The invention relates to a short-time traffic flow prediction method based on a graph convolution recurrent neural network, which comprises the following steps: firstly, obtaining road network information, converting the road network information into a road network graph which is composed of a plurality of road nodes and edges connected with the road nodes; then, a short-time traffic flow prediction model is constructed, and the short-time traffic flow prediction model predicts the traffic flows of a plurality of time slices in the future according to the road network map and the traffic flow characteristics of a plurality of historical time slices by learning a mapping function; the short-time traffic flow prediction model comprises a serial spatial feature extraction module, a time feature extraction module and a full convolutional network; wherein the spatial feature extraction module and the time feature extraction module carry out normalization processing on input features through a space-time hybrid normalization layer, and the space-time hybrid normalization layer comprises three operations of time normalization, space normalization and hybrid normalization; and finally, training the short-term traffic flow prediction model by using the historical traffic flow data, and using the trained short-term traffic flow prediction model to predict the traffic flow. According to the method, the spatial-temporal correlation and dependency of the traffic flow data are more fully mined, and the prediction accuracy is improved.

    本发明为一种基于图卷积循环神经网络的短时交通流预测方法,首先获取路网信息,将路网信息转换为路网图,路网图由多个道路节点以及连接道路节点的边组成;然后,构建短时交通流预测模型,短时交通流预测模型通过学习映射函数,根据路网图和多个历史时间片的交通流特征来预测未来多个时间片的交通流;短时交通流预测模型包括串行的空间特征提取模块、时间特征提取模块以及全卷积网络;其中,空间特征提取模块和时间特征提取模块分别通过时空混合归一层对输入特征进行归一化处理,时空混合归一化层包括时间归一化、空间归一化以及混合归一化三种操作;最后,利用历史交通流数据对短时交通流预测模型进行训练,将训练后的短时交通流预测模型用于预测交通流。该方法更充分挖掘交通流数据的时空相关性及依赖性,提高了预测准确性。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Short-term traffic flow prediction method based on graph convolution recurrent neural network


    Weitere Titelangaben:

    基于图卷积循环神经网络的短时交通流预测方法


    Beteiligte:
    GU JUNHUA (Autor:in) / GUO RUIZHE (Autor:in) / HE WENYING (Autor:in) / WU JINGUANG (Autor:in) / ZHANG XIYANG (Autor:in) / HUANG TIANBO (Autor:in) / ZHENG HAIFEI (Autor:in)

    Erscheinungsdatum :

    2024-04-12


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06Q Datenverarbeitungssysteme oder -verfahren, besonders angepasst an verwaltungstechnische, geschäftliche, finanzielle oder betriebswirtschaftliche Zwecke, sowie an geschäftsbezogene Überwachungs- oder Voraussagezwecke , DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Optimized Graph Convolution Recurrent Neural Network for Traffic Prediction

    Guo, Kan / Hu, Yongli / Qian, Zhen et al. | IEEE | 2021


    Short-term traffic flow prediction with LSTM recurrent neural network

    Kang, Danqing / Lv, Yisheng / Chen, Yuan-yuan | IEEE | 2017


    Traffic flow prediction method of recurrent neural network based on dynamic diffusion diagram convolution

    LI YUN / TIAN YUDOU / JING PEIGUANG | Europäisches Patentamt | 2022

    Freier Zugriff