Discovering the Precursors of Traffic Breakdowns Using Spatiotemporal Graph Attribution Networks

2025-04-23

AI Summary

This study addresses early-warning for abrupt traffic flow deterioration by identifying and interpreting physically meaningful precursors to traffic breakdown. We propose a novel framework integrating spatiotemporal graph neural networks (ST-GNNs) with graph-structured Shapley value attribution—marking the first extension of Shapley-based interpretability to dynamic spatiotemporal graph modeling. Our method enables precise alignment between black-box predictions and physical causative factors, such as road network topological bottlenecks and localized emergency braking events. Leveraging multi-source floating-car data, we quantitatively validate the framework on the I-24 field dataset: topological bottlenecks and localized emergency braking emerge as the two dominant, interpretable precursors, with attribution results demonstrating high consistency and empirical verifiability. This work establishes a new paradigm for explainable intelligent traffic warning systems, bridging predictive accuracy with actionable, physics-grounded insights.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Understanding and predicting the precursors of traffic breakdowns is critical for improving road safety and traffic flow management. This paper presents a novel approach combining spatiotemporal graph neural networks (ST-GNNs) with Shapley values to identify and interpret traffic breakdown precursors. By extending Shapley explanation methods to a spatiotemporal setting, our proposed method bridges the gap between black-box neural network predictions and interpretable causes. We demonstrate the method on the Interstate-24 data, and identify that road topology and abrupt braking are major factors that lead to traffic breakdowns.

Problem

• Identify precursors of traffic breakdowns for road safety
• Interpret traffic breakdown causes using ST-GNNs and Shapley values
• Analyze road topology and abrupt braking as key factors

Innovation

• ST-GNNs for traffic breakdown precursor detection
• Shapley values for spatiotemporal interpretation
• Identifies road topology and braking factors