Liu, Yilin (Beijing university of posts and telecommunications), Sartoretti, Guillaume (National University of Singapore)

Ride the Flow: Dynamic Tensor and Adaptive Modeling for Short-Term Traffic Prediction

Scheduled for presentation during the Regular Session "S30a-Intelligent Modeling and Prediction of Traffic Dynamics" (TH-LM-T30), Thursday, November 20, 2025, 11:30−11:50, Gold Coast

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 18, 2025

Abstract

Accurate short-term traffic flow prediction plays a pivotal role in enabling reliable and intelligent decision-making for automated transportation systems. However, the inherent complexity of spatiotemporal traffic patterns and the frequent presence of missing data present critical challenges to prediction accuracy and system resilience. In this paper, we present a dynamic tensor-based forecasting framework tailored for intelligent transportation scenarios. We first employ a hierarchical clustering strategy to identify key intersections with strong interdependencies, enhancing the modeling of spatial correlations. Subsequently, we introduce a dynamic “Location–X–Time” tensor structure that captures multi-scale temporal dependencies and adapts in real-time through a sliding window mechanism, ensuring robustness to data incompleteness. Our approach improves the reliability of downstream applications such as adaptive traffic control and vehicle-infrastructure coordination. Experimental results on large-scale real-world datasets demonstrate that our proposed method consistently outperforms state-of-the-art models, supporting the advancement of secure and intelligent urban mobility systems.