The invention discloses an automatic driving behavior decision-making method and device based on a game theory and reinforcement learning, and relates to the technical field of automatic driving. The method comprises the following steps: based on state information of an own vehicle and state information of a plurality of other agents at a historical moment, calculating an average probability that the plurality of other agents select robbing at the historical moment, and calibrating parameters of a game model; based on the calibrated game model, calculating the probability that other agents select robbing at historical moments; training a reinforcement learning model based on the state information of the own vehicle at the historical moment, the state information of the multiple other intelligent agents, the own vehicle longitudinal action set, the probability that the other intelligent agents select robbing at the historical moment, a preset state transition model and a preset reward function; and inputting the state information of the own vehicle at the current moment, the state information of other intelligent agents and the probability of selecting robbing by other intelligent agents into the trained reinforcement learning model to obtain the longitudinal action of the own vehicle at the next moment. The implementation mode is suitable for different scenes.

    本发明公开了一种基于博弈论和强化学习的自动驾驶行为决策方法和装置,涉及自动驾驶技术领域。该方法包括:基于历史时刻自车的状态信息和多个其他智能体的状态信息,计算历史时刻多个其他智能体选择抢行的平均概率,标定博弈模型的参数;基于标定的博弈模型,计算历史时刻各个其他智能体选择抢行的概率;基于历史时刻自车的状态信息、多个其他智能体的状态信息、自车纵向动作集合和历史时刻各个其他智能体选择抢行的概率、预设的状态转移模型和预设的奖励函数训练强化学习模型;将当前时刻自车的状态信息、其他智能体的状态信息和其他智能体选择抢行的概率,输入训练好的强化学习模型,得到下一时刻自车的纵向动作。该实施方式适用于不同场景中。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Automatic driving behavior decision-making method and device based on game theory and reinforcement learning


    Weitere Titelangaben:

    基于博弈论和强化学习的自动驾驶行为决策方法和装置


    Beteiligte:
    LYU YANG (Autor:in) / LYU QIANG (Autor:in) / MIAO QIANKUN (Autor:in)

    Erscheinungsdatum :

    2023-12-15


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



    Automatic driving behavior decision-making method based on deep reinforcement learning

    YANG MINGZHU / LIU XIANGWEI / LI ZHUOLUO | Europäisches Patentamt | 2020

    Freier Zugriff

    Automatic driving behavior integrated decision-making method based on deep reinforcement learning

    PENG JIANKUN / ZHANG SIYU / WANG YUCHENG et al. | Europäisches Patentamt | 2022

    Freier Zugriff

    Automatic driving decision-making method based on deep reinforcement learning

    LIU CHENGQI / LIU SHAOWEIHUA / ZHANG YUJIE et al. | Europäisches Patentamt | 2024

    Freier Zugriff

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

    LI DONGCHEN / ZHANG YIHAO / WEI CHONGSHAN et al. | Europäisches Patentamt | 2022

    Freier Zugriff

    Automatic driving decision-making method based on rule-assisted reinforcement learning

    ZHENG KAI / SU HAN / ZENG XIMU | Europäisches Patentamt | 2023

    Freier Zugriff