The invention discloses a traffic flow prediction method based on federated learning and a graph neural network, and the method comprises the steps: carrying out the initialization of a global model for traffic flow prediction, and transmitting the parameters of the global model to all clients; each client replaces the local model parameter with the obtained global model parameter, and further obtains local traffic flow data to train a local model; each client adds noise to the local model parameters through a differential privacy technology, and then uploads the local model parameters; the local model parameters uploaded by all the clients are aggregated, a new global model is obtained through average pooling operation, then the updated global model parameters are sent to all the clients, and the global model is continuously updated until convergence or the maximum number of iterations is reached; and inputting the traffic flows of the first M time points into the trained global model, and outputting traffic flow predicted values of the future N time points. According to the invention, privacy protection of the traffic flow data can be realized and the traffic flow prediction precision can be improved.

    本发明公开了一种基于联邦学习和图神经网络的交通流量预测方法,包括:对交通流量预测的全局模型进行初始化,并将全局模型的参数发送给各个客户端;各个客户端将局部模型参数替换为获取的全局模型参数,进而获取本地的交通流量数据来训练局部模型;各个客户端通过差分隐私技术对局部模型参数加入噪声,然后进行上传;聚合各个客户端上传的局部模型参数,使用平均池化操作得到新的全局模型,进而将更新后的全局模型参数发送给各个客户端,不断更新全局模型直至收敛或达到最大迭代次数;将前M个时间点的交通流量输入训练好的全局模型中输出未来N个时间点的交通流量预测值。本发明能够实现交通流数据的隐私保护并提高交通流预测的精度。


    Access

    Download


    Export, share and cite



    Title :

    Traffic flow prediction method based on federated learning and graph neural network


    Additional title:

    一种基于联邦学习和图神经网络的交通流量预测方法


    Contributors:
    MI BO (author) / ZENG RAN (author) / HUANG DARONG (author) / ZENG SIYUAN (author) / WANG FUYUAN (author) / SUN LINA (author) / LUO QIUJU (author)

    Publication date :

    2024-10-25


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / 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



    Digital twin traffic flow prediction system based on graph federated learning

    XING LING / GAO JIANPING / LI BING et al. | European Patent Office | 2024

    Free access

    Graph neural network traffic flow prediction method based on deep learning

    CHENG XIAOHUI / HE YUHAO / LU QIU | European Patent Office | 2023

    Free access

    Traffic flow prediction method based on graph neural network

    PENG LAIHU / WU BAOWEN / QI YUBAO et al. | European Patent Office | 2024

    Free access


    Node adaptive learning graph attention neural network traffic flow prediction method

    LOU JUNGANG / HUANG XUXIANG / SHEN QING et al. | European Patent Office | 2024

    Free access