The invention discloses a light-weight traffic flow prediction method based on deep learning, and the method comprises the steps: building a plurality of space-time attention mechanisms to excavate the space-time dependence of traffic flow, carrying out the processing of output data through a convolution module, carrying out the light-weight processing of a model through employing a pruning and distillation technology, removing the redundant parameters in a prediction model, and carrying out the prediction of the light-weight traffic flow. And a lightweight prediction model with relatively high accuracy is obtained. The method solves the problem that a microcontroller equipped for traffic infrastructures in a real road network environment is small in memory and is insufficient to deploy the most advanced traffic flow prediction model, realizes balance between traffic flow prediction precision and model calculation cost, is high in prediction precision and small in calculation memory, and is suitable for large-scale popularization and application. And a theoretical basis is laid and technical support is provided for high-precision traffic flow prediction in the microcontroller.

    本发明公开了一种基于深度学习的轻量化交通流预测方法,通过构建多个时空注意力机制挖掘交通流的时空依赖性,并利用卷积模块对输出数据进行处理,最后使用剪枝和蒸馏技术对模型进行轻量化处理,去除预测模型中的冗余参数,得到具有较高准确率的轻量预测模型。本发明解决了真实路网环境下交通基础设施配备的微控制器内存较小,不足以部署最先进的交通流预测模型的问题,实现了交通流预测精度和模型计算成本之间的平衡,不仅预测精度高且计算占用内存小,为实现在微控制器中进行精度较高的交通流预测奠定理论基础并提供了技术支持。


    Access

    Download


    Export, share and cite



    Title :

    Lightweight traffic flow prediction method based on deep learning


    Additional title:

    一种基于深度学习的轻量化交通流预测方法


    Contributors:
    ZHENG RUI (author) / DONG CHUNJIAO (author) / GENG QINGQIAO (author) / LI PENGHUI (author) / WANG JUNYUE (author)

    Publication date :

    2024-07-26


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    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 / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Traffic flow prediction method based on deep learning

    XIE GANG / WANG HAIYING / XIE RUIQI | European Patent Office | 2024

    Free access

    Traffic flow prediction method based on hybrid deep learning

    XIA DAWEN / CHEN YAN / LI HUAQING et al. | European Patent Office | 2023

    Free access

    Urban traffic flow prediction method based on deep learning

    WANG RONGXIU / DING YANQIU / ZHENG YAQIAN | European Patent Office | 2025

    Free access

    Traffic Flow Prediction Model Based on Deep Learning

    Wang, Bowen / Wang, Jingsheng / Zhang, Zeyou et al. | British Library Conference Proceedings | 2022


    Short-term traffic flow prediction method based on deep learning

    XIAO HONGBO / XIAO JIANHUA / DING LIMING et al. | European Patent Office | 2025

    Free access