空间控制技术与应用2025，Vol.51Issue(6)：63-74,12.DOI:10.3969/j.issn.1674-1579.2025.06.006
基于大模型对话协商的空间装配人机协作方法
A Dialogue-Based Negotiation Method for Human-Robot Collaboration in Space Assembly Using Large Language Models
摘要
Abstract
To address the challenges of efficiency,safety,and robustness in human-robot collaborative assembly tasks within complex space environments,a human-robot collaborative assembly method that centers on large model planning and dialogue negotiation is proposed.In this method,initial task planning and allocation are first performed by a large model integrated with scene understanding.Subsequently,a human-robot dialogue mechanism based on the large model is established to enable natural language communication,through which specific action steps are collaboratively planned and refined by both humans and robots.The innovations of this method are threefold.First,a large language model is utilized for communicative planning,whereby the robustness and flexibility of human-robot collaboration are enhanced.The core intent of the task can thus be more clearly comprehended by both parties,which effectively addresses the complex and variable nature of on-orbit operating conditions.Second,dialogue-based negotiation is supported during task execution,by which the risk of operational errors is mitigated and the stringent safety requirements of space operations are met.The feasibility and effectiveness of the proposed method were validated through simulations and physical experiments that involved a mock space camera assembly task.The significant potential of large language model technology to advance the intelligence and capabilities of human-robot collaboration in complex space assembly missions is demonstrated by these results.关键词
空间站/人机协作/大语言模型/相机装配/任务规划/智能交互Key words
space station/human-robot collaboration/large language models/camera assembly/task planning/intelligent interaction分类
信息技术与安全科学引用本文复制引用
LYU Wenxiang,CAI Yinghao,LI Yuhan,LIU Nailong,ZHOU Xiaodong,WANG Zhenyu,LU Tao..基于大模型对话协商的空间装配人机协作方法[J].空间控制技术与应用,2025,51(6):63-74,12.基金项目
思源联盟开放基金资助项目和国家自然科学基金资助项目(62473366) Siyuan Alliance Open Ended Fund and National Natural Science Foundation of China(62473366) （62473366）
