EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Vision‐Based Queue Estimation and Prediction for Ramp Metering: A Hybrid Computer Vision and Queuing Theory Approach

Sushant Tiwari, Sanjay Luitel, Abbas Rashidi, Nikola Marković

Journal of Advanced Transportation · 2026-01 · 2026 (1)

DOI: 10.1155/atr/2604474 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Advanced Transportation, online erschienen am 19.08.2026. Verfasst von Sushant Tiwari und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Utah Department of Transportation, Mountain-Plains Consortium.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Ramp metering is a critical tool for improving freeway efficiency, but its performance relies on real‐time queue‐state estimation. Conventional sensor‐based methods (e.g., inductive loops) often fail under congestion, leading to unreliable queue data. This study proposes a hybrid computer vision and queuing theory framework to overcome these limitations. The computer vision‐based framework integrates a segmentation model for ramp identification, object detection and tracking (YOLO + DeepSORT) for vehicle flow extraction, and a time‐varying queuing model to compute queue length and delay. To support proactive control, a random forest‐based predictor forecasts arrivals over a 10‐min horizon. Evaluated on four metered ramps in Utah under diverse lighting and seasonal conditions, the system achieved 93.0% accuracy for queue‐length estimation in clear conditions and 77%–87% under adverse weather conditions. The hybrid adaptive vision–queuing approach reduced annualized estimation error by 49% (queue length) and 30% (queuing time) compared to a vision‐only baseline. By leveraging existing surveillance infrastructure, this work provides a scalable, robust solution for real‐time ramp management, demonstrating the operational value of fusing vision‐based sensing with analytical models for proactive ramp management.

Förderung

  • Utah Department of Transportation
  • Mountain-Plains Consortium

Zitieren

APA

Tiwari, S., Luitel, S., Rashidi, A., & Marković, N. (2026). Vision‐Based Queue Estimation and Prediction for Ramp Metering: A Hybrid Computer Vision and Queuing Theory Approach. Journal of Advanced Transportation, 2026(1). https://doi.org/10.1155/atr/2604474

Vancouver

Tiwari S, Luitel S, Rashidi A, Marković N. Vision‐Based Queue Estimation and Prediction for Ramp Metering: A Hybrid Computer Vision and Queuing Theory Approach. Journal of Advanced Transportation. 2026;2026(1). doi:10.1155/atr/2604474

BibTeX
@article{Tiwari2026,
  author = {Tiwari, Sushant and Luitel, Sanjay and Rashidi, Abbas and Marković, Nikola},
  title = {Vision‐Based Queue Estimation and Prediction for Ramp Metering: A Hybrid Computer Vision and Queuing Theory Approach},
  journal = {Journal of Advanced Transportation},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/atr/2604474}
}
Kostenlos starten →