The invention provides an automatic driving non-signalized intersection decision generation method based on hierarchical reinforcement learning, and relates to the technical field of vehicle driving decisions. The method comprises the following steps: acquiring a real driving environment where a vehicle and an opposite vehicle are located; the driving environment comprises an environment observation value and an action space; inputting the real driving environment into a path decision-making module to obtain a decision-making track; inputting the decision track into a speed decision module to obtain a decision speed; the path decision-making module and the speed decision-making module are trained based on a simulated driving environment, a vehicle track, an end condition and a reward function; and the vehicle executes a left-turn task according to the decision track and the decision speed. The speed decision and the path decision are trained respectively, the speed decision model can focus on learning how to select the optimal left-turn acceleration according to the current environment and other factors, the path decision model focuses on finding the proper radius capable of completing crossing, and the model performance and the accuracy of decision generation are improved.

    本公开提出一种基于分层强化学习的自动驾驶无信号交叉口决策生成方法,涉及车辆驾驶决策技术领域。方法包括:获取自车和对向来车所在的真实驾驶环境;所述驾驶环境包括环境观测值和动作空间;将真实驾驶环境输入路径决策模块,得到决策轨迹;将决策轨迹输入速度决策模块,得到决策速度;所述路径决策模块和速度决策模块基于模拟驾驶环境、自车轨迹、结束条件和奖励函数进行训练;自车根据决策轨迹和决策速度执行左转任务。将速度决策和路径决策分别训练,可使速度决策模型专注于学习如何根据当前环境和其他因素选择最佳左转加速度,而路径决策模型则专注于找到可完成穿越的合适半径,提高模型性能以及生成决策的准确性。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Automatic driving non-signalized intersection decision generation method based on hierarchical reinforcement learning


    Weitere Titelangaben:

    基于分层强化学习的自动驾驶无信号交叉口决策生成方法


    Beteiligte:
    CHEN XUEMEI (Autor:in) / TANG YUNHAO (Autor:in) / HAO JIACHEN (Autor:in) / LIU YUEZE (Autor:in) / TIAN YIHONG (Autor:in) / XIAO LONG (Autor:in) / DONG XIANYUAN (Autor:in) / ZHAO XIAOXUAN (Autor:in) / YAO CHENGDA (Autor:in) / GAO CONGZHENG (Autor:in)

    Erscheinungsdatum :

    18.10.2024


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    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 / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung



    Automatic driving vehicle non-signalized intersection coordination passing control system

    CUI JIANXUN / QU MINGCHENG | Europäisches Patentamt | 2021

    Freier Zugriff

    Non-signalized intersection vehicle cooperation method based on asynchronous deep reinforcement learning

    WANG SHUAI / DING YUHAO | Europäisches Patentamt | 2025

    Freier Zugriff

    Cooperative passing method and system for automatic driving motorcade at non-signalized intersection

    CHEN SHOUHAI / ZENG WEILIANG | Europäisches Patentamt | 2024

    Freier Zugriff

    Automatic driving decision-making control method and system based on hierarchical reinforcement learning

    CAI YINGFENG / YANG SHAOQING / TENG CHENGLONG et al. | Europäisches Patentamt | 2020

    Freier Zugriff

    Two-Level Signalized Intersection

    Shin, Eon Kyo | Online Contents | 2008