SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Models
and real-world setups, where the pre-learned action grids are re-discretized to capture robot-specific spatial action movements of new setups.The superior results from extensive evaluations demonstrate the exceptional in-distribution generalization and out-of-distribution adaptation capability, highlighting the crucial benefit of the proposed spatial-aware representations for generalist robot policy learning.All the details and codes are opensourced.
Authors
- Zhigang Wang (ORCID: https://orcid.org/0000-0002-8711-235X)
- Yan Ding (ORCID: https://orcid.org/0000-0002-7949-4351)
- Qizhi Chen (ORCID: https://orcid.org/0000-0002-8661-5596)
- Bin Zhao (ORCID: https://orcid.org/0009-0006-6965-1903)
- Haoming Song (ORCID: https://orcid.org/0009-0003-3759-8966)
- Dong Wang (ORCID: https://orcid.org/0009-0008-7738-1688)
- Xinyi Ye (ORCID: https://orcid.org/0009-0009-1126-3336)
- Jiayuan Gu
- Yuanqi Yao
- Delin Qu
Publication Details
- Published
- 2025-06-21
- DOI
- https://doi.org/10.15607/rss.2025.xxi.011
- Citations
- 12
- Primary Topic
- Multimodal Machine Learning Applications
- Type
- article
- Field-Weighted Citation Impact
- 15.31
Funders
- National Natural Science Foundation of China