Modeling of left-turn vehicle trajectories and exit lane selection at signalized intersections

Abstract

Abstract Under the urban road system, the most frequent locations for traffic accidents are concentrated at road intersections, where left-turning traffic is one of the most direct and important factors causing intersection conflicts. There are relatively few applied models that can effectively explain the trajectory change characteristics and exit lane selection behavior of left-turning vehicles, making it more difficult to achieve accurate analysis and prediction of intersection vehicle conflict behavior. This study extracted trajectories of left-turning vehicles at the intersection based on the video captured by the UAV, and a vehicle left-turning trajectory model was proposed based on the curvature change of the vehicle trajectory. Also, based on the geometric parameters of the intersection and video data, a vehicle exits lane selection model with intersection turning angle, vehicle type, vehicle position at the entrance lane, vehicle speed and following clearance, and front car position at the exit lane as variables was developed using random utility theory. The results show that the left-turn trajectory could be fitted with a flat curve, and the curvature of the trajectory varies with the aspect ratio of the intersection in a quadratic parabolic relationship. The left-turn trajectory model established in the paper could effectively characterize the driving trajectory of left-turning vehicles at different intersections. The exit lane selection model could effectively predict the distribution of left-turning vehicles in the exit lane position, as the turning angle, vehicle position at the entrance lane and vehicle speed have a significant effect on the choice of exit lane for left-turning vehicles.