深度神经网络在船舶自动舵中的应用Application of deep neural networks in ship autopilot rudder
李少伟,王胜正
摘要(Abstract)：
为了改进现有船舶自动舵的控制精度,提高自动舵的自适应能力,提出一种基于深度置信网络(DBN)的自动舵控制算法。首先,利用对比散度算法,结合上海海事大学高级船员考试系统中记录的数据,对组成DBN的每一层受限玻尔兹曼机(RBM)模型依次进行预训练,并将结果作为深度神经网络权重的初值。在此基础上,使用反向传播算法,进行多层深度结构的微调训练。仿真实验表明,该方法与资深船长的模拟操船误差仅为5.2%。
关键词(KeyWords)： 自动舵;深度置信网络;对比散度算法;受限波尔兹曼机;深度神经网络;反向传播算法
基金项目(Foundation): 国家自然科学基金(51379121);国家自然科学基金(61304230);; 上海市曙光人才计划项目资助(15SG44);; 江汉大学博士启动基金(1008-06600001);; 湖北省教育厅科学研究计划指导性项目(B2018254)~~
作者(Author): 李少伟,王胜正
DOI: 10.16652/j.issn.1004-373x.2018.24.010
参考文献(References)：
- [1]乔富强.基于DSP的舰船PID自动舵监控系统的研究[J].舰船科学技术,2017,39(14):127-129.QIAO Fuqiang. Research on PID automatic rudder monitoring system based on DSP technology[J]. Ship science and technology,2017,39(14):127-129.
- [2] KULA K S. Model-based controller for ship track-keeping using neural network[C]//Proceedings of 2nd International Conference on Cybernetics. Gdynia:IEEE,2015:178-183.
- [3] HUO Xingxing,HU Jiangqiang,YU Ting. Fuzzy PID autopilot for ship steering based on stratification and sub fuzzy controller[J]. Journal of Dalian Maritime University,2013,39(4):71-74.
- [4]常静.基于BP神经网络的PID在船舶自动舵中的应用与研究[J].舰船科学技术,2017,39(6):105-107.CHANG Jing. Application and research of the fuzzy PID autopilot based on BP neural network[J]. Ship science and technology,2017,39(6):105-107.
- [5] WANG Renqiang,ZHAO Yuelin,SUN Jianming. Application of optimized RBF neural network in ship′s autopilot design[C]//Proceedings of IEEE Advanced Information Management,Communicates,Electronic and Automation Control Conference.Xi’an:IEEE,2017:1642-1646.
- [6] WANG Y Y,NGUYEN H D,CHAI S,et al. Radial basis function neural network based rudder roll stabilization for ship sailing in waves[C]//Proceedings of 5th Australian Control Conference. Gold Coast:IEEE,2015:158-163.
- [7] WANG M H,YU Y Q,ZENG B. Study of intelligent control system for ship autopilot design[C]//Proceedings of 8th IEEE/ACIS International Conference on Computer and Information Science. Shanghai:IEEE,2009:701-705.
- [8] WU Y X,XUE M J,WU P J. Study on fuzzy neural networkbased ship autopilot[C]//Proceedings of 6th International Conference on Natural Computation. Yantai:IEEE,2010:1393-1396.
- [9] KAPITANYUK Y A,PROSKURNIKOV A V,CAO M. Optimal controllers for rudder roll damping with an autopilot in the loop[J]. IFAC-papersonline,2016,49(23):562-567.
- [10] HINTON G E,OSINDERO S,TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural computation,2006,18(7):1527-1554.
- [11] BENGIO Y. Learning deep architectures for AI[J]. Foundations&trends in machine learning,2009,2(1):1-127.
- [12]刘建伟,刘媛,罗雄麟.深度学习研究进展[J].计算机应用研究,2014,31(7):1921-1930.LIU Jianwei,LIU Yuan,LUO Xionglin. Research and development on deep learning[J]. Application research of computers,2014,31(7):1921-1930.
- [13] LAROCHELLE H,BENGIO Y,LOURADOUR J,et al. Exploring strategies for training deep neural networks[J]. Journal of machine learning research,2009,10:1-40.
- [14] PALM R B. Deep learn toolbox[EB/OL].[2018-01-04]. http://www.pudn.com/Download/item/id/3341483.html.