Robot Navigation via Foundation Language Models: A Review
Recently, with advances in Large Language Models(LLMs), robot navigation models have demonstrated superior generalization capabilities across environment perception, decision-making, reasoning, planning, instruction understanding, and human-robot interaction. In this paper, we systematically review recent LLM-based robot navigation research papers and categorize them into a novel taxonomy comprising perception, planning, control, interaction, and coordination. We also present an overview of the principal datasets, simulations, and metrics used in robot navigation, analyzing the distinctive characteristics of the datasets and the performance of the main LLM-based methods. Furthermore, we discuss the challenges hindering the integration of LLMs into robot navigation and provide opportunities and potential directions for future development.
Authors
- Muyu Wang (ORCID: https://orcid.org/0009-0002-8009-3966)
- Jinpeng Mi (ORCID: https://orcid.org/0000-0002-0506-9707)
- Jinqun Zhou
- Lin Zhao (ORCID: https://orcid.org/0000-0003-4498-6054)
- Ye Liu (ORCID: https://orcid.org/0000-0002-8473-6538)
- Li Zhang (ORCID: https://orcid.org/0000-0003-1152-8962)
- Xi Zhang (ORCID: https://orcid.org/0000-0002-4638-3997)
- Peidong Liang
- Xian Wei
- Ke Li
- Jian Yang
- Xiong You
- Jinbo Yang
- Shibo Huang
- Haotian Pan
- Jie Yang
- Mingsong Chen
Institutions
- University of Shanghai for Science and Technology (CN)
- PLA Information Engineering University (CN)
- Harbin Institute of Technology (CN)
- Machine Intelligence Research Institute (US)
- Embedded Systems (United States) (US)
- Minnan University of Science and Technology (CN)
- East China Normal University (CN)
Publication Details
- Journal
- ACM Computing Surveys
- Published
- 2026-03-17
- DOI
- https://doi.org/10.1145/3802539
- Citations
- 1
- Primary Topic
- Multimodal Machine Learning Applications
- Type
- article
- Field-Weighted Citation Impact
- 46.84