STMGCN: A Spatiotemporal Multi-Graph Convolutional Networks for Real-Time Intelligent Transportation Traffic Flow Prediction

The Internet of Things (IoT) and artificial intelligence (AI) are essential for intelligent transportation systems (ITS) because they allow for real-time decision-making, predictive capabilities, and data-driven optimization. With the use of real-time analytics and predictive modeling, artificial intelligence (AI) evaluates the massive volumes of real-time data that IoT devices gather from sensors and automobiles to optimize routes, improve traffic management, increase safety, and enable autonomous driving. This study introduces a spatiotemporal multi-graph convolutional network (STMGCN) for traffic flow prediction. The framework constructs dynamic and static urban traffic networks by leveraging multi-source data, including road checkpoints, meteorological conditions, and points of interest. A relational evolving graph convolutional network is employed for knowledge embedding, while a knowledge fusion module integrates traffic representations with real-time traffic flow data. The fused embeddings are processed through a spatiotemporal multi-graph convolutional module to capture complex dependencies across multiple semantic topology graphs. Experimental validation using a real-world Hangzhou traffic dataset demonstrates that STMGCN outperforms state-of-the-art baselines by 5.76%–10.71%, while robustness tests confirm its resilience to noise and disruptions. This research highlights the potential of virtual, adaptive models that enhance predictive accuracy, scalability, and reliability in IoT for next-generation ITS applications.