Forecasting traffic has been considered as the foundation for many applications such as traffic control, trip planning, and vehicle routing in intelligent transportation system. It is typically a prediction of traffic speed or flow from time-series data. It can be modeled as predicting traffic in next n time steps given the previous k traffic observations. Different forecasting methods proposed in the literature are model-driven and data-driven such as deep learning models, machine learning models, and statistical methods. Timely precise traffic forecast is critical for traffic control and guidance. Traditional methods fail in precise prediction due to high complexity and nonlinearity of traffic data and negligence of temporal and spatial dependencies available in the data. Graph neural networks (GNNs) emerged recently as state-of-the-art methods to forecast traffic as they are better suitable for traffic forecasting systems with graph models. Spatiotemporal graph modeling is very important task for analysis of spatial relations and temporal dependencies to forecast traffic flow. The objectives of this paper are three fold to discuss; graph construction from spatiotemporal traffic data for GNNs, review of spatiotemporal GNN models to forecast traffic, comparative analysis of performance of the models in predicting traffic flow ahead of 15, 30, and 45 min on different datasets. Comparative analysis of various neural network-based models used in traffic prediction on different datasets highlighted the strengths and limitations of each model.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Spatiotemporal Graph Neural Networks for Traffic Forecasting: A Comparative Analysis


    Additional title:

    Lect. Notes Electrical Eng.



    Conference:

    International Conference on Advanced Engineering Optimization Through Intelligent Techniques ; 2023 ; Surat, India September 28, 2023 - September 30, 2023



    Publication date :

    2024-10-15


    Size :

    13 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English





    Multiadaptive Spatiotemporal Flow Graph Neural Network for Traffic Speed Forecasting

    Xu, Yaobin / Liu, Weitang / Mao, Tingyun et al. | Transportation Research Record | 2022


    GSTGAT: Gated spatiotemporal graph attention network for traffic demand forecasting

    Shuilin Yao / Huizhen Zhang / Chenxi Wang et al. | DOAJ | 2024

    Free access

    GSTGAT: Gated spatiotemporal graph attention network for traffic demand forecasting

    Yao, Shuilin / Zhang, Huizhen / Wang, Chenxi et al. | Wiley | 2024

    Free access