore the competition on time and space between these two flow leads to highly incidence of conflict between pedestrian and right-turning vehicle. These conflicts have significantly influence on the traffic safety at intersections. The present studies on conflict between pedestrian and right-turning vehicle are mostly built around pedestrian crossing characteristic and right-turning traffic flow characteristic. However, the discussion about pedestrian and right-turning vehicle safety using traffic conflict technology is given few consideration during this stage of research production. Simultaneously research production in the past lost sight of the dynamic interaction between pedestrian and right-turning vehicle during conflict process and exploration of influential factors of the conflict severity level.The paper takes conflict between pedestrian and right-turning vehicle at signalized intersection as the research objects. At the beginning part of this paper, we select suitable conflict indicators for quantifying conflict between pedestrian and right-turning vehicle based on characteristic of each traffic conflict indicator, that is Time-To-Collision (TTC), Deceleration-To-Safety Time (DST), Time Advantage (TADV), and come up with the calculation model of these three traffic conflict indicators; Then, paper designs reasonable traffic investigation, combined with individual image tracking and trajectory calibration procedure to get trajectory data of pedestrian and right-turning vehicle, furthermore achieved data preprocessing and data secondary processing to get the final traffic indicator values by using MATLAB software; Thirdly, all the conflict samples are divided into potential conflicts, minor conflicts and serious conflict according to conflict severity by using K-means cluster analysis method which regards traffic indicators as evaluation indicators. At the same time, through Ordered probit statistical analysis model we explore the significant factors of pedestrian and right-turning conflict severity level which include the right-turning traffic volume, pedestrian traffic volume, speed changes on right-turning vehicle, passing situation of pedestrian and right-turning vehicle, pedestrian violation, and calculate the marginal benefit of each variable. Eventually, pedestrian and right-turning vehicle safety control method which is on the basis of above research conclusion about conflict severity level influence factors is put forward. |
