UMI-Underwater: Learning Underwater Manipulation without Underwater Teleoperation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914429068640256 |
|---|---|
| author | Li, Hao Chung, Long Yin Goler, Jack Zhang, Ryan Xie, Xiaochi Ha, Huy Song, Shuran Cutkosky, Mark |
| author_facet | Li, Hao Chung, Long Yin Goler, Jack Zhang, Ryan Xie, Xiaochi Ha, Huy Song, Shuran Cutkosky, Mark |
| contents | Underwater robotic grasping is difficult due to degraded, highly variable imagery and the expense of collecting diverse underwater demonstrations. We introduce a system that (i) autonomously collects successful underwater grasp demonstrations via a self-supervised data collection pipeline and (ii) transfers grasp knowledge from on-land human demonstrations through a depth-based affordance representation that bridges the on-land-to-underwater domain gap and is robust to lighting and color shift. An affordance model trained on on-land handheld demonstrations is deployed underwater zero-shot via geometric alignment, and an affordance-conditioned diffusion policy is then trained on underwater demonstrations to generate control actions. In pool experiments, our approach improves grasping performance and robustness to background shifts, and enables generalization to objects seen only in on-land data, outperforming RGB-only baselines. Code, videos, and additional results are available at https://umi-under-water.github.io. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_27012 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | UMI-Underwater: Learning Underwater Manipulation without Underwater Teleoperation Li, Hao Chung, Long Yin Goler, Jack Zhang, Ryan Xie, Xiaochi Ha, Huy Song, Shuran Cutkosky, Mark Robotics Artificial Intelligence Underwater robotic grasping is difficult due to degraded, highly variable imagery and the expense of collecting diverse underwater demonstrations. We introduce a system that (i) autonomously collects successful underwater grasp demonstrations via a self-supervised data collection pipeline and (ii) transfers grasp knowledge from on-land human demonstrations through a depth-based affordance representation that bridges the on-land-to-underwater domain gap and is robust to lighting and color shift. An affordance model trained on on-land handheld demonstrations is deployed underwater zero-shot via geometric alignment, and an affordance-conditioned diffusion policy is then trained on underwater demonstrations to generate control actions. In pool experiments, our approach improves grasping performance and robustness to background shifts, and enables generalization to objects seen only in on-land data, outperforming RGB-only baselines. Code, videos, and additional results are available at https://umi-under-water.github.io. |
| title | UMI-Underwater: Learning Underwater Manipulation without Underwater Teleoperation |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2603.27012 |