The invention discloses a vehicle track prediction method based on a multi-vehicle space-time interaction relationship. The method comprises the following steps: firstly, generating a speed relation graph, an acceleration relation graph and a driving intention relation graph of a target vehicle and surrounding vehicles according to a road driving data set, and then generating a new environment feature graph aggregating neighbor features through SuperGATConv fusion; secondly, inputting the new environment feature map into an environment intention encoder constructed by an expanded convolutional network to generate space encoding features; and inputting the vehicle data information of the predicted target vehicle as an input feature into the LSTM to obtain a time coding feature. And then splicing the space coding features and the time coding features, and inputting the spliced features into a space-time decoder to obtain future state features. And finally, generating a predicted future trajectory through a prediction module constructed by the LSTM and the MPL according to the future state features. Rich interaction information between vehicles is obtained, local space information is guaranteed not to be lost, and the prediction performance of the vehicle track is improved.

    本发明公开了一种基于多车时空交互关系的车辆轨迹预测方法。本发明首先根据公路行车数据集,生成目标车辆与周围车辆的速度关系图、加速度关系图以及驾驶意图关系图,再通过SuperGATConv融合生成聚合了邻居特征新的环境特征图;其次将新的环境特征图输入到由扩张卷积网络构建的环境意图编码器中生成空间编码特征;将被预测的目标车辆的车辆数据信息作为输入特征,输入到LSTM中得到时间编码特征。然后将空间编码特征和时间编码特征进行拼接,并输入到时空解码器中,得到未来状态特征。最后未来状态特征,通过由LSTM和MPL构建的预测模块,生成预测的未来轨迹。本发明获得车辆之间丰富的交互信息,保证局部空间信息不丢失,提高车辆轨迹的预测性能。


    Access

    Download


    Export, share and cite



    Title :

    Vehicle track prediction method based on multi-vehicle space-time interaction relationship


    Additional title:

    一种基于多车时空交互关系的车辆轨迹预测方法


    Contributors:
    ZHAI CHUNJIE (author) / XIAO JIAOLIE (author)

    Publication date :

    2024-11-01


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / 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



    Vehicle track prediction method based on intention perception space-time attention network

    GUO JINGHUA / WANG JINGYAO / HE ZHIFEI et al. | European Patent Office | 2023

    Free access

    Vehicle track prediction method based on heterogeneous node time-space domain perception

    GAO MINGYU / ZENG XUEPU / YANG JUNJIE et al. | European Patent Office | 2023

    Free access


    Vehicle track prediction method based on dynamic interaction graph convolution

    SHEN GUOJIANG / LI PENGFEI / KONG XIANGJIE et al. | European Patent Office | 2022

    Free access

    Vehicle - Track Interaction

    Mau, F. / Olsen, P. / O Brien, A. et al. | British Library Conference Proceedings | 2001