Image Data Analysis on Traffic Congestion at Merging Section of Urban Expressway

Abstract

This paper attempts to figure out the mechanism of traffic congestion occurrence by using the video image data. Video observation has been carried out at the Moriguchi-Kanjo merging section of Hanshin Expressway, and vehicle trajectories are extracted from the video by using image processing technique. Macroscopic traffic flow characteristics such as traffic flow rate, space mean density and velocity are calculated from the vehicle trajectories. We also focus on the vehicle platoon characteristics such as average and variance of the vehicle velocities in the platoon. Using the macroscopic measures, microscopic vehicle trajectories as well as vehicle platoon characteristics, the mechanism of traffic congestion occurrence and the possible countermeasures to relax the congestion are discussed. We have found that when the vehicle platoons are coming both from inside lanes of Moriguchi and Kanjo lines the velocity on the vehicle platoons from Moriguchi Line generally decreases. By this velocity reduction, it gets more difficult to merge from Moriguchi Line, and the congestion remains on the Moriguchi Line. It is therefore needed to carry out some measures to discourage using the inner lane of Kanio line to avoid conflict against Moriguchi Line, and to recover the velocity of Moriguchi line.