The invention relates to the technical field of traffic control, in particular to a traffic signal lamp timing regulation and control method and system based on deep reinforcement learning. The invention discloses a traffic signal lamp timing regulation and control method based on deep reinforcement learning, and the method comprises the steps: 1, initializing two Q-value networks of the same structure, taking one of the Q-value networks as an action value network # imgabs0 #, and taking the other Q-value network as a target Q-value network # imgabs1 #; 2, taking the target road intersection as a target environment, and periodically collecting the state of the target environment; and step 3, performing multiple rounds of iterative operation on # imgabs2 # and # imgabs3 # based on double-Q learning, and synchronously performing phase control on traffic lights in the target environment until the maximum iteration round number N is reached. According to the method, a double-Q learning method is adopted, experience playback is introduced, a target Q value network is used for evaluating the Q value, the process of selecting the optimal action and the process of evaluating the optimal action are separated, and the optimal strategy can be learned better and more stably.

    本发明涉及交通控制技术领域,具体涉及基于深度强化学习的交通信号灯配时调控方法及系统。本发明公开了基于深度强化学习的交通信号灯配时调控方法,包括:步骤一,初始化两个相同结构的Q值网络,并将其中一个Q值网络作为动作价值网络#imgabs0#、另一个Q值网络作为目标Q值网络#imgabs1#;步骤二,将目标道路交叉口作为目标环境,并周期性采集目标环境的状态;步骤三,基于双Q学习对#imgabs2#、#imgabs3#进行多轮迭代运算,并同步对目标环境中的交通信号灯进行相位控制,直至达到最大迭代轮数N。本发明采用双Q学习方法,通过引入经验回放、使用目标Q值网络来评估Q值,将选择最优动作的过程和评估最优动作的过程分开,可以更好、更稳定性的学习到最优策略。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Traffic signal lamp timing regulation and control method and system based on deep reinforcement learning


    Weitere Titelangaben:

    基于深度强化学习的交通信号灯配时调控方法及系统


    Beteiligte:
    GUO LIQUAN (Autor:in) / XING HONGYE (Autor:in) / WANG BO (Autor:in) / LI CHEN (Autor:in) / SONG WENJING (Autor:in)

    Erscheinungsdatum :

    2024-04-02


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen



    Traffic signal lamp timing method, device and equipment based on deep reinforcement learning

    ZHOU JIANHONG / HUANG YUMIN / WANG YUNXIANG et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    Traffic signal lamp control system and method based on reinforcement learning and dynamic timing

    LIU BINGYI / SONG BIN / WU PENGCHUN et al. | Europäisches Patentamt | 2021

    Freier Zugriff


    Crossroad traffic signal lamp control method based on reinforcement learning

    WU LIBING / WANG MIN / ZHANG RUI et al. | Europäisches Patentamt | 2020

    Freier Zugriff

    Traffic signal lamp control method and system based on reinforcement learning

    LUO JUAN / ZHENG YANLIU | Europäisches Patentamt | 2021

    Freier Zugriff