Real time traffic management using convolutional neural networks in smart city IoT frameworks

Résumé

Rapid urbanization and increasing vehicular density have led to severe traffic congestion, excessive fuel consumption, and increased pollution levels in modern cities. Traditional traffic management systems lack real-time adaptability and often fail to optimize traffic flow effectively, resulting in significant delays and inefficiencies. To address these challenges, this study presents a Convolutional Neural Network (CNN)-based real-time traffic management system integrated into an Internet of Things (IoT) framework. The proposed approach utilizes real-time video surveillance and sensor data, processed by CNN models for vehicle detection, traffic density estimation, and anomaly detection. The system dynamically adjusts traffic signals and optimizes vehicle movement using edge computing to minimize latency. Experimental results demonstrate a 29.4% reduction in congestion, a 35.7% improvement in vehicle detection accuracy compared to traditional systems, and a 21.8% enhancement in traffic signal efficiency. Additionally, fuel consumption and carbon emissions were reduced by 18.2% and 15.9%, respectively. These findings highlight the effectiveness of AI-driven traffic control in improving urban mobility, reducing environmental impact, and enhancing road safety. The integration of CNNs with IoT-based infrastructure provides a scalable and adaptive solution for smart city transportation networks.