Network-wide Freeway Traffic Estimation Using Sparse Sensor Data: A Dirichlet Graph Auto-Encoder Approach

2025-03-20

AI Summary

Estimating full-network traffic states on highways with sparse sensor deployments remains challenging—particularly due to zero-filling bias on sensor-free road segments, difficulty in modeling directed graph topology, and capturing mode-dependent (e.g., congested vs. free-flow) propagation dynamics. Method: We propose the Directed Graph Autoencoder (DGAE), a Dirichlet-based graph neural architecture explicitly designed for directed traffic networks. Contribution/Results: DGAE introduces (1) Directed Dirichlet Energy Feature Propagation (DEFP4D), a theoretically grounded, physics-informed mechanism integrating dual-mode (congestion/free-flow) propagation; and (2) zero-filling–free unsupervised cross-city transferability. Evaluated on three real-world highway datasets, DGAE consistently outperforms state-of-the-art methods—even at only 5% sensor coverage—while demonstrating exceptional generalization across diverse topologies and traffic dynamics, confirming strong robustness to both structural and dynamical variations.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Planning, Routing, and Scheduling: Model-Based Reasoning
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Network-wide Traffic State Estimation (TSE), which aims to infer a complete image of network traffic states with sparsely deployed sensors, plays a vital role in intelligent transportation systems. With the development of data-driven methods, traffic dynamics modeling has advanced significantly. However, TSE poses fundamental challenges for data-driven approaches, since historical patterns cannot be learned locally at sensor-free segments. Although inductive graph learning shows promise in estimating states at locations without sensor, existing methods typically handle unobserved locations by filling them with zeros, introducing bias to the sensitive graph message propagation. The recently proposed Dirichlet Energy-based Feature Propagation (DEFP) method achieves State-Of-The-Art (SOTA) performance in unobserved node classification by eliminating the need for zero-filling. However, applying it to TSE faces three key challenges: inability to handle directed traffic networks, strong assumptions in traffic spatial correlation modeling, and overlooks distinct propagation rules of different patterns (e.g., congestion and free flow). We propose DGAE, a novel inductive graph representation model that addresses these challenges through theoretically derived DEFP for Directed graph (DEFP4D), enhanced spatial representation learning via DEFP4D-guided latent space encoding, and physics-guided propagation mechanisms that separately handles congested and free-flow patterns. Experiments on three traffic datasets demonstrate that DGAE outperforms existing SOTA methods and exhibits strong cross-city transferability. Furthermore, DEFP4D can serve as a standalone lightweight solution, showing superior performance under extremely sparse sensor conditions.

Problem

• Estimates network-wide traffic states using sparse sensor data.
• Addresses challenges in traffic spatial correlation modeling.
• Handles distinct propagation rules for congestion and free flow.

Innovation

• Dirichlet Graph Auto-Encoder for traffic estimation
• DEFP4D handles directed traffic networks effectively
• Physics-guided propagation for congestion and free-flow