Chinese Journal of Ship Research, Volume. 20, Issue 1, 172(2025)

Autonomous decision-making method of unmanned ship based on improved DDPG algorithm

Wei GUAN... Shuhui HAO, Zhewen CUI and Miaomiao WANG |Show fewer author(s)
Author Affiliations
  • Navigation College, Dalian Maritime University, Dalian 116026, China
  • show less
    Figures & Tables(13)
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    [in Chinese]
    • Table 1. The principle parameters of the unmanned vehicle

      View table
      View in Article

      Table 1. The principle parameters of the unmanned vehicle

      参数数值
      船长/m52.5
      船宽/m8.6
      吃水深度/m2.29
      最大指令舵角/(°)35
      额定速度/kn14
    • Table 2. The parameters of improved DDPG algorithm and reward functions

      View table
      View in Article

      Table 2. The parameters of improved DDPG algorithm and reward functions

      参数数值参数数值
      折扣率$\gamma $0.99制导奖励权重${\lambda _{\rm{g}}}$5
      策略网络学习率0.000 3航向修正奖励权重${\lambda _{\rm{yaw}}}$−2.5
      价值网络学习率0.000 3COLREGs奖励权重$ \lambda_{\mathrm{C}\mathrm{ }} $6.5
      LSTM 隐藏层数量256雷达线长度$ {r_{\rm{1} }} $/n mile3.5
      软更新系数0.01船舶领域半径$ {r_{\rm{d}}} $/n mile1.4
      批次大小256角度权重$\varepsilon $0.23
      波风向/(°)31.4有效波高/m0.17
      波浪周期/s6.2平均风速/(m·s−1)0.38
    • Table 3. The initial data of two-ship encounter situation

      View table
      View in Article

      Table 3. The initial data of two-ship encounter situation

      会遇局面无人船起始位置/n mile目标位置/n mile速度/kn
      对遇OS(−1.0, −1.0)(5.0, 6.0)14.00
      TS(3.0, −3.0)(−2.0, −2.0)9.18
      右交叉OS(1.0, −4.0)(−1.0, 6.0)14.00
      TS(4.0, 0)(−4.0, 0)9.60
      左交叉OS(−4.0, 0)(1.0, 8.0)14.00
      TS(−6.0, 4.0)(2.0, 4.0)9.60
      追越OS(−1.0, −4.5)(−.01, 8.0)14.00
      TS(−1.0, −2.0)(−1.0, 4.0)6.00
      大角度右交叉OS(−2.2, −3.1)(0.8, 2.7)14.00
      TS(0.5, −3.4)(−0.2, 1.3)9.90
      被追越OS(4.5, 3.6)(4.5, 5.9)14.00
      TS(4.5, 2.6)(4.5, 6.8)19.00
    • Table 4. The initial information of there-ship encounter situation

      View table
      View in Article

      Table 4. The initial information of there-ship encounter situation

      无人船起始位置/n mile目标位置/n mile速度/kn
      OS(0, −1.20)(−0.40, 1.40)14.00
      TS1(−1.25, 1.70)(1.35, −0.40)13.75
      TS2(1.50, 0.60)(−1.00, −0.50)11.20
      TS3(−1.50, −0.80)(1.40, −0.80)11.90
    • Table 5. The trajectories comparison for two algorithms

      View table
      View in Article

      Table 5. The trajectories comparison for two algorithms

      算法实验次数路径长度/n mile决策时间/s最小会遇距离/n mile
      改进DDPG算法13.403875.00.760
      23.352862.00.900
      33.348861.00.650
      43.316853.00.730
      53.432882.00.840
      传统DDPG算法3.596924.60.620
      平均值3.370866.60.776
    Tools

    Get Citation

    Copy Citation Text

    Wei GUAN, Shuhui HAO, Zhewen CUI, Miaomiao WANG. Autonomous decision-making method of unmanned ship based on improved DDPG algorithm[J]. Chinese Journal of Ship Research, 2025, 20(1): 172

    Download Citation

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

    Category: Planning and Decision-making

    Received: May. 14, 2024

    Accepted: --

    Published Online: Mar. 13, 2025

    The Author Email:

    DOI:10.19693/j.issn.1673-3185.03929

    Topics