面向认知赋能的人机协作:进展、挑战和展望  

Cognitive Empowerment for Human-robot Collaboration: Research Progress and Challenges

作　　者：寇逸群 杨晔 刘颉[2] 胡友民[1] 李林 俞百川 徐家和 胡中旭 史铁林[1] KOU Yiqun;YANG Ye;LIU Jie;HU Youmin;LI Lin;YU Baichuan;XU Jiahe;HU Zhongxu;SHI Tielin(School of Mechanical Science&Engineering,Huazhong University of Science and Technology,Wuhan 430074;School of Civil and Hydraulic Engineering,Huazhong University of Science and Technology,Wuhan 430074;Nari Group Corporation,State Grid Electric Power Research Institute,Wuhan 430223)

机构地区：[1]华中科技大学机械科学与工程学院,武汉430074 [2]华中科技大学土木与水利工程学院,武汉430074 [3]国网电力科学研究院武汉南瑞有限责任公司,武汉430223

出　　处：《机械工程学报》2025年第3期1-22,共22页Journal of Mechanical Engineering

基　　金：国家自然科学基金(52205104);国家重点研发计划(2023YFD2100905,2024YFE030008)资助项目。

摘　　要：在工业4.0向工业5.0的发展过程中,以人为本逐渐成为智能制造领域关注的焦点之一。当前的人机协作不仅强调要聚焦于技术的进步与效率的提升,更强调将人类的高阶认知思维与机器的计算能力相结合,实现认知赋能。基于此,梳理人机协作中认知赋能在交互感知、任务规划与执行、技能学习等关键领域的现有研究,揭示了多模态信息整合、任务推理、动态决策与技能知识表征的挑战。进一步,提出通过应用知识图谱构建的相关技术来支持人与其机器认知对齐的方法,以及通过应用知识图谱推理的相关技术来支持复杂环境下人机协作的任务优化和动态决策。在分析现有人机协作认知赋能研究局限性的基础上,展望未来智能制造环境下的深度认知协同的发展方向。In the transition from Industry 4.0 to Industry 5.0,a human-centered approach has gradually emerged as a focal point in the field of smart manufacturing.Current human-machine collaboration not only emphasizes technological advancements and efficiency improvements but also stresses the integration of human higher-order cognitive thinking with machine computational capabilities to achieve cognitive empowerment.Based on this premise,this study reviews existing research on cognitive empowerment in human-machine collaboration,focusing on key areas such as interactive perception,task planning and execution,and skill learning.The challenges of multimodal information integration,task reasoning,dynamic decision-making,and skill knowledge representation are highlighted.Furthermore,methods are proposed to support human-machine cognitive using knowledge graph construction technologies,as well as to optimize tasks and facilitate dynamic decision-making in complex environments through the application of knowledge graph reasoning techniques.Building upon an analysis of the limitations in current research on cognitive empowerment in human-machine collaboration,this study also forecasts the future directions for deep cognitive collaboration within intelligent manufacturing environments.

关 键 词：人机协作 认知赋能 知识图谱 知识表示 多模态感知 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]