The invention briefly describes a collision avoidance method for a dynamic obstacle of a driverless automobile. A fuzzy control rule base is established mainly through deep reinforcement learning of the driverless automobile on human collision avoidance operation examples or simulation platform test data, then fuzzy processing is carried out according to the unmanned driving speed collected by a sensor or camera equipment and the longitudinal or transverse distance between the driverless automobile and the obstacle, rule reasoning is carried out according to the fuzzy rule base, a fuzzy processing result is output, and according to defuzzification processing, a specific operation behavior result: speed invariance, deceleration, deceleration + lane change and other behavior operations is determined.

    本发明简述了一种无人驾驶汽车动态障碍物的避撞方法,主要通过无人驾驶汽车对人类避撞操作实例或仿真平台测试数据的深度强化学习,建立模糊控制规则库。然后根据传感器或摄像设备所采集到无人驾驶速度、与障碍物之间纵向或横向之间距离,进行模糊化处理,根据模糊规则库,进行规则推理,输出一个模糊处理结果,再根据反模糊化处理,确定一个具体操作行为结果:速度不变、减速以及减速+换道等行为操作。


    Access

    Download


    Export, share and cite



    Title :

    Collision avoidance method and system for dynamic obstacle of driverless automobile


    Additional title:

    一种无人驾驶汽车动态障碍物避撞的方法和系统


    Contributors:
    SHENG JING (author) / LIU GUOMAN (author)

    Publication date :

    2021-01-01


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion



    Automatic obstacle avoidance driverless vehicle driverless system

    PANG QUNYING | European Patent Office | 2021

    Free access

    OBSTACLE AVOIDANCE METHOD AND DEVICE USED FOR DRIVERLESS VEHICLE

    WANG YUE / SHEN DONGHUI / CHENG LIE et al. | European Patent Office | 2021

    Free access

    Driverless automobile obstacle recognition and detection system based on deep learning

    JIA RUI / CUI SHENGGUANG / WANG ZHIXUE et al. | European Patent Office | 2020

    Free access