The invention discloses a reinforcement learning automatic driving decision switching method and system based on sustainable learning, and relates to the technical field of reinforcement learning automatic driving decision switching based on sustainable learning, and the method comprises the steps: creating a multi-scene decision environment, building a decision switching model through a graph neural network, and carrying out decision switching. And adjusting each parameter in the decision switching model, acquiring information of the vehicle and the current scene in the sensing information of the automatic driving vehicle in real time, performing sustainable learning setting on the decision switching model, and driving the automatic driving vehicle into an unknown scene. According to the method, the characteristic of uncertainty of the model to the scene can be expressed through reinforcement learning of the action value function, and the reciprocal of the action value function is used as a label basis for decision switching in the face of a new scene; according to the value, a label relation between an automatic corresponding scene and a decision can be realized during driving in an unknown scene, and the label is used as a training fitting label of a decision switching model for a new scene.

    本发明公开了一种基于可持续学习的强化学习自动驾驶决策切换方法及系统,涉及基于可持续学习的强化学习自动驾驶决策切换技术领域,包括创建多场景的决策环境,通过图神经网络搭建决策切换模型,调整决策切换模型中的各个参数,实时获取自动驾驶汽车感知信息中的车辆以及当前场景的信息,对决策切换模型进行可持续学习的设置,自动驾驶汽车驶入未知场景。本发明所述方法通过强化学习动作值函数能够表示该模型对于场景的不确定性的特性,将其动作值函数的倒数作为面对新场景时进行决策切换的标签依据,根据该值能够实现在未知场景中行驶时实现自动对应场景与决策之间的标签联系,并作为决策切换模型针对新场景的训练拟合标签。


    Access

    Download


    Export, share and cite



    Title :

    Reinforcement learning automatic driving decision switching method based on sustainable learning


    Additional title:

    一种基于可持续学习的强化学习自动驾驶决策切换方法


    Contributors:
    PEI XIAOFEI (author) / LU SONGXIN (author) / YANG BO (author)

    Publication date :

    2024-11-01


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



    Automatic driving decision-making method based on deep reinforcement learning

    LIU CHENGQI / LIU SHAOWEIHUA / ZHANG YUJIE et al. | European Patent Office | 2024

    Free access

    Automatic driving decision uncertainty estimation method based on reinforcement learning

    ZOU QIANYING / LIAO YUAN | European Patent Office | 2023

    Free access

    Automatic driving vehicle microscopic decision-making method based on reinforcement learning

    ZHENG KAN / LIU JIE / ZHAO LONG | European Patent Office | 2020

    Free access

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

    YANG MINGZHU / LIU XIANGWEI / LI ZHUOLUO | European Patent Office | 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