文献类型：期刊文献

英文题名：LLM-Based Heuristic Task Planning for Robotic Arm Proximity Operations with Non-Cooperative Spacecraft

作者：Jin, Yifeng[1]; Huang, Wenchao[1]; Yang, Chaonan[1]; Wu, Yanyan[2]; Yi, Jianjun[1]

机构：[1] Department of Mechanical Engineering, East China University of Science and Technology, Shanghai, China; [2] Department of Information Engineering, East China University of Science and Technology, Shanghai, China

年份：2025

期号：2025

起止页码：100

外文期刊名：International Conference on Intelligent Robotics and Control Engineering, IRCE

收录：EI(收录号：20260920152646)

语种：英文

外文关键词：Computation theory - Heuristic algorithms - Knowledge management - Orbits - Robotic arms - Spacecraft

摘要：When performing operational tasks such as capturing or repairing non-cooperative spacecraft, space robotic arms are typically constrained by incomplete prior knowledge and limited global visual observation conditions, which prevents them from autonomously completing tasks requiring structural reasoning about targets. This study focuses on typical space robotic arm operation tasks under partial visibility and incomplete prior knowledge. It establishes key structural recognition for non-cooperative spacecraft through multi-modal perceptual information and primitive perception algorithms. By leveraging the expert knowledge reasoning capabilities regarding common spacecraft structures constructed using Large Language Models (LLMs), the method guides space robotic arms to achieve heuristic active perception and incremental target understanding within limited visual fields, ultimately accomplishing specific operational tasks. Extensive experiments demonstrate that this approach achieves advanced performance across different types of operational tasks. ?2025 IEEE.