The invention discloses a traffic flow prediction method based on an adaptive lightweight time convolution network, and the method comprises the steps: carrying out the down-sampling of an input historical traffic flow in a context sensing embedding layer, then providing context information for the traffic flow of each time step after the down-sampling, and carrying out the embedded aggregation with time, and obtaining processed data; and in an output layer, taking the data after space embedding and processing and the time embedding as the input of stacked space-time layers, carrying out jump connection on the output of the adaptive lightweight time convolution module in each space-time layer to obtain an aggregation feature, and then obtaining the predicted traffic flow through convolution operation. According to the method, the adaptive lightweight time convolution module is adopted to deeply extract time features, and model parameters can be adaptively updated in a lightweight and fine-grained manner, so that the processing capability of traffic flow in multiple time modes is improved. And the method can be conveniently expanded to other space-time prediction tasks without priori knowledge.

    本发明公开了一种基于自适应轻量级时间卷积网络的交通流量预测方法,在上下文感知嵌入层中,对输入的历史交通流量进行下采样,然后为下采样后的每个时间步的交通流量提供上下文信息,并与时间嵌入聚合,得到处理后的数据;将空间嵌入、处理后的数据和时间嵌入作为堆叠的时空层的输入,在输出层,将每一个时空层中自适应轻量级时间卷积模块的输出进行跳跃连接得到聚合特征,然后通过卷积操作得到预测的交通流量。本发明采用自适应轻量级时间卷积模块以深入提取时间特征,能以轻量级、细粒度的方式自适应地更新模型参数,从而提高交通流量多种时间模式的处理能力。无需先验知识,可以便捷地扩展到其他的时空预测任务。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Traffic flow prediction method based on adaptive lightweight time convolution network


    Weitere Titelangaben:

    一种基于自适应轻量级时间卷积网络的交通流量预测方法


    Beteiligte:
    ZHANG SHUAI (Autor:in) / YIN XIANG (Autor:in) / ZHANG WENYU (Autor:in) / JING XIN (Autor:in) / JIN CHENGLEI (Autor:in) / MAO JIAYI (Autor:in)

    Erscheinungsdatum :

    02.01.2024


    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



    Periodic adaptive graph convolution circulation network traffic flow prediction method

    WANG BIN / LONG ZHENDAN / SHENG JINFANG et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    Road section traffic flow prediction method based on time convolution neural network

    JIN SHENG / CHANG WEI | Europäisches Patentamt | 2020

    Freier Zugriff


    Traffic flow prediction method based on graph attention convolution network

    ZHENG HONG / ZHANG SIKAI / LIU JIAMOU et al. | Europäisches Patentamt | 2020

    Freier Zugriff

    Traffic flow short-term prediction optimization application method based on time convolution network

    CUI TIANZE / DOU XUN / LUO HAIFENG et al. | Europäisches Patentamt | 2021

    Freier Zugriff