摘要
机器人正朝着智能化方向快速发展,深度学习作为人工智能最前沿的一个分支,其突破性发展促进了人工智能领域的爆发性增长,已广泛应用于机器人学的感知、决策、控制等相关技术领域。为加强科技发展战略谋划和系统布局,利用SciVal文献分析平台对深度学习与机器人学的交叉学科的发展态势进行了系统梳理和分析,并对深度学习重要算法的应用进行了综述。结果表明:中国的学术产出已达第一,但学术影响力落后于欧美,并缺少高产出的学者;当前的研究重点是计算机视觉,研究热点是卷积神经网络用于机器人对象检测及深度Q学习用于辅助决策等;深度学习元学习、分布式深度学习、三维深度CNN、深度学习融合应用展现出有良好的前景,将推动机器人学向人机混合的增强智能与智能自主化方向发展;通过比较中国与欧美发达国家的优劣势,提出了相应的对策建议。研究结果展现了深度学习与机器人学交叉学科领域的全貌,为研究人员提供了有价值的参考。
As one of the most cutting-edge technologies,deep learning has promoted the explosive growth of artificial intelligence and has been widely used in the perception,decision-making and control of robots.In order to strengthen the strategic planning and systematic layout of science and technology development,the trends and hotspots of deep learning in robotics is systematically analyzed using SciVal,and the applications of deep learning algorithms are reviewed.The results show that Chinese scholarly output has reached the first place in the world,but its academic influence lags behind that of US and Europe,and lacks scholars with high output.The current research focuses on computer vision,and the research hotspots are convolutional neural network for robotic object detection and deep Q-learning for decision assistance.Deep learning meta-learning,distributed deep learning,3D deep CNN,and deep learning fusion applications show promising prospects that will drive robotics toward human-machine hybrid augmented intelligence and intelligent autonomy.Through the comparative study of the advantages and disadvantages of China and other countries,the corresponding countermeasures and suggestions are put forward.The results present the full picture of the interdisciplinary field of deep learning and robotics,and provide a valuable reference for researchers.
作者
仇秋飞
周武源
雷良育
吴叶青
崔银江
陈登
QIU Qiu-fei;ZHOU Wu-yuan;LEI Liang-yu;WU Ye-qing;CUI Yin-jiang;CHEN Deng(Zhejiang Academy of Science and Technology Information,Hangzhou 310006,China;School of Engineering,Zhejiang A&F University,Hangzhou 310006,China)
出处
《计算机技术与发展》
2021年第11期208-215,共8页
Computer Technology and Development
基金
国家自然科学基金面上项目(61976192)
浙江省软科学研究计划项目(2019C35016,2019C35070)。
