Research Reports
Report Details
Final Report Abstract
Understanding ramp queue length and queuing time is important for transportation agencies to manage and operate the ramps with optimum performance. Since these data are collected with conventional sensor systems such as coils, they are prone to error, especially during traffic congestion. The increased deployment of cameras and recent advancements in artificial intelligence, such as deep learning and computer vision, provides an opportunity to employ traffic surveillance camera videos for ramp management. This study employed four location surveillance video cameras to develop and evaluate the framework developed using the object detection and tracking algorithms. This framework uses existing video cameras as input to the framework and determines the queueing parameters of highways on ramps, such as queue length and queuing time, which provide important information to freeway management teams to optimize signal timing. Additionally, this study provides a detailed implementation plan for computer vision and optimum location of the camera installation and hardware requirements.
How to Cite
Tiwari, Sushant, Abbas Rashidi, and Nikola Markovic. Evaluating Different Methods for Estimating Queue Length on Access Ramps, MPC-23-507. North Dakota State University - Upper Great Plains Transportation Institute, Fargo: Mountain-Plains Consortium, 2023.