The invention discloses a trajectory prediction method under scene fusion based on a space-time structure causal model, and the method comprises the steps: firstly eliminating a mixed path in a feature extraction process through a front door criterion, and merging a scene context into a space-time state; embedding the dynamic space-time causal relationship into the constructed space-time causal graph, and extracting a space-time dependency relationship through causal convolution; and finally, deducing the spatial-temporal characteristics of the trajectory into a future traffic scene by using an anti-fact representation reasoning method to predict the trajectory. The invention provides a causal reasoning-based trajectory prediction structure causal model, explains a real reason for performance improvement brought by scene context in trajectory prediction, analyzes hybrid path interference limiting model performance, realizes dynamic interaction between trajectory prediction and scenes, and has good robustness.

    本发明公开一种基于时空结构因果模型的场景融合下轨迹预测方法,首先通过前门准则消除特征提取过程中的混杂路径,将场景上下文合并到时空状态;接着将动态时空因果关系嵌入所构建的时空因果图,通过因果卷积来提取时空依赖关系;最后利用反事实表征推理的方法,将轨迹的时空特征推演到未来交通场景中来进行轨迹预测。本发明提出了基于因果推理的轨迹预测结构因果模型,解释了场景上下文在轨迹预测中带来性能提升的真正原因,分析了限制模型性能的混杂路径干扰,实现轨迹预测与场景间的动态交互并具有良好的鲁棒性。


    Access

    Download


    Export, share and cite



    Title :

    Track prediction method under scene fusion based on space-time structure causal model


    Additional title:

    一种基于时空结构因果模型的场景融合下轨迹预测方法


    Contributors:
    WANG XIAODING (author) / QUE YOUXIONG (author) / ZHAO MIN (author) / WU QIBIN (author) / LIU JIANMIN (author) / LIN HUI (author)

    Publication date :

    2024-04-26


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Traffic prediction method based on dynamic causal analysis and space-time adaptive fusion graph learning

    NING NIANWEN / ZHANG WENBO / CHEN BAICHUAN et al. | European Patent Office | 2024

    Free access

    Causal Scene Understanding

    Cooper, P. R. / Birnbaum, L. A. / Brand, M. E. | British Library Online Contents | 1995


    Traffic flow prediction method based on space-time causal attention network

    ZHAO YUAN / LEI HANG / MOON SE-HEE et al. | European Patent Office | 2024

    Free access

    Airport scene track prediction method based on multi-head attention mechanism

    SHI XIAOZHU / JIANG ZHIQIAN / WANG ERSHEN et al. | European Patent Office | 2022

    Free access

    Intersection traffic scene radar video real-time high-precision track fusion and connection method

    YU XIAOLU / CHEN YINGYING / WEI XIAOYONG et al. | European Patent Office | 2023

    Free access