The invention discloses a vehicle track prediction method and system considering multiple target points, and the system comprises a vehicle feature and lane center line feature extraction network, a feature fusion network, a target point prediction network, and a feature decoding network. Comprising the following steps of S1, respectively extracting historical track features and lane center line features of all vehicles in a current traffic scene; s2, spreading and updating lane center line features as environmental constraints, fusing the lane center line features with vehicle historical trajectory features by adopting a spatial cross attention mechanism, and then realizing vehicle-vehicle interaction with constraints by adopting a spatial self-attention mechanism; s3, respectively predicting target points of each vehicle, including a midpoint and a plurality of terminal points, and endowing lane environment features near the midpoint and the terminal point with the highest confidence coefficient into the vehicle features; s4, decoding and outputting a plurality of predicted tracks and confidence coefficients, and smoothing; according to the method, the accuracy of a trajectory prediction result is improved, and multi-possibility prediction of future prediction is reserved.

    本发明公开了一种考虑多目标点的车辆轨迹预测方法及系统,包括车辆特征与车道中心线特征提取网络,特征融合网络,目标点预测网络和特征解码网络,包括以下步骤:S1.分别提取当前交通场景下所有车辆的历史轨迹特征和车道中心线特征;S2.对车道中心线特征进行传播和更新作为环境约束,采用空间交叉注意力机制与车辆历史轨迹特征进行融合,再对其采用空间自注意力机制实现带约束的车车交互;S3.分别预测各车辆的目标点,包括中点和多个终点,并将中点和置信度最高的终点附近的车道环境特征赋予车辆特征中;S4.解码输出预测的多条轨迹及置信度,并进行平滑处理;本发明增加了轨迹预测结果的准确性,并保留了未来预测的多可能性预测。


    Access

    Download


    Export, share and cite



    Title :

    Vehicle track prediction method and system considering multiple target points


    Additional title:

    一种考虑多目标点的车辆轨迹预测方法及系统


    Contributors:
    REN HONGBIN (author) / ZHAO SHIYUAN (author) / ZHOU GAOLI (author) / ZHAO YUZHUANG (author)

    Publication date :

    2023-12-01


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06V / 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 deep learning prediction method considering physical constraint

    XU FANG / HAN SHUAI / GUO ZHONGYI et al. | European Patent Office | 2023

    Free access

    Hybrid electric vehicle prediction energy management method considering SOC track

    ZENG XIAOHUA / WANG YUNYE / SONG DAFENG et al. | European Patent Office | 2023

    Free access

    Target vehicle track prediction method for multi-lane scene

    ZUO ZHIQIANG / GUO SONGLIN / WANG YIJING et al. | European Patent Office | 2023

    Free access

    Vehicle track prediction model construction method and vehicle track prediction method and system

    CHEN ZHE / WANG YUEHAN / LIU HUAN et al. | European Patent Office | 2020

    Free access

    System and method for displaying vehicle surrounding target detection and track prediction

    ZHI RONG / WANG BAOFENG | European Patent Office | 2022

    Free access