Traffic congestion remains a critical issue worldwide due to the increasing number of vehicles and the limitations of fixed infrastructure. This study proposes a dynamic traffic signal management system leveraging deep learning techniques to optimize the traffic flow. The system integrates YOLO for accurate real-time vehicle detection and counting, enabling adaptive signal timing adjustments based on traffic density at intersections. Comparative analysis with R-CNN and SSD models has identified YOLO as the optimal choice, balancing speed and precision. Real-time data processing, dynamically extends green light durations for high-density lanes, minimizing vehicle wait time and improving overall flow efficiency. The proposed method utilizes standard traffic cameras and OpenCV for image processing, offering a cost-effective and scalable alternative to infrastructure expansion. Simulation results demonstrate significant reductions in vehicle wait times and emissions, emphasizing the system's environmental and practical benefits. This approach not only enhances traffic management but also aligns with urban sustainability goals by reducing fuel consumption and improving air quality. Designed for adaptability, the system accommodates varying traffic conditions, making it a viable solution for diverse urban settings. Future developments integrating intelligent infrastructure and autonomous vehicles could further revolutionize urban mobility.
Dynamic Signal for Traffic Congestion Using CNN
25.06.2025
533124 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Protecting buses from congestion using traffic signal control
British Library Conference Proceedings | 1995
|Protecting buses from congestion using traffic signal control
IET Digital Library Archive | 1995
|Europäisches Patentamt | 2022
|