Incident-Guided Spatiotemporal Traffic Forecasting

2026-01-27

arXiv.org

AI Summary

Existing traffic prediction methods struggle to model the significant and non-stationary spatiotemporal disturbances caused by突发事件 such as accidents or adverse weather. To address this limitation, this work proposes the IGSTGNN framework, which explicitly captures the spatiotemporal impact mechanisms of such events on traffic flow for the first time. Specifically, an Incident Context Spatial Fusion (ICSF) module models the initial heterogeneous spatial effects, while a Temporal Impact Decay (TIID) module characterizes the dynamic dissipation process over time. The study also constructs and publicly releases the first large-scale dataset with precise alignment between incident records and traffic time series. The proposed modules are designed to be readily integrated into various backbone models, achieving state-of-the-art performance on this new benchmark and demonstrating both effectiveness and strong generalization capability.

Abstract

Recent years have witnessed the rapid development of deep-learning-based, graph-neural-network-based forecasting methods for modern intelligent transportation systems. However, most existing work focuses exclusively on capturing spatio-temporal dependencies from historical traffic data, while overlooking the fact that suddenly occurring transportation incidents, such as traffic accidents and adverse weather, serve as external disturbances that can substantially alter temporal patterns. We argue that this issue has become a major obstacle to modeling the dynamics of traffic systems and improving prediction accuracy, but the unpredictability of incidents makes it difficult to observe patterns from historical sequences. To address these challenges, this paper proposes a novel framework named the Incident-Guided Spatiotemporal Graph Neural Network (IGSTGNN). IGSTGNN explicitly models the incident's impact through two core components: an Incident-Context Spatial Fusion (ICSF) module to capture the initial heterogeneous spatial influence, and a Temporal Incident Impact Decay (TIID) module to model the subsequent dynamic dissipation. To facilitate research on the spatio-temporal impact of incidents on traffic flow, a large-scale dataset is constructed and released, featuring incident records that are time-aligned with traffic time series. On this new benchmark, the proposed IGSTGNN framework is demonstrated to achieve state-of-the-art performance. Furthermore, the generalizability of the ICSF and TIID modules is validated by integrating them into various existing models.

Problem

• traffic forecasting
• transportation incidents
• spatiotemporal dependencies
• external disturbances
• incident impact

Innovation

• Incident-Guided Forecasting
• Spatiotemporal Graph Neural Network
• Incident Impact Modeling
• Traffic Prediction
• External Disturbance