Network-Wide Traffic Volume Estimation from Speed Profiles using a Spatio-Temporal Graph Neural Network with Directed Spatial Attention

2025-12-15

AI Summary

Accurately estimating full-network traffic flow in sensor-sparse urban environments remains challenging due to the scarcity of real-time traffic measurements. Method: This paper proposes an inductive, network-level estimation framework that requires no real-time flow data. We design a Hybrid Directed Attention Spatio-Temporal Graph Neural Network (HDA-STGNN), which introduces a novel directed spatial attention mechanism to explicitly model directional traffic propagation. The framework jointly learns spatio-temporal dependencies via graph-structured modeling, decouples spatial and temporal dynamics, and integrates topology-aware feature fusion for end-to-end daily traffic volume profile prediction—using only floating-car speeds, static road attributes, and network topology. Contribution/Results: Our method achieves significant performance gains over state-of-the-art baselines on real-world city-scale datasets. Ablation studies confirm the critical roles of directed attention and topological information. Moreover, the framework supports zero-shot cross-region generalization, demonstrating strong transferability across heterogeneous urban areas.

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: Graph neural networks and deep learning approaches for Web-related graphs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

Existing traffic volume estimation methods typically address either forecasting traffic on sensor-equipped roads or spatially imputing missing volumes using nearby sensors. While forecasting models generally disregard unmonitored roads by design, spatial imputation methods explicitly address network-wide estimation; yet this approach relies on volume data at inference time, limiting its applicability in sensor-scarce cities. Unlike traffic volume data, probe vehicle speeds and static road attributes are more broadly accessible and support full coverage of road segments in most urban networks. In this work, we present the Hybrid Directed-Attention Spatio-Temporal Graph Neural Network (HDA-STGNN), an inductive deep learning framework designed to tackle the network-wide volume estimation problem. Our approach leverages speed profiles, static road attributes, and road network topology to predict daily traffic volume profiles across all road segments in the network. To evaluate the effectiveness of our approach, we perform extensive ablation studies that demonstrate the model's capacity to capture complex spatio-temporal dependencies and highlight the value of topological information for accurate network-wide traffic volume estimation without relying on volume data at inference time.

Problem

• Estimates network-wide traffic volumes from speed data
• Uses spatio-temporal graph neural networks with attention
• Predicts daily traffic without relying on volume sensors

Innovation

• Leverages speed profiles and road attributes for estimation
• Uses spatio-temporal graph neural network with directed attention
• Predicts traffic volumes without volume data at inference