The invention provides a construction method of a traffic flow prediction model based on a space-time multi-scale graph convolutional network and a traffic flow prediction method, and specifically, a fine-grained traffic graph is given, and a coarse-grained traffic graph is generated through spectral clustering; then, respectively extracting the spatial-temporal correlation of the fine-grained traffic map and the coarse-grained traffic map, and fully mining the spatial correlation, including static invariant features of the region and dynamic spatial correlation; secondly, in order to relieve the negative influence of traffic flow fluctuation in the fine-grained traffic map, information diffusion between the fine-grained traffic map and the coarse-grained traffic map is realized by adopting cross-scale fusion; and finally, performing an experiment on the two traffic data sets in the real world, selecting an optimal parameter, and determining a final result. According to the method, the existing traffic flow prediction model is improved, so that the model can utilize the characteristics in the coarse-grained traffic map, the negative influence of traffic flow fluctuation in the fine-grained traffic map is relieved, and the prediction precision is improved to a certain extent.

    本发明提供了一种基于时空多尺度图卷积网络的交通流预测模型的构建方法以及交通流预测方法,具体来说,给定一个细粒度交通图,首先通过谱聚类生成一个粗粒度交通图;然后分别提取细粒度和粗粒度交通图的时空相关性,本发明充分挖掘了空间相关性,包括区域的静态不变特征和动态的空间相关性;其次,为了缓解细粒度交通图中交通流波动的负面影响,采用跨尺度融合来实现细粒度交通图和粗粒度交通图之间的信息扩散;最后,在两个真实世界的交通数据集上进行实验,选择最优参数,确定最终的结果。本发明改进了现有的交通流预测模型,使得模型可以利用粗粒度交通图中的特征,缓解细粒度交通图中交通流波动的负面影响,一定程度上提高了预测精度。


    Access

    Download


    Export, share and cite



    Title :

    Construction method of traffic flow prediction model based on space-time multi-scale graph convolutional network and traffic flow prediction method


    Additional title:

    一种基于时空多尺度图卷积网络的交通流预测模型的构建方法以及交通流预测方法


    Contributors:
    WU LIBING (author) / CAO SHUQIN (author) / ZHANG RUI (author) / WANG MIN (author) / ZHANG ZHUANGZHUANG (author)

    Publication date :

    2022-06-10


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Multi-scale traffic flow prediction method based on graph convolutional neural network

    ZHANG MEIYUE / WANG SENZHANG / MIAO HAO et al. | European Patent Office | 2021

    Free access


    Multi-view fusion space-time dynamic graph convolutional network urban traffic flow prediction method

    YUAN GUAN / ZHAO WENZHU / ZHANG YANMEI et al. | European Patent Office | 2023

    Free access

    Temporal Multi-Graph Convolutional Network for Traffic Flow Prediction

    Lv, Mingqi / Hong, Zhaoxiong / Chen, Ling et al. | IEEE | 2021


    Dynamic multi-graph convolutional network traffic flow prediction method based on graph transformation

    HU YONGLI / PENG TING / GUO KAN et al. | European Patent Office | 2022

    Free access