留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于机器学习的实海域无人艇避碰算法智能演进方法

楼建坤 王鸿东 王检耀 易宏

楼建坤, 王鸿东, 王检耀, 等. 基于机器学习的实海域无人艇避碰算法智能演进方法[J]. 中国舰船研究, 2021, 16(1): 65–73 doi: 10.19693/j.issn.1673-3185.02116
引用本文: 楼建坤, 王鸿东, 王检耀, 等. 基于机器学习的实海域无人艇避碰算法智能演进方法[J]. 中国舰船研究, 2021, 16(1): 65–73 doi: 10.19693/j.issn.1673-3185.02116
LOU J K, WANG H D, WANG J Y, et al. Intelligent evolution method for obstacle-avoidance algorithm of unmanned surface vehicles in real sea trial based on machine learning [J]. Chinese Journal of Ship Research, 2021, 16(1): 65–73 doi: 10.19693/j.issn.1673-3185.02116
Citation: LOU J K, WANG H D, WANG J Y, et al. Intelligent evolution method for obstacle-avoidance algorithm of unmanned surface vehicles in real sea trial based on machine learning [J]. Chinese Journal of Ship Research, 2021, 16(1): 65–73 doi: 10.19693/j.issn.1673-3185.02116

基于机器学习的实海域无人艇避碰算法智能演进方法

doi: 10.19693/j.issn.1673-3185.02116
基金项目: 国家自然科学基金资助项目(51909162)
详细信息
    作者简介:

    楼建坤,男,1997年生,博士生。研究方向:无人艇智能控制。E-mail:eternal1530@sjtu.edu.cn

    王鸿东,男,1989年生,副研究员,博士生导师。研究方向:基于海洋装备动力学特征的智能控制。E-mail:whd302@sjtu.edu.cn

    易宏,男,1962年生,教授,博士生导师。研究方向:舰船可靠性工程,海洋智能装备与系统。E-mail:yihong@sjtu.edu.cn

    通信作者:

    王鸿东

  • 中图分类号: U664.82

Intelligent evolution method for obstacle-avoidance algorithm of unmanned surface vehicles in real sea trial based on machine learning

知识共享许可协议
基于机器学习的实海域无人艇避碰算法智能演进方法楼建坤,等创作,采用知识共享署名4.0国际许可协议进行许可。
  • 摘要:   目的  无人艇(USV)效能是指在给定时间内的特定海域完成指定任务的能力,是多层次多技术节点耦合作用的结果,而针对单一技术节点的传统优化方法,对无人艇效能的提升效果有限。  方法  针对无人艇自主系统的特点,从智能算法的角度,提出无人艇智能演进的2种主要形式:一是算法函数;二是算法参数。在此基础上,给出基于机器学习的无人艇智能演进方法,设计一种可演进的无人艇自主系统控制体系架构,并在实海域测试。  结果  以无人艇避碰算法为例,基于实海域测试结果,初步验证了所提方法在提升无人艇效能方面的可行性与有效性。  结论  基于机器学习的无人艇智能演进方法是持续提升无人艇效能的有效途径,具有较高研究价值和应用意义。
  • 图  1  无人艇智能演进方法应用流程

    Figure  1.  The application process of USV intelligent evolution

    图  2  可演进的无人艇自主系统控制体系

    Figure  2.  The evolution-capable control architecture of USV autonomous system

    图  3  无人艇运动状态量传递模型

    Figure  3.  Transfer model of dynamic motion states of USV

    图  4  基于代理模型的算法参数演进流程

    Figure  4.  The evolution process of algorithm parameter based on the surrogated model

    图  5  “追梦三号”无人艇

    Figure  5.  Zhuimeng 3 USV

    图  6  虚拟障碍物设置

    Figure  6.  The virtual obstacle setting

    图  7  规划路径与实海域测试路径对比

    Figure  7.  The comparison between the planned paths and paths in real sea trial

    图  8  高斯代理模型拟合数据

    Figure  8.  The fitting data based on Gaussian surrogated model

    图  9  采用RRT算法基于最优参数集生成的避碰路径

    Figure  9.  The obstacle-avoidance path generated by RRT algorithm based on optimized parameters

    表  1  无人艇自主系统技术节点层次划分及相关特征

    Table  1.   The hierarchical division and corresponding features of tecnical node for USV autonomous system

    层次示例描述形式测评方法系统属性环境影响
    指标层推进功率物理量客观计算硬件层,单硬件作用不受环境影响
    行为层战术半径物理量客观计算硬件层,多硬件作用受环境影响较小
    功能层图像识别功能物理量与评分客观计算与主观描述子系统层,多算法作用受环境影响较大
    任务层巡逻任务评分主观描述与客观参考总体系统层受环境影响显著
    下载: 导出CSV

    表  2  虚拟障碍物坐标及相应避碰半径

    Table  2.   The location of virtual obstacles and corresponding avoiding radius

    障碍物圆心/m避碰半径/m障碍物圆心/m避碰半径/m
    (−180,180)80(−370,400)50
    (−310,250)35(−200,475)95
    (−150,330)30(−450,220)95
    下载: 导出CSV

    表  3  实海域试验路径测试结果

    Table  3.   The results of path in real sea trial

    步长D/m目标偏向概率P路径长度/m路径转角/(°)目标函数值
    150 0.1 739.274 3.687 79.001
    150 0.2 703.877 3.723 77.329
    150 0.3 748.569 3.656 79.226
    160 0.1 753.047 5.035 93.177
    160 0.25 654.352 3.719 74.450
    170 0.05 648.220 3.406 70.990
    170 0.15 636.477 3.118 67.457
    170 0.25 643.532 3.430 70.960
    180 0.25 872.046 3.976 89.484
    下载: 导出CSV
  • [1] 金克帆, 王鸿东, 易宏, 等. 海上无人装备关键技术与智能演进展望[J]. 中国舰船研究, 2018, 13(6): 1–8.

    JIN K F, WANG H D, YI H, et al. Key technologies and intelligence evolution of maritime UV[J]. Chinese Journal of Ship Research, 2018, 13(6): 1–8 (in Chinese).
    [2] 严新平, 刘佳仑, 范爱龙, 等. 智能船舶技术发展与趋势简述[J]. 船舶工程, 2020, 42(3): 15–20.

    YAN X P, LIU J L, FAN A L, et al. The development and tendency of intelligent vessel techniques[J]. Ship Engineering, 2020, 42(3): 15–20 (in Chinese).
    [3] LIU Z X, ZHANG Y M, YU X, et al. Unmanned surface vehicles: an overview of developments and challenges[J]. Annual Reviews in Control, 2016, 41: 71–93. doi: 10.1016/j.arcontrol.2016.04.018
    [4] 国防科学技术工业委员会. 装备费用−效能分析: GJB 1364-92[S]. 北京: 国防科学技术工业委员会, 1992.

    Commission of Science, Technology and Industry for National Defence. Cost-effectiveness analysis for materiel: GJB 1364-92[S]. Beijing: Commission of Science, Technology and Industry for National Defence, 1992 (in Chinese).
    [5] HOOTMAN J C, WHITCOMB D C. A military effectiveness analysis and decision making framework for naval ship design and acquisition[J]. Naval Engineers Journal, 2005, 117(3): 43–61.
    [6] BRISTOW D A, THARAYIL M, ALLEYNE A G. A survey of iterative learning control[J]. IEEE Control Systems Magazine, 2006, 26(3): 96–114. doi: 10.1109/MCS.2006.1636313
    [7] FOSSEN T I. Handbook of marine craft hydrodynamics and motion control[M]. Chichester, Hoboken: Wiley, 2011.
    [8] PRATT W K, ADAMS J E. Digital image processing[J]. Journal of Electronic Imaging, 2007, 16(2): 029901.
    [9] 欧阳子路, 王鸿东, 王检耀, 等. 基于改进Bi-RRT的无人水面艇自动避碰算法[J]. 中国舰船研究, 2019, 14(6): 8–14.

    OUYANG Z L, WANG H D, WANG J Y, et al. Automatic collision avoidance algorithm for unmanned surface vessel based on improved Bi-RRT algorithm[J]. Chinese Journal of Ship Research, 2019, 14(6): 8–14 (in Chinese).
    [10] 韩京清. 自抗扰控制技术[J]. 前沿科学, 2007(1): 24–31. doi: 10.3969/j.issn.1673-8128.2007.01.004

    HAN J Q. Auto disturbances rejection control technique[J]. Frontier Science, 2007(1): 24–31 (in Chinese). doi: 10.3969/j.issn.1673-8128.2007.01.004
    [11] RUSSELL S J, NORVIG P. Artificial intelligence: a modern approach (international edition)[M]. [S.1.]: Pearson Education Inc., 2002.
    [12] 余凯, 贾磊, 陈雨强, 等. 深度学习的昨天、今天和明天[J]. 计算机研究与发展, 2013, 50(9): 1799–1804. doi: 10.7544/issn1000-1239.2013.20131180

    YU K, JIA L, CHEN Y Q, et al. Deep learning: yesterday, today and tomorrow[J]. Journal of Computer Research and Development, 2013, 50(9): 1799–1804 (in Chinese). doi: 10.7544/issn1000-1239.2013.20131180
    [13] 王鸿东, 黄一, 赵恺, 等. 建设实海域智能船艇测试场亟待合力推进[J]. 中国船检, 2020(1): 64–67.

    WANG H D, HUANG Y, ZHAO K, et al. Work together to establish the intelligent ship test field in real sea area[J]. China Ship Survey, 2020(1): 64–67 (in Chinese).
    [14] BROOKS R A. A robust layered control system for a mobile robot[J]. IEEE Journal on Robotics and Automation, 1986, 2(1): 14–23.
    [15] FANG K T, LIN D K J, WINKER P, et al. Uniform design: theory and application[J]. Technometrics, 2000, 42(3): 237–248. doi: 10.1080/00401706.2000.10486045
    [16] ALPAYDIN E. Neural networks and deep learning[M]//ALPAYDIN E. Machine Learning: the New AI. Cambridge, Mass: MIT Press, 2016: 85–109.
    [17] HAN Z H, ZHANG K S. Surrogate-based optimization[M]//ROEVA O. Real-World Applications of Genetic Algorithms. Rijek, Crotia: InTech, 2012.
    [18] RASMUSSEN C E, WILLIAMS C K I. Gaussian processes for machine learning[M]. Cambridge, Mass: The MIT Press, 2006.
    [19] MALLIPEDDI R, LEE M. An evolving surrogate model-based differential evolution algorithm[J]. Applied Soft Computing, 2015, 34: 770–787. doi: 10.1016/j.asoc.2015.06.010
    [20] STEIN M L. A kernel approximation to the kriging predictor of a spatial process[J]. Annals of the Institute of Statistical Mathematics, 1991, 43(1): 61–75.
  • ZG2116_en.pdf
  • 加载中
图(9) / 表(3)
计量
  • 文章访问数:  865
  • HTML全文浏览量:  315
  • PDF下载量:  242
  • 被引次数: 0
出版历程
  • 收稿日期:  2020-09-17
  • 修回日期:  2020-11-20
  • 网络出版日期:  2021-01-18
  • 刊出日期:  2021-02-28

目录

    /

    返回文章
    返回