The invention discloses a multi-AUV underwater target attack method and system based on deep reinforcement learning, and belongs to the technical field of underwater vehicles. Comprising the following steps: constructing an AUV kinematic model based on a mathematical expression of underwater motion of the AUV; constructing an underwater confrontation model based on the AUV kinematic model; designing a state space, an action space and a reward function in reinforcement learning to obtain a reinforcement learning algorithm; constructing a single AUV decision model based on the underwater confrontation model and a reinforcement learning algorithm; obtaining a multi-AUV collaborative decision-making model based on interaction among an intelligent agent, a state action space and an optimal strategy in the single AUV decision-making model; and performing iterative training on the multi-AUV collaborative decision model according to the hyper-parameters of the multi-AUV collaborative decision model to obtain an optimal multi-AUV collaborative decision model. According to the invention, the training speed and cooperative capability of each AUV are improved, and the agents can be promoted to act in a coordinated manner, so that the strike planning of the whole cluster is more efficient and intelligent.

    本发明公开了一种基于深度强化学习的多AUV水下目标攻击方法及系统,属于水下航行器技术领域。包括:基于AUV水下运动的数学表达式构建AUV运动学模型;基于所述AUV运动学模型构建水下对抗模型;并设计强化学习中的状态空间、动作空间和奖励函数,得到强化学习算法;基于所述水下对抗模型和强化学习算法构建单AUV决策模型;基于所述单AUV决策模型中的智能体、状态动作空间以及最优策略之间的交互,得到多AUV协同决策模型;根据所述多AUV协同决策模型的超参数,对所述多AUV协同决策模型进行迭代训练,得到最优多AUV协同决策模型。本发明提高每艘AUV的训练速度和协同能力,能够促使智能体协调行动,从而使得整个集群的打击规划更加高效和智能化。


    Access

    Download


    Export, share and cite



    Title :

    Multi-AUV underwater target attack method and system based on deep reinforcement learning


    Additional title:

    一种基于深度强化学习的多AUV水下目标攻击方法及系统


    Contributors:
    XU YANG (author) / ZHANG KAI (author) / ZHAO WEN (author)

    Publication date :

    2024-08-23


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G05D SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES , Systeme zum Steuern oder Regeln nichtelektrischer veränderlicher Größen / B63G OFFENSIVE OR DEFENSIVE ARRANGEMENTS ON VESSELS , Angriffs- oder Verteidigungsanordnungen auf Schiffen / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen



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

    XU DONGWEI / WANG DA / LI CHENGBIN | European Patent Office | 2022

    Free access

    Hybrid electric vehicle ecological driving method based on multi-target deep reinforcement learning

    PENG JIANKUN / FAN YI / YU SICHEN et al. | European Patent Office | 2023

    Free access

    Deep Learning Based Approach for Underwater Acoustic Target Detection

    Akin Kucuk, Fatma K. / Bilgin, Mustafa / Kaymaz, Mehmetcan et al. | AIAA | 2025


    Automatic driving system backdoor attack method based on deep reinforcement learning and related device

    LIU JIAJIA / YU YINBO / YAN SAIHAO | European Patent Office | 2023

    Free access

    Multi-UAV Cooperative Target Assignment Method Based on Reinforcement Learning

    Yunlong Ding / Minchi Kuang / Heng Shi et al. | DOAJ | 2024

    Free access