hia-gat: A Heterogeneous Interaction-Aware Graph Attention Network For Frame-Level Traffic Conflict Risk Prediction On Freeways

2026-06-25

AI Summary

This study addresses the real-time prediction of trajectory-based traffic conflict risk in highway scenarios by formulating it as a binary classification task on a multi-agent scene graph. The authors propose a dual-stream Heterogeneous Interaction-Aware Graph Attention Network (HIA-GAT) that constructs a heterogeneous relation graph integrating longitudinal interactions within lanes and lateral interactions across adjacent lanes. A conflict-type-aware gating mechanism is introduced to enable interpretable fusion of these two interaction types and attribute the dominant conflict category. The method leverages physics-informed edge features, time-to-collision (TTC) and post-encroachment time (PET) conflict metrics, and event-level gate supervision. Evaluated on the NGSIM I-80 and US-101 datasets, it achieves state-of-the-art average AUC scores of 0.835 and 0.867, respectively, with particularly notable performance gains in PET-based lateral conflict scenarios, underscoring the critical role of graph structure in modeling lateral conflicts.

Abstract

This paper formulates frame-level freeway risk assessment as a multi-agent scene graph-level binary classification problem, where each video or trajectory frame is labeled risky if any TTC- or PET-based conflict violates a specified severity threshold. We construct a relation-aware graph per frame with vehicles as nodes and two interaction types as edges: same-lane (longitudinal) and adjacent-lane (lateral), augmented with physics-informed edge features aligned to rear-end and lane-change conflict mechanisms. Building on a structured benchmarking suite of non-graph models and graph baselines, we propose HIA-GAT, a dual-stream heterogeneous graph attention network that processes longitudinal and lateral interactions through dedicated attention pathways and fuses them via a conflict-type-aware gating mechanism with event-level gate supervision derived from SSM conflict attribution. Experiments on the NGSIM I-80 and US-101 freeway datasets across nine TTC and PET threshold configurations show that HIA-GAT achieves the best average risk-ranking performance (AUC 0.835 on I-80 and 0.867 on US-101), with the largest gains on PET-only (lane-change) settings where relational structure is essential. Beyond accuracy, the learned gate provides interpretable per-vehicle attribution of dominant conflict type, supporting actionable, real-time freeway safety monitoring. We show that graph structure is critical for modeling lateral conflict risk, while longitudinal risk can often be captured by non-relational aggregation.

Problem

• traffic conflict risk prediction
• frame-level risk assessment
• freeway safety
• multi-agent scene graph
• heterogeneous interactions

Innovation

• heterogeneous graph attention network
• traffic conflict prediction
• relation-aware scene graph
• conflict-type-aware gating
• frame-level risk assessment