The invention discloses a traffic speed prediction method based on multi-graph cross attention fusion, and the method comprises the following steps: 1, constructing multiple space-time diagrams in the form of an adjacent matrix according to the topological structure of a directed road network, the multiple space-time diagrams comprising a distance diagram, a directional diagram and a time correlation diagram; 2, adopting a multi-layer spatial-temporal feature extractor which comprises a multi-layer time feature extractor and a multi-layer space-time diagram feature extractor, and extracting time features and space-time diagram features from the road signal sequence, the distance diagram, the directional diagram and the time correlation diagram by the multi-layer spatial-temporal feature extractor; and step 3, splicing the time features and the space-time diagram features and then inputting the spliced features to a full connection layer, and outputting a traffic speed prediction result by the full connection layer. According to the method, the spatial and temporal correlation among the representation nodes is defined, the information loss caused by the deep network is reduced, and the accuracy of traffic speed prediction is improved.

    本发明公开了一种基于多图交叉注意力融合的交通速度预测方法,包括以下步骤:步骤1、根据有向道路网络的拓扑结构构建邻接矩阵形式的多重时空图,多重时空图包括距离图、方向图和时间相关性图;步骤2、采用多层时空特征提取器,多层时空特征提取器包括多层时间特征提取器、多层时空图特征提取器,由多层时空特征提取器从道路信号序列以及距离图、方向图和时间相关性图中提取时间特征、时空图特征;步骤3、将时间特征、时空图特征进行拼接后输入至全连接层,由全连接层输出交通速度预测结果。本发明明确了表示节点之间的空间和时间相关性,减少了深度网络导致的信息损失,有利于提高交通速度预测的准确性。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Traffic speed prediction method based on multi-graph cross attention fusion


    Weitere Titelangaben:

    基于多图交叉注意力融合的交通速度预测方法


    Beteiligte:
    MA TIAN (Autor:in) / CHEN YUE (Autor:in) / REN YILONG (Autor:in)

    Erscheinungsdatum :

    2023-12-01


    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



    High-speed traffic flow prediction method based on multi-modal fusion and graph attention mechanism

    RUAN TAO / WU DEXING / XU LEI | Europäisches Patentamt | 2020

    Freier Zugriff

    Asymmetric Long-Term Graph Multi-Attention Network for Traffic Speed Prediction

    Hwang, Jiyoung / Noh, Byeongjoon / Jin, Zhixiong et al. | IEEE | 2022




    Traffic flow prediction method of space-time attention graph convolutional network based on multi-feature fusion

    CHEN YAJUN / DING ZHIMING / GUO LIMIN | Europäisches Patentamt | 2023

    Freier Zugriff