Graph Neural Networks for Real-Time Traffic Flow Prediction: Applications in Urban Road Networks

Authors- Assistant Professor Dr. Pankaj Malik, Atharva Sharma, Aditya Thakur, Kavish Jha, Pranjal Patidar

Abstract-Accurate and real-time traffic flow prediction is essential for effective urban road network management, minimizing congestion, and optimizing transportation systems. Traditional traffic prediction methods struggle to capture the complex spatiotemporal dependencies inherent in urban traffic data. This paper proposes a novel approach leveraging Graph Neural Networks (GNNs) to address these challenges. GNNs are well-suited for modeling traffic networks due to their ability to handle graph-structured data, where road intersections are represented as nodes and road segments as edges. The proposed framework integrates dynamic graph construction, temporal attention mechanisms, and adaptive learning to model the evolving nature of urban traffic patterns. Real-world traffic datasets are used to validate the framework, demonstrating its superiority in prediction accuracy, scalability, and robustness compared to baseline models. The results indicate that GNN-based models can effectively capture both short-term and long-term dependencies, providing actionable insights for traffic control systems, urban planners, and smart city applications. This research highlights the transformative potential of advanced machine learning techniques in tackling real-world traffic management challenges.

DOI: /10.61463/ijset.vol.12.issue6.402