The invention discloses a deep reinforcement learning traffic signal control poisoning attack method based on Trojan attack, which comprises the following steps: training a traffic intersection signal lamp control model according to an existing deep reinforcement learning DQN algorithm and Trojan attack, and generating abnormal traffic state data by modifying vehicle information at a specified position; and setting abnormal actions adopted by the model under the abnormal traffic state data, and finally, setting a relatively high reward value for the abnormal actions adopted under the abnormal state to realize poisoning attack of a deep reinforcement learning DQN algorithm. According to the method, the vehicle waiting time of the intersection is prolonged in the model testing process, the performance of the model is greatly reduced, and the vehicle circulation efficiency of the traffic intersection is greatly reduced.

    本发明公开了一种基于木马攻击的深度强化学习交通信号控制中毒攻击方法,根据已有的深度强化学习DQN算法并加入木马攻击对交通路口信号灯控制模型进行训练,通过修改指定位置的车辆信息生成异常交通状态数据,再设定异常交通状态数据下模型采取的异常动作,最后对该异常状态下采取的异常动作设定较高的奖励值实现深度强化学习DQN算法的中毒攻击。本发明最后在模型测试过程中增加路口的车辆等待时间,大幅降低模型的性能,大大降低交通路口的车辆流通效率。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Deep reinforcement learning traffic signal control poisoning attack method based on Trojan horse attack


    Weitere Titelangaben:

    基于木马攻击的深度强化学习交通信号控制中毒攻击方法


    Beteiligte:
    XU DONGWEI (Autor:in) / WANG DA (Autor:in) / LI CHENGBIN (Autor:in)

    Erscheinungsdatum :

    2022-12-02


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    H04L TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION , Übertragung digitaler Information, z.B. Telegrafieverkehr / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Deep Reinforcement Learning-based Traffic Signal Control

    Ruan, Junyun / Tang, Jinzhuo / Gao, Ge et al. | IEEE | 2023


    Yugoslav Trojan Horse

    Online Contents | 2000


    The Trojan Horse reconstruction

    Chondros, Thomas G | Online Contents | 2015



    Traffic signal control method based on deep reinforcement learning

    LIU DUANYANG / SHEN SI / SHEN GUOJIANG et al. | Europäisches Patentamt | 2021

    Freier Zugriff