The invention relates to a cloud-fog collaborative personalized forward collision risk early warning method based on federated learning, and the method specifically comprises the following steps: S1, a cloud end carries out the training of a long-short term memory neural network according to time sequence characteristic numerical data, generates an initial vehicle driving forward collision risk early warning model, and transmits the model to a fog end; S2, the fog end obtains local data of accumulated driving of a driver, updates the initial forward collision risk early warning model before driving of the vehicle through a gradient descent algorithm, and uploads the updated model parameters to the cloud end; S3, the cloud side aggregates the returned model parameters, updates the initial vehicle driving forward collision risk early warning model and downloads the model to the fog side; and S4, the fog end carries out parameter adjustment on driver characteristics through a transfer learning algorithm, and a personalized forward collision risk early warning model is formed. Compared with the prior art, the method has the advantages that personalized high-precision early warning is carried out for forward collision, the utilization rate of an advanced driving assistance system (ADAS) is improved, and the occurrence rate of rear-end collision accidents is reduced.

    本发明涉及一种基于联邦学习的云雾协同个性化前向碰撞风险预警方法,具体包括以下步骤:S1、云端根据时序特征数值数据,通过长短时记忆神经网络,训练生成初始车辆行驶前向碰撞风险预警模型并下传至雾端;S2、雾端获取驾驶员累计行驶的本地数据,通过梯度下降算法对初始车辆行驶前向碰撞风险预警模型进行更新,模型参数更新后上传至云端;S3、云端对回传的模型参数进行聚合,更新初始车辆行驶前向碰撞风险预警模型并下传至雾端;S4、雾端通过迁移学习算法对驾驶员特性进行调参,形成个性化前向碰撞风险预警模型。与现有技术相比,本发明针对前向碰撞进行个性化高精度预警,具有提高高级驾驶辅助系统(ADAS)的利用率,降低追尾事故发生率等优点。


    Access

    Download


    Export, share and cite



    Title :

    Cloud-fog collaborative personalized forward collision risk early warning method based on federated learning


    Additional title:

    基于联邦学习的云雾协同个性化前向碰撞风险预警方法


    Contributors:
    YU RONGJIE (author) / ZHANG RUICI (author) / AI HAO'AN (author)

    Publication date :

    2022-01-11


    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 / G06K Erkennen von Daten , RECOGNITION OF DATA / 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 / H04L TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION , Übertragung digitaler Information, z.B. Telegrafieverkehr



    Collision early warning method based on adaptive federated learning

    TANG LUN / WEN MINGYAN / SHAN ZHENZHEN et al. | European Patent Office | 2023

    Free access

    Forward collision early warning method and device

    LIU XIAO | European Patent Office | 2023

    Free access

    Forward collision early warning method and system

    LI ANPEI / WU ZHONGWEN / ZHANG JIANWEI et al. | European Patent Office | 2020

    Free access

    Early warning information credibility prediction method based on overspeed early warning and forward collision early warning

    NIU SHIFENG / GUO ZIHAO / HU YANTING et al. | European Patent Office | 2024

    Free access

    Microwave radar-based forward collision-preventing early warning method

    HE ZHAODAN / WANG XINGANG | European Patent Office | 2015

    Free access