Truck Parking Usage Prediction with Decomposed Graph Neural Networks

2024-01-23

arXiv.org

AI Summary

Imbalanced supply and demand for truck parking spaces along freight corridors—coupled with unsafe illegal parking—poses significant safety risks; existing single-point forecasting methods fail to capture spatiotemporal dependencies across geographically distributed parking facilities, primarily due to insufficient multi-source collaborative data. Method: This paper proposes the first statewide multi-site collaborative forecasting framework for truck parking occupancy rates, introducing a Regional Decomposition strategy for data partitioning and a novel RegT-GCN spatiotemporal graph neural network that explicitly models geographic distribution characteristics, spatial topological relationships, and temporal dynamics. Contribution/Results: Evaluated on a real-world statewide dataset, the proposed method achieves over 20% improvement in prediction accuracy compared to state-of-the-art baseline models, thereby enabling more effective dynamic allocation of parking resources and data-informed safety regulation decisions.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Application Domains: Transportation
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Security and Privacy: Data transparency and provenance

Abstract

Truck parking on freight corridors faces the major challenge of insufficient parking spaces. This is exacerbated by the Hour-of-Service (HOS) regulations, which often result in unauthorized parking practices, causing safety concerns. It has been shown that providing accurate parking usage prediction can be a cost-effective solution to reduce unsafe parking practices. In light of this, existing studies have developed various methods to predict the usage of a truck parking site and have demonstrated satisfactory accuracy. However, these studies focus on a single parking site, and few approaches have been proposed to predict the usage of multiple truck parking sites considering spatio-temporal dependencies, due to the lack of data. This paper aims to fill this gap and presents the Regional Temporal Graph Neural Network (RegT-GCN) to predict parking usage across the entire state to provide more comprehensive truck parking information. The framework leverages the topological structures of truck parking site locations and historical parking data to predict the occupancy rate considering spatio-temporal dependencies across a state. To achieve this, we introduce a Regional Decomposition approach, which effectively captures the geographical characteristics of the truck parking locations and their spatial correlations. Evaluation results demonstrate that the proposed model outperforms other baseline models, improving performance by more than 20%.

Problem

• Predict truck parking usage across multiple sites
• Address spatio-temporal dependencies in parking data
• Improve accuracy with Regional Temporal Graph Convolutional Network

Innovation

• Uses decomposed graph neural networks
• Predicts parking usage across regions
• Captures spatio-temporal dependencies effectively