The invention discloses an intersection right-turn vehicle and pedestrian interaction trajectory prediction and risk assessment method, and relates to the technical field of road traffic safety, and the method comprises the steps: obtaining the motion data of pedestrians and vehicles in an intersection, screening and extracting the motion data of right-turn vehicles and pedestrians in interaction, building a deep learning-interaction trajectory prediction model, and carrying out the risk assessment of the right-turn vehicles and pedestrians in the intersection. Future trajectories of the man and the vehicle are obtained through prediction of the prediction model, and man-vehicle conflict risk quantification is carried out on trajectory coordinate positions at each moment on the future trajectories; a long-short-term memory neural network model of an encoder-decoder framework enables the model to describe a change rule of time sequence type data to be more fit, and potential characteristics of behaviors adopted for interaction by a person and a vehicle can be mined, so that the accuracy of predicting a person-vehicle interaction condition in a period of time in the future is obviously improved; and the post-invasion time PET is obtained according to the predicted future trajectory of the person and the vehicle, so that a relatively reasonable conflict risk quantification level is obtained.

    本发明公开了一种交叉口右转车辆与行人交互轨迹预测与风险评估方法,涉及道路交通安全技术领域,包括获取交叉口内的行人与车辆的运动数据,筛选、提取出存在交互的右转车辆和行人运动数据,建立深度学习‑交互轨迹预测模型,通过预测模型预测得到人车双方的未来轨迹,未来轨迹上每一时刻的轨迹坐标位置进行人车冲突风险量化;编码器‑解码器框架的长短期记忆神经网络模型,使得模型描述时间序列型数据的变化规律更加拟合,能够挖掘人车双方对于交互所采用行为的潜在特点,从而使得预测未来一段时间内的人车交互情况的精确性得到明显提高;且根据预测得到人车双方的未来轨迹得到后侵犯时间PET,从而得出相对合理的冲突风险量化等级。


    Access

    Download


    Export, share and cite



    Title :

    Intersection right-turn vehicle and pedestrian interaction trajectory prediction and risk assessment method


    Additional title:

    一种交叉口右转车辆与行人交互轨迹预测与风险评估方法


    Contributors:
    JIANG XU (author) / HUANG CONGJUN (author) / ZHANG DEHAO (author) / WANG KAN (author)

    Publication date :

    2024-01-09


    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 / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06V



    No-signal right-turn intersection pedestrian trajectory prediction method based on game theory

    LI WENLI / TANG YUANHANG / ZHANG YINAN et al. | European Patent Office | 2023

    Free access


    INTERSECTION RIGHT-TURN PEDESTRIAN ALERT SYSTEM

    MCDANIEL GEVIN JOSEPH | European Patent Office | 2023

    Free access