The invention provides a traffic prediction method based on graph sampling aggregation and a space-time attention mechanism. The accuracy of traffic state prediction can timely dispatch traffic vehicles to reduce the congestion rate, and traffic accidents are pre-judged to avoid traffic abnormal conditions. Although the current method obtains a good progress, most of the methods only pay attention to the time sequence of traffic data during modeling, and the long-term and short-term dependence in a modeling space and space information extraction of a large image are still insufficient. Therefore, in order to better extract the spatial information of a larger traffic network, the method utilizes graph sampling and aggregation acquisition spatial embedding. And a multi-head attention mechanism is utilized to fuse the time periodicity and the space embedding of the traffic data to obtain the spatial-temporal characteristics of the traffic data. In addition, the method also shows the long-term and short-term dependence of the spatial-temporal characteristics, so that the method has better capability of representing the real data spatial-temporal level, has better prediction capability on complex data, and is suitable for most traffic prediction scenes.

    本发明提供了一种基于图采样聚合与时空注意力机制的交通预测方法。交通状态预测的精确度可以及时调度交通车辆以降低拥堵率、预判交通事故以避免交通异常状况。虽然当前的方法取得很好的进展,但是这些方法多数在建模时只关注交通数据的时序性,在建模空间的长短期依赖性和提取大图的空间信息仍显不足。为此,为了更好的提取较大交通网络的空间信息,本发明利用图采样与聚合采集空间嵌入。利用多头注意力机制融合交通数据的时间周期性与空间嵌入得到交通数据的时空特征。除此以外,本发明也表现了时空特征的长短期依赖性,使其具有更好的表示真实数据时空层面的能力,在复杂数据上有更好的预测能力,适用于大多数交通预测场景。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Traffic prediction method based on graph sampling aggregation and space-time attention mechanism


    Weitere Titelangaben:

    基于图采样聚合与时空注意力机制的交通预测方法


    Beteiligte:
    LU ZIBAO (Autor:in) / AN CHEN (Autor:in) / DING ZIQIONG (Autor:in) / ZHANG BOTAO (Autor:in) / TANG GUANGLI (Autor:in) / ZHANG JIALI (Autor:in) / CAO HAO (Autor:in)

    Erscheinungsdatum :

    2024-03-29


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen



    Traffic flow prediction method based on space-time attention mechanism

    WANG JIAYING / YANG HENG / SHAN JING et al. | Europäisches Patentamt | 2024

    Freier Zugriff

    Road traffic safety prediction method based on dynamic graph attention space-time network

    JIA ZHAOHONG / YUAN JIE / YANG BO | Europäisches Patentamt | 2024

    Freier Zugriff

    Traffic flow prediction method and device based on time-space synchronization graph aggregation

    LYU YISHENG / WEI ZEBING / WANG FEIYUE | Europäisches Patentamt | 2022

    Freier Zugriff

    Traffic flow prediction method based on cyclic space-time attention mechanism

    XU XIAO / ZHANG LEI / LIU BAILONG et al. | Europäisches Patentamt | 2024

    Freier Zugriff

    Traffic flow prediction method and device based on space-time attention graph neural network

    HUANG WENBING / YUAN JIRUI / TIAN CHUJIE et al. | Europäisches Patentamt | 2024

    Freier Zugriff