Statistical analysis of vehicle-vehicle conflicts with a LIDAR sensor in a signalized intersection

LIDAR sensor is capable of recording traffic data including the number of passing vehicles, pedestrians, and bicyclists, the speed of vehicles, and the number of conflicts among different road users. As part of intelligent mobility and in order to collect real-time traffic data and investigate the safety of different road users, particularly motorized vehicles and pedestrians in different approaches of the intersection, a LIDAR sensor was installed at Cold Spring Ln – Hillen Rd intersection in Baltimore city. The intersection was chosen based on substantial vehicle traffic, proximity to Morgan State University, and the history of recent crashes. One of the efficient capabilities of the installed LIDAR sensor is recording the “Post Encroachment Threshold (PET)” between two vehicles, vehicle-pedestrian, and vehicle-bicyclists. The paper aims to concentrate on vehicle-vehicle (including car-car, car-bus, car-truck, and bus-truck) conflicts. The frequency and severity of conflicts were analyzed and the results highlighted that 857 conflicts were recorded during a 5-month time interval. The frequency and severity of conflicts were investigated by three methods including leading-following vehicles, right-turn and left-turn movements, and in different phases of traffic signal. The critical zones were recognized, and K-means clustering was performed on severity of conflicts subcategories while selecting the optimal number of clustering. The significant error of each subcategory was specified, and the error values were compared with Bonferroni test error values. The results demonstrated that the error values from k-means clustering are consistent with the error values from the Bonferroni test. Furthermore, the statistical analysis showed that: 1) In terms of conflict frequency and severity, the eastern and middle zones are crucial 2) all errors are consistent with 95% confidence interval, and 3) the severity of conflicts by different movements has a significant correlation with speed at conflict point, hourly time interval, weather, and the entrance vehicle volume to each zone.