The invention relates to an automatic driving reinforcement learning motion planning method and system considering uncertainty, and the method comprises the following steps: training a pre-constructed multi-head value function network through employing a deep reinforcement learning method based on obtained historical driving data; environment state information at the current moment is obtained, a motion track based on data fitting and a motion track based on rules are obtained through the multi-head value function network obtained through training and an automatic driving system of the automatic driving vehicle, and the optimal motion track of the automatic driving vehicle is obtained after comparison. By combining bootstrap uncertainty estimation, the uncertainty possibly caused by the black box characteristic of the reinforcement learning algorithm in the calculation process is considered, dangerous decisions possibly caused by the uncertainty are avoided through a rule-based method, and the use safety and reliability of the method in an intelligent vehicle can be improved. Therefore, the method can be widely applied to the field of intelligent vehicles.

    本发明涉及一种考虑不确定性的自动驾驶强化学习运动规划方法和系统,包括以下步骤:基于获取的历史驾驶数据,使用深度强化学习方法对预先构建的多头值函数网络进行训练;获取当前时刻的环境状态信息,并分别利用训练得到的多头值函数网络和自动驾驶车辆的自动驾驶系统得到基于数据拟合的运动轨迹和基于规则的运动轨迹,比较后得到自动驾驶车辆的最优运动轨迹。通过结合自举式不确定性估计,考虑了强化学习算法在计算过程中可能由于其黑箱特性导致的不确定性,通过基于规则的方法避免了这种不确定性可能造成的危险决策,有助于提升该类方法在智能车辆中使用的安全性和可靠性。因此,本发明可以广泛应用于智能车辆领域。


    Access

    Download


    Export, share and cite



    Title :

    Automatic driving reinforcement learning motion planning method and system considering uncertainty


    Additional title:

    考虑不确定性的自动驾驶强化学习运动规划方法和系统


    Contributors:
    YANG DIANGE (author) / JIANG KUN (author) / ZHOU WEITAO (author) / CAO ZHONG (author) / DENG NANSHAN (author) / LIU XIAOYU (author)

    Publication date :

    2023-05-16


    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 driving decision uncertainty estimation method based on reinforcement learning

    ZOU QIANYING / LIAO YUAN | European Patent Office | 2023

    Free access

    Motion Planning Approach Considering Localization Uncertainty

    Artuñedo, Antonio / Villagra, Jorge / Godoy, Jorge et al. | BASE | 2020

    Free access

    Automatic driving decision planning method based on deep reinforcement learning and deep learning

    YANG LU / ZHANG HAO / TAN YANSONG et al. | European Patent Office | 2024

    Free access

    Automatic driving automobile decision planning method based on maximum entropy reinforcement learning

    TANG XIAOLIN / HUANG BING / LIU TENG et al. | European Patent Office | 2022

    Free access

    Automatic driving vehicle decision planning method considering interactive game

    XU CAN / ZHAO WANZHONG / LI LIN et al. | European Patent Office | 2021

    Free access