The invention discloses a trajectory prediction method for adaptive pseudo tag improvement of a behavior prediction domain. Inputting the source domain into a predictor for training to obtain a trained predictor; inputting the data of the target domain into a first neural network obtained by optimization of the trained predictor for re-training to obtain a trained first neural network; and finally, extracting and establishing a prediction network from the trained first neural network, and performing identification and prediction processing on real-time to-be-detected data. According to the method, the quality of the pseudo tag can be improved to improve the generalization performance of a model network, the diversity of target domain data can be ensured, the model is prevented from falling into local optimum, the prediction performance of the pseudo tag is improved, and the trajectory prediction capability is further improved.

    本发明公开了一种用于行为预测域自适应的伪标签改进的轨迹预测方法。将源域输入到预测器中进行训练,获得训练后的预测器;然后将目标域的数据输入到由训练后的预测器优化获得的第一神经网络中进行再次训练,获得训练后的第一神经网络;最后从训练后的第一神经网络提取组建预测网络,对实时待测的数据实进行识别预测处理。本发明能够提升伪标签的质量以提升模型网络的泛化性能,能够确保目标域数据的多样性以及防止模型陷入局部最优,提升了伪标签的预测性能,进而提升轨迹预测能力。


    Access

    Download


    Export, share and cite



    Title :

    Pseudo-label improved trajectory prediction method for behavior prediction domain self-adaption


    Additional title:

    一种用于行为预测域自适应的伪标签改进的轨迹预测方法


    Contributors:
    ZHANG ZHANWEI (author) / LAI SHENQI (author) / PENG LIANG (author) / QIAN WEI (author) / YANG ZHENG (author) / HE XIAOFEI (author)

    Publication date :

    2023-10-27


    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



    Uncertainty-Aware Pseudo Labels for Domain Adaptation in Pedestrian Trajectory Prediction

    Poibrenski, Atanas / Nozarian, Farzad / Rezaeianaran, Farzaneh et al. | IEEE | 2023


    Transverse control method based on trajectory boundary self-adaption

    KUANG RUI / LUO MAN / KUANG ZONGXU et al. | European Patent Office | 2024

    Free access

    TRAJECTORY REPRESENTATION IN BEHAVIOR PREDICTION SYSTEMS

    REFAAT KHALED / ROSS STEPHANE | European Patent Office | 2022

    Free access

    Trajectory prediction method, trajectory prediction device and storage medium

    DUAN ZHIXIANG / YANG KUIYUAN | European Patent Office | 2022

    Free access

    TRAJECTORY PREDICTION

    HUANG CHAO / YU JIANING / XIE MIN et al. | European Patent Office | 2022

    Free access