LeLaN: Learning A Language-Conditioned Navigation Policy from In-the-Wild Videos
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914964765147136 |
|---|---|
| author | Hirose, Noriaki Glossop, Catherine Sridhar, Ajay Shah, Dhruv Mees, Oier Levine, Sergey |
| author_facet | Hirose, Noriaki Glossop, Catherine Sridhar, Ajay Shah, Dhruv Mees, Oier Levine, Sergey |
| contents | The world is filled with a wide variety of objects. For robots to be useful, they need the ability to find arbitrary objects described by people. In this paper, we present LeLaN(Learning Language-conditioned Navigation policy), a novel approach that consumes unlabeled, action-free egocentric data to learn scalable, language-conditioned object navigation. Our framework, LeLaN leverages the semantic knowledge of large vision-language models, as well as robotic foundation models, to label in-the-wild data from a variety of indoor and outdoor environments. We label over 130 hours of data collected in real-world indoor and outdoor environments, including robot observations, YouTube video tours, and human walking data. Extensive experiments with over 1000 real-world trials show that our approach enables training a policy from unlabeled action-free videos that outperforms state-of-the-art robot navigation methods, while being capable of inference at 4 times their speed on edge compute. We open-source our models, datasets and provide supplementary videos on our project page (https://learning-language-navigation.github.io/). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_03603 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | LeLaN: Learning A Language-Conditioned Navigation Policy from In-the-Wild Videos Hirose, Noriaki Glossop, Catherine Sridhar, Ajay Shah, Dhruv Mees, Oier Levine, Sergey Robotics The world is filled with a wide variety of objects. For robots to be useful, they need the ability to find arbitrary objects described by people. In this paper, we present LeLaN(Learning Language-conditioned Navigation policy), a novel approach that consumes unlabeled, action-free egocentric data to learn scalable, language-conditioned object navigation. Our framework, LeLaN leverages the semantic knowledge of large vision-language models, as well as robotic foundation models, to label in-the-wild data from a variety of indoor and outdoor environments. We label over 130 hours of data collected in real-world indoor and outdoor environments, including robot observations, YouTube video tours, and human walking data. Extensive experiments with over 1000 real-world trials show that our approach enables training a policy from unlabeled action-free videos that outperforms state-of-the-art robot navigation methods, while being capable of inference at 4 times their speed on edge compute. We open-source our models, datasets and provide supplementary videos on our project page (https://learning-language-navigation.github.io/). |
| title | LeLaN: Learning A Language-Conditioned Navigation Policy from In-the-Wild Videos |
| topic | Robotics |
| url | https://arxiv.org/abs/2410.03603 |