The invention provides a traffic flow prediction method and system based on hierarchical attention, and the method comprises the steps: capturing traffic flow space-time modes of different periods through three independent channels, and employing a coupled graph convolution network CGCN and residual image attention RGAT in a circulation layer of a scale channel, the method aims to obtain aggregated edge weighted flow states and dynamic learning node weights, after prediction of three channels is obtained, residual channel attention RCAT is used for integration, and a final prediction result is calculated. According to the traffic flow prediction method provided by the invention, through the three independent channels, the CGCN, the RGAT and the RCAT used for fusing time features and generating final prediction, modeling is effectively carried out on space, time and semantic correlation among roads in a traffic network, a real dynamic flow scene is better fitted, and the traffic flow prediction efficiency is improved. Therefore, the traffic flow state in different time and space can be dynamically learned from multiple angles, and the accuracy of traffic flow prediction is further improved.

    本发明提出一种基于分层注意力的交通流量预测方法及系统,该方法为:利用三个独立的通道来捕获不同周期的交通流量时空模式,在尺度通道的循环层采用耦合图卷积网络CGCN和残差图注意力RGAT,旨在获得聚合的边加权流量状态和动态学习节点权重,在获得三个通道的预测后,应用残差通道注意力RCAT进行整合,计算出最终的预测结果。本发明提出的一种交通流量预测方法,通过三个独立通道、CGCN、RGAT以及用于融合时间特征并生成最终预测的RCAT,有效地对交通网络中道路之间的空间、时间和语义相关性进行建模,更贴合真正的动态流量场景,实现多个角度动态学习不同时空下的交通流量状态,由此进一步提高了预测交通流量的准确度。


    Access

    Download


    Export, share and cite



    Title :

    Traffic flow prediction method and system based on hierarchical attention


    Additional title:

    一种基于分层注意力的交通流量预测方法及系统


    Contributors:
    LIAO HUICHUAN (author) / ZHOU YUNFEI (author) / LAN YUANCHUN (author) / LING JIAHAO (author) / HUANG XIAOHUI (author)

    Publication date :

    2023-12-08


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    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



    Traffic speed prediction method based on hierarchical clustering and hierarchical attention mechanism

    LIU DUANYANG / XU XINBO / TANG LONGFENG et al. | European Patent Office | 2020

    Free access

    Traffic flow prediction method based on attention mechanism

    SUN LIJUN / LIU MINGZHI / LIU GUANFENG et al. | European Patent Office | 2022

    Free access

    Traffic flow prediction method based on LSTM-Attention

    QIN XIAOLIN / LIU JIACHEN / SONG LIXIANG et al. | European Patent Office | 2021

    Free access

    Traffic flow speed prediction method based on feature attention

    SUN YONG / ZHANG ANQIN / CHEN JINGJING | European Patent Office | 2023

    Free access

    Traffic flow prediction method based on space-time attention network

    MA CHUANG / YAN LI / LIU SHUAIWU et al. | European Patent Office | 2023

    Free access