摘要
工业5.0强调以人为本、弹性制造和可持续性,促进了工业个性化产品发展,此类产品灵活多变的非结构化生产环境需要高度智能的人机协作。当前“空间感知-过程分析-决策推理-任务执行”框架驱动的人机协作方法已取得显著进展,但仍受限于交互语义理解不足和工艺知识缺失,难以应对复杂多变的制造环境。因此,提出工艺知识和交互语义增强的主动式人机协作方法。为增强空间感知能力,设计知识增强自适应微调方法,提升大模型对制造领域知识的理解与迁移。为增强过程分析与决策推理能力,构建操作员意图多样化预测模型,深化人机交互分析。为增强任务执行能力,设计深度强化学习方法,实现持续学习与任务规划。在人机协作装配案例中,所提方法可以根据操作员指令和装配空间感知,保证机器人成功完成任务,可证明其可行性与有效性。
Industry 5.0 emphasizes human centricity,flexible manufacturing and sustainability,fostering the advancement of personalized industrial products.The unstructured and dynamically changing production environments of such products demand highly intelligent Human-Robot Collaboration(HRC).Although existing HRC frameworks driven by the“spatial perception-process analysis-decision reasoning-task execution”paradigm had made significant progress,they remain constrained by limited interaction semantic understanding and insufficient process knowledge,making it difficult to adapt to complex and variable manufacturing conditions.To address these challenges,a process knowledge and interaction semantics-enhanced proactive human-robot collaboration method was proposed.Specifically,a knowledge-augmented adaptive fine-tuning approach was developed to enhance spatial perception and improve large models’understanding and transferability of manufacturing knowledge.A diverse operator intention prediction model was constructed to strengthen process analysis and decision reasoning by deepening human-robot interaction understanding.Furthermore,a deep reinforcement learning-based approach was designed to improve task execution through continual learning and task planning.In a human-robot collaborative assembly case,the proposed method enabled robots to successfully complete tasks by interpreting operator instructions and perceiving the assembly space,demonstrating its feasibility and effectiveness.
作者
丁鹏飞
张洁
张朋
DING Pengfei;ZHANG Jie;ZHANG Peng(College of Mechanical Engineering,Donghua University,Shanghai 201620,China;Engineering Research Center of Artificial Intelligence for Textile Industry,Ministry of Education,Institute of Artificial Intelligence,Donghua University,Shanghai 201620,China;Shanghai Engineering Research Center of Industrial Big Data and Intelligent System,Institute of Artificial Intelligence,Donghua University,Shanghai 201620,China)
出处
《计算机集成制造系统》
2025年第12期4542-4555,共14页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金面上资助项目(52375485)
国家自然科学基金资助项目(52005099)
国家自然科学基金区域创新发展联合基金重点支持资助项目(U24A20262)。
关键词
工业5.0
智能制造
人本智造
主动式人机协作
Industry 5.0
smart manufacturing
human-centric smart manufacturing
proactive human-robot collaboration
