Innovative Method for Vehicle Queue Estimation at Traffic Signals Using Roadside Sensors

Abstract

Rapidly escalating urban traffic congestion demands intelligent, sensor-driven approaches to traffic signal management. Conventional fixed-timer traffic signals allocate uniform phase durations irrespective of prevailing vehicle density, generating unnecessarily prolonged queues, excessive fuel consumption, and heightened greenhouse-gas emissions. This paper presents a structured survey of ten contemporary research works spanning 2019 to 2026 that address vehicle queue estimation and adaptive signal control at urban intersections through roadside sensing, computer vision, and artificial intelligence. The surveyed studies encompass YOLO-based deep learning detection frameworks, fuzzy logic inference controllers, reinforcement learning optimization agents, and hybrid IoT-integrated architectures. Each work is reviewed for its methodological novelty, performance outcomes, simulation environments, and operational limitations. A comparative analysis table consolidates key metrics across studies, followed by identification of prevalent research gaps including adverse-weather robustness, edge-deployment scalability, multi-intersection coordination, and emergency-vehicle preemption. The survey provides a foundational reference for engineers and researchers engaged in next-generation adaptive traffic management system design.

Introduction

Traffic congestion is a major urban problem caused by inefficient management of signalised intersections, where fixed-time traffic lights fail to adapt to real-time variations in vehicle flow. Since traffic demand fluctuates due to peak hours, events, and random changes, accurate vehicle queue estimation is essential for enabling adaptive traffic signal control systems that respond dynamically to actual road conditions. Modern approaches increasingly replace traditional inductive loop sensors with camera-based systems using deep learning models like YOLO, which can detect and count vehicles in real time under diverse conditions.

The study reviews methods for vehicle detection, queue estimation, and adaptive signal control. It highlights CNN-based techniques (especially YOLO variants) as highly effective for real-time detection, along with methods such as density-based and ROI-based queue estimation. It also discusses control strategies including reinforcement learning, fuzzy logic, and hybrid AI systems, which improve traffic efficiency by adjusting signal timings based on live data.

Several recent studies are compared, showing that YOLO-based and AI-driven systems significantly reduce waiting times and congestion compared to fixed-time signals. However, challenges remain, including sensitivity to poor weather, high computational requirements, lack of real-world deployment, and difficulties in scaling adaptive systems across complex intersections.

Conclusion

This paper has surveyed ten contemporary research contributions addressing vehicle queue estimation and adaptive traffic signal control at urban intersections, drawing on works published between 2019 and 2026. The reviewed studies collectively demonstrate that computer-vision-based queue detection—particularly through YOLO deep learning models—combined with intelligent control algorithms, can deliver meaningful improvements in intersection performance metrics including reduced vehicle waiting time, lower queue lengths, and improved throughput relative to conventional fixed-time signal operation. Deep learning-based vehicle detection has emerged as a versatile, hardware-light approach that leverages existing surveillance infrastructure without requiring intrusive pavement sensors. Reinforcement learning provides a principled optimisation framework capable of learning complex multi-phase control policies through simulation experience. Fuzzy logic controllers offer interpretable, low-complexity adaptive mechanisms particularly suited to deployments with limited computational resources. The proposed project—an innovative method for vehicle queue estimation at traffic signals using roadside sensors—aligns directly with the frontiers identified in this survey: combining YOLO-based real-time detection with ESP32 microcontroller-based signal switching to deliver a cost-effective, hardware-accessible adaptive control solution. Key future enhancement directions include adverse-weather robustness, edge-optimised model deployment, multi-intersection coordination, and integration of emergency vehicle preemption. The survey establishes a comprehensive bibliographic and analytical foundation supporting the continued development of intelligent adaptive traffic management systems.

References

[1] S. Nandhini, M. R. Magesh Kumar, M. Mownika, K. Ranganayagi, and R. J. Pavithra, \"Smart Traffic Surveillance System with Adaptive Traffic Control Signal using YOLO,\" Int. J. Eng. Res. Technol. (IJERT), vol. 13, no. 5, 2025. [2] H. Jain, S. Thete, R. Shahare, and P. Kumbhare, \"Computer Vision-Based Adaptive Traffic Signal Control System,\" Int. J. Comput. Tech. (IJCT), vol. 13, no. 1, pp. 324–328, Jan.–Feb. 2026. [3] S. Kamble, M. Patil, and R. Deshmukh, \"AI-Driven Smart Traffic Management System: An Adaptive Approach Using YOLO and OpenCV,\" ResearchGate Publication, 2025. [4] N. Fahrunnisa, A. Setiawan, and B. Pratama, \"Adaptive Traffic Light Signal Control Using Fuzzy Logic,\" J. Inf. Technol. Electr. Eng. (JITEKI), 2024. [5] G. Michailidis, K. Papageorgiou, and E. Kosmatopoulos, \"Traffic Signal Control via Reinforcement Learning: A Review,\" MDPI Electronics, vol. 14, no. 3, 2025. [6] A. Agrahari, M. M. Dhabu, P. S. Deshpande, A. Tiwari, M. A. Baig, and A. D. Sawarkar, \"Artificial Intelligence-Based Adaptive Traffic Signal Control System: A Comprehensive Review,\" Electronics, vol. 13, no. 19, p. 3875, Sep. 2024. [7] Y. Xiao, L. Zhang, and H. Wang, \"Intelligent Traffic Signal Control Based on Reinforcement Learning,\" Artif. Intell. Rev., Springer, 2026. [8] A. Sakhare, P. Kulkarni, and S. Patil, \"Smart Traffic: Integrating Machine Learning and YOLO for Efficient Traffic Management,\" Int. J. Intell. Syst. Appl. Eng. (IJISAE), 2023. [9] C. Wang, B. Zheng, and C. Li, \"Efficient Traffic Sign Recognition using YOLO for Intelligent Transport Systems,\" Sci. Rep., vol. 15, p. 13657, Apr. 2025. [10] X H. Wei, G. Zheng, V. Gayah, and Z. Li, \"A Survey on Traffic Signal Control Methods,\" arXiv:1904.08117, Jul. 2019.

Copyright

Copyright © 2026 Sheela Rani C M, Chalitha MN Jain, Kruthika TJ, Hithashree G, Mahalakshmi M Ratnakar. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.