Chinese Journal of Ship Research, Volume. 16, Issue 1, 105(2021)

Tracking control of intelligent ship based on deep reinforcement learning

Kang ZHU1, Zhen HUANG1, and Xuming WANG2
Author Affiliations
  • 1School of Automation, Wuhan University of Technology, Wuhan 430070, China
  • 2Intelligent Transport System Research Center, Wuhan University of Technology, Wuhan 430063, China
  • show less
    Figures & Tables(14)
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    • Table 1. Parameters of a KVLCC2 tanker

      View table
      View in Article

      Table 1. Parameters of a KVLCC2 tanker

      参数数值参数数值
      船长Lpp/m7方形系数${C_{\rm{b}}}$0.809 8
      船宽Bwl/m1.168 8浮心坐标/m0.244 0
      型深D/m0.656 3螺旋桨直径Dp/m0.216 0
      排水体积/m33.272 4舵面积/m20.053 9
    • Table 2. Critic network parameters

      View table
      View in Article

      Table 2. Critic network parameters

      参数赋值
      输入层状态向量$ {\boldsymbol{S}}\left( t \right)$
      第1个隐层300
      第1层激活函数Relu
      第2个隐层200
      第2层激活函数Relu
      输出层动作$ \delta \left( t \right)$
      输出层激活函数Tanh
      参数初始化Xavier初始化
      学习率0.000 1
      优化器Adam
    • Table 3. Actor network parameters

      View table
      View in Article

      Table 3. Actor network parameters

      参数赋值
      输入层状态向量$ {\boldsymbol{S}}\left( t \right)$,动作$ \delta \left( t \right)$
      第1个隐层300
      第1层激活函数Relu
      第2个隐层200
      第2层激活函数Relu
      输出层$Q\left( { {\boldsymbol{S}}\left( i \right),\delta \left( i \right)} \right)$
      输出层激活函数Linear
      参数初始化Xavier初始化
      学习率0.001
      优化器Adam
    • Table 4. Control performance

      View table
      View in Article

      Table 4. Control performance

      控制器RMSE
      BP-PID控制器13.585 0
      DDPG控制器6.911 96
    Tools

    Get Citation

    Copy Citation Text

    Kang ZHU, Zhen HUANG, Xuming WANG. Tracking control of intelligent ship based on deep reinforcement learning[J]. Chinese Journal of Ship Research, 2021, 16(1): 105

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Apr. 29, 2020

    Accepted: --

    Published Online: Mar. 27, 2025

    The Author Email:

    DOI:10.19693/j.issn.1673-3185.01940

    Topics