The invention discloses an intelligent automobile collision probability prediction method based on multi-dimensional uncertainty perception, and relates to the technical field of automobile collision safety prediction, and the method comprises the following steps: obtaining information of a main automobile and surrounding automobiles, processing the information through an extended Kalman filter, carrying out the optimal estimation of the state of the automobile, and obtaining the collision probability of the automobile. According to the method, the influence of lane line curvature on vehicle motion is considered through a weight distribution method, CV/CT multiple models are selected to track linear motion and curvilinear motion of traffic participating vehicles respectively, kinematics models are switched according to the states of the traffic vehicles, and trajectory prediction of a main vehicle and surrounding vehicles in future 2s based on an uncertainty perception physical kinematics model is achieved; time discretization is carried out on the tracks of the two vehicles, and vehicle uncertainty track prediction is output in combination with Gaussian probability density; the method comprises the following steps: performing heuristic deterministic collision detection by taking a transverse and longitudinal safe distance as a detection condition, detecting whether two vehicles are likely to collide at each moment in 2s in the future, and performing collision probability calculation by applying an uncertainty prediction trajectory of Monte Carlo simulation combined with a Gaussian probability function. The automobile collision probability prediction method can adapt to various collision scenes, guarantees the real-time performance and accuracy of the probability prediction result, and has very high application value.

    本发明公开了一种多维度不确定性感知的智能汽车碰撞概率预测方法,涉及车碰撞安全预测技术领域,该方法包括以下步骤:获取主车和周边车辆的信息,使用扩展卡尔曼滤波器处理信息,对车辆状态进行最优估计,通过权重分配方法考虑了车道线曲率对车辆运动的影响,选取CV/CT多模型分别追踪交通参与车的直线运动和曲线运动,根据交通车的状态切换运动学模型,实现基于不确定性感知物理运动学模型的主车与周边车辆未来2s内轨迹预测;将两车轨迹进行时间离散化,联合高斯概率密度输出车辆不确定性轨迹预测;以横纵向安全距离为检测条件,进行启发式确定性碰撞检测,面向未来2s内每一个时刻检测两车是否有可能发生碰撞,运用蒙特卡洛模拟联合高斯概率函数的不确定性预测轨迹进行碰撞概率计算。本发明汽车碰撞概率预测方法可适应多种碰撞场景,同时保障了概率预测结果的实时性与准确性,具有很高的应用价值。


    Access

    Download


    Export, share and cite



    Title :

    Intelligent automobile collision probability prediction method based on multi-dimensional uncertainty perception


    Additional title:

    一种多维度不确定性感知的智能汽车碰撞概率预测方法


    Contributors:
    GAO FEI (author) / LIAN JIAJUN (author) / ZHAO RUI (author) / GAO ZHENHAI (author)

    Publication date :

    2023-09-15


    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



    Collision probability assessment based on uncertainty prediction of ship trajectories

    Rong, H. / Teixeira, A. P. / Soares, C. Guedes | TIBKAT | 2020


    Intelligent vehicle lane changing collision probability evaluation method based on track prediction

    DUAN YUCHEN / DENG YUCONG / QI WEIWEI et al. | European Patent Office | 2021

    Free access


    Evaluating Probability of Collision (Pc) Uncertainty

    Hejduk, M. D. / Johnson, L. C. | NTRS | 2016


    Automobile intelligent anti-collision system

    WANG CAIXIA | European Patent Office | 2020

    Free access