摘要
无人艇具有高智能、低成本、耐单调、可执行危险任务和不会造成人员伤亡等优点,在海事监管、地质调查、环境监测、军事任务和水文调查等领域均具有广阔的应用前景.然而,无人艇在如何适应复杂海洋环境和执行多元化任务方面,仍面临着诸多挑战.近年来,以深度学习为代表的人工智能技术迅速发展,并被应用于无人艇的研发中.论文从归纳无人艇自主化所需具备的能力入手,简要回顾了深度学习的代表性技术,阐述了深度学习在无人艇多源感知、智能分析、动态决策以及精准控制等方面的应用前景,同时也探讨了深度学习在应用于无人艇上所面临的挑战和难点,为无人艇智能化研究提供参考.
Unmanned surface vehicle(USV)is suitable for boring and dangerous tasks because of its high intelligence,low cost and zero casualty.It has great application prospects in maritime supervision cruise,environmental monitoring,and marine military operations.However,USV is still faced with many challenges in a complex sea environment.Recently,artificial intelligence technology,especially deep learning,advances rapidly and is applied in many fields.This article described the intelligent capability required for achieving the autonomy of the USV,and reviewed several representative techniques of deep learning briefly,as well as analyzed the application prospect of deep learning in multi-source perception,intelligent analysis,dynamic analysis,decision-making and precise control of the USV.Challenges and difficulties of applying deep learning to USV are discussed.
作者
钱正芳
陆江
孙小帅
QIAN Zhengfang;LU Jiang;SUN Xiaoshuai(China Marine Development and Research Center,Beijing 100161,China)
出处
《中国造船》
EI
CSCD
北大核心
2020年第S01期6-13,共8页
Shipbuilding of China
关键词
无人艇
人工智能
深度学习
多源感知
精准控制
unmanned surface vehicle
artificial intelligence
deep learning
multi-source perception
precise control
作者简介
钱正芳,男,1973年生,工学博士,正高级工程师。主要从事智能船舶总体和流体力学方面的研究工作;陆江,男,1992年生,工学博士,工程师。主要从事人工智能和机器学习方面的研究工作;孙小帅,男,1990年生,工学博士,工程师。主要从事智能船舶总体和流体力学方面的研究工作。
