A Comparative Review of AI-Based Approaches for Real-Time Wrong-Way Vehicle Detection in Intelligent Highway Monitoring Systems

DOI

10.2991/978-94-6239-789-7_35How to use a DOI?

Keywords

Wrong-Way Vehicle Detection; YOLOv10; Detection Transformer (DETR); Faster R-CNN; Multi-Object Tracking; ByteTrack; Trajectory Analysis; Real-Time Traffic Monitoring; UA-DETRAC

Abstract

Wrong-way driving represents one of the most vital traffic offenses that considerably increase the frequency of severe accidents and road fatalities. Thus, efficient real-time object detection techniques should be implemented within intelligent traffic environments to detect any cases of wrong-way driving. In this work, a comparative analysis of three recent state-of-the-art object detection models, namely, YOLOv10, Detection Transformer (DETR), and Faster R-CNN, will be presented in relation to real-time vehicle detection based on the usage of the UA-DETRAC dataset.

For a unified and consistent assessment, a combined system of vehicle detection, multi-object tracking, and trajectory-based analysis of direction will be considered. According to the experimental findings, YOLOv10 is found to provide the best balance of accuracy and efficiency, delivering mAP of 91.3%, achieving real-time performance of 42 frames per second, and demonstrating minimal latency of 23 milliseconds. Although DETR demonstrates the highest performance in terms of detection accuracy (93.1% mAP) and tracking (90.2% MOTA) due to global contextual modeling, it requires more computational power with 12 frames per second being achieved along with 83 milliseconds of latency. Finally, Faster R-CNN demonstrates average performance of 89.6% mAP and 18 frames per second.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Sagar Choudhary
AU  - Brij Mohan Singh
PY  - 2026
DA  - 2026/10/08
TI  - A Comparative Review of AI-Based Approaches for Real-Time Wrong-Way Vehicle Detection in Intelligent Highway Monitoring Systems
BT  - Proceedings of theInternational Conference on Innovation and Practices in Artificial Intelligence (ICIPAI 2026)
PB  - Atlantis Press
SP  - 507
EP  - 524
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-789-7_35
DO  - 10.2991/978-94-6239-789-7_35
ID  - Choudhary2026
ER  -