STGAN: Spatial-temporal Graph Autoregression Network for Pavement Distress Deterioration Prediction

Shilin Tong, Difei Wu, Xiaona Liu, Le Zheng, Yuchuan Du, Difan Zou

Abstract

Pavement distress significantly compromises road integrity and poses risks to drivers. Accurate prediction of pavement distress deterioration is essential for effective road management, cost reduction in maintenance, and improvement of traffic safety. However, real-world data on pavement distress is usually collected irregularly, resulting in uneven, asynchronous, and sparse spatial-temporal datasets. This hinders the application of existing spatial-temporal models, such as DCRNN, since they are only applicable to regularly and synchronously collected data. To overcome these challenges, we propose the Spatial-Temporal Graph Autoregression Network (STGAN), a novel graph neural network model designed for accurately predicting irregular pavement distress deterioration using complex spatial-temporal data. Specifically, STGAN integrates the temporal domain into the spatial domain, creating a larger graph where nodes are represented by spatial-temporal tuples and edges are formed based on a similarity-based connection mechanism. Furthermore, based on the constructed spatiotemporal graph, we formulate pavement distress deterioration prediction as a graph autoregression task, i.e., the graph size increases incrementally and the prediction is performed sequentially. This is accomplished by a novel spatial-temporal attention mechanism deployed by STGAN. Utilizing the ConTrack dataset, which contains pavement distress records collected from different locations in Shanghai, we demonstrate the superior performance of STGAN in capturing spatial-temporal correlations and addressing the aforementioned challenges. Experimental results further show that STGAN outperforms baseline models, and ablation studies confirm the effectiveness of its novel modules. Our findings contribute to promoting proactive road maintenance decision-making and ultimately enhancing road safety and resilience.

Introduction

Maintaining road infrastructure is a multi-billion dollar challenge. While we have sensors and cameras to detect cracks and potholes, the data they produce is rarely "clean." Most existing spatial-temporal AI models (like DCRNN) require data to be collected at regular intervals (e.g., every hour) and across all locations simultaneously.

In reality, pavement data is often collected via crowdsourcing—inspection vehicles or buses driving different routes at different times. This results in datasets that are irregular, asynchronous, and sparse. The Spatial-Temporal Graph Autoregression Network (STGAN) addresses the gap between these "messy" real-world datasets and the rigid requirements of standard deep learning models by treating time as a dimension within a unified graph.

The Challenge of Irregular Pavement Data

Crowdsourced inspection data is notoriously difficult to model because inspections do not happen on a fixed schedule. Location A might be checked on Monday, while Location B is checked on Wednesday, and many road segments have very few historical records. This asynchronicity makes traditional time-series forecasting models inapplicable.

Asynchronous and irregular data collection across different geographic locations in Shanghai.

Methodology: The STGAN Architecture

The core innovation of STGAN is how it structures the data. Instead of a static spatial graph that changes over time steps, the model proposes a unified spatial-temporal graph where each node represents a specific (location, time) tuple. If a road segment is inspected five times, it appears as five distinct nodes in the graph.

To handle the inherent sparsity of this data, STGAN utilizes two types of connection mechanisms:

  • Hard Connections: Links nodes that are physically close and close in time based on a fixed threshold.
  • TOP Connections: Ensures every node connects to at least its $K$ nearest neighbors, preventing "isolated" data points in sparse regions.
The STGAN unified spatial-temporal graph structure featuring hard and TOP connections.

Prediction is treated as an autoregressive task. To predict a future state, a "new" node is added to the graph, and the model uses a specialized spatial-temporal attention mechanism to aggregate information from its historical "neighbor" nodes. This mechanism recognizes that more recent data is generally more relevant for predicting future decay.

Experimental Results

The study utilized the ConTrack dataset, which covers over 350 km of urban roads and highways in Shanghai. The dataset includes 11 features across spatial, temporal, environmental (humidity, wind, precipitation), and road distress dimensions.

STGAN was compared against several baselines, including MLPs, GCNs, and Graph Attention Networks (GAT). STGAN consistently achieved the lowest error rates. For example, in the "0-2000" data segment, STGAN achieved a Mean Absolute Error (MAE) of 0.1759, compared to 0.4609 for the next-best model (GAT). When used to classify road health (Healthy, Good, Severe, Very Severe), the model achieved an AUC of 0.93 for the first three classes.

Conclusion and Future Work

By embedding the temporal domain directly into the spatial graph, STGAN bypasses the need for synchronized time steps, making it naturally resilient to irregular sampling. While the model currently struggles with predicting rare "Very Severe" cases due to data imbalance, it represents a significant advancement in pavement management. Future work will focus on integrating multi-scale data fusion, such as traffic load and maintenance records, to further improve long-term forecasting accuracy.

Summary updated

Abstract

Pavement distress significantly compromises road integrity and poses risks to drivers. Accurate prediction of pavement distress deterioration is essential for effective road management, cost reduction in maintenance, and improvement of traffic safety. However, real-world data on pavement distress is usually collected irregularly, resulting in uneven, asynchronous, and sparse spatial-temporal datasets. This hinders the application of existing spatial-temporal models, such as DCRNN, since they are only applicable to regularly and synchronously collected data. To overcome these challenges, we propose the Spatial-Temporal Graph Autoregression Network (STGAN), a novel graph neural network model designed for accurately predicting irregular pavement distress deterioration using complex spatial-temporal data. Specifically, STGAN integrates the temporal domain into the spatial domain, creating a larger graph where nodes are represented by spatial-temporal tuples and edges are formed based on a similarity-based connection mechanism. Furthermore, based on the constructed spatiotemporal graph, we formulate pavement distress deterioration prediction as a graph autoregression task, i.e., the graph size increases incrementally and the prediction is performed sequentially. This is accomplished by a novel spatial-temporal attention mechanism deployed by STGAN. Utilizing the ConTrack dataset, which contains pavement distress records collected from different locations in Shanghai, we demonstrate the superior performance of STGAN in capturing spatial-temporal correlations and addressing the aforementioned challenges. Experimental results further show that STGAN outperforms baseline models, and ablation studies confirm the effectiveness of its novel modules. Our findings contribute to promoting proactive road maintenance decision-making and ultimately enhancing road safety and resilience.
BibTeX
@article{tong2025stgan,
  author = {Shilin Tong and Difei Wu and Xiaona Liu and Le Zheng and Yuchuan Du and Difan Zou},
  title = {{STGAN: Spatial-temporal Graph Autoregression Network for Pavement Distress Deterioration Prediction}},
  journal = {IEEE Trans. Intell. Transp. Syst.},
  year = {2025},
  doi = {10.1109/tits.2025.3547883},
}
Figure