文献类型：期刊文献

英文题名：Formation control and path planning of multi-robot systems via large language models

作者：Xue, Dong[1];Zhou, Xuanjie[1];Wang, Ming[1];Liu, Fangzhou[2]

机构：[1]East China Univ Sci & Technol, Key Lab Smart Mfg Energy Chem Proc, Minist Educ, Shanghai 200237, Peoples R China;[2]Harbin Inst Technol, Res Inst Intelligent Control & Syst, Sch Astronaut, Harbin 150001, Peoples R China

年份：2025

卷号：68

期号：5

外文期刊名：SCIENCE CHINA-INFORMATION SCIENCES

收录：;EI(收录号：20251818339476);WOS:【SCI-EXPANDED(收录号:WOS:001479679700007)】；

基金：This work was supported in part by National Key Research and Development Program of China (Grant No. 2023YFB3307800), National Natural Science Foundation of China (Grant Nos. 62173147, 62373123), Major Science and Technology Project of Xinjiang (Grant No. 2022A01006-4), and Fundamental Research Funds for the Centra Universities (Grant No. 222202517006).

语种：英文

外文关键词：multi-robot systems; formation control; path planning; obstacle avoidance; large language models

摘要：Existing path planning and coordination control methods for multi-robot systems (MRS) typically rely on predefined rules and rudimentary algorithms. However, these methods often struggle to adapt flexibly to complex environments and to adjust motion targets appropriately. To address this challenge, this study presents a large language model (LLM)-assisted framework. By integrating textual descriptions of complex motion constraints, robot information, and local environmental data as inputs, LLMs generate motion objectives and translate them into executable control commands for the robots, thereby achieving coordinated control and path planning. This framework facilitates the generation, maintenance, and reshaping of formations in MRSs during path planning, applicable to both obstacle-free and obstacle-avoidance environments. Simulation results demonstrate that LLM-based control strategies enhance the autonomy, adaptability, flexibility, and robustness of MRS by processing complex information, making intelligent decisions, adapting to environmental changes, and handling disturbances and uncertainties.