• Classified by Topic • Classified by Publication Type • Sorted by Date • Sorted by First Author Last Name • Classified by Funding Source •
Bottom-Up Skill Discovery from Unsegmented Demonstrations for Long-Horizon Robot Manipulation.
Yifeng
Zhu, Peter Stone, and Yuke
Zhu.
IEEE Robotics and Automation Letters (RA-L), 7:4126–33, April 2022.
Project page
Code
We tackle real-world long-horizon robot manipulation tasks through skill discovery. We present a bottom-up approach to learning a library of reusable skills from unsegmented demonstrations and use these skills to synthesize prolonged robot behaviors. Our method starts with constructing a hierarchical task structure from each demonstration through agglomerative clustering. From the task structures of multi-task demonstrations, we identify skills based on the recurring patterns and train goal-conditioned sensorimotor policies with hierarchical imitation learning. Finally, we train a meta controller to compose these skills to solve long-horizon manipulation tasks. The entire model can be trained on a small set of human demonstrations collected within 30 minutes without further annotations, making it amendable to real-world deployment. We systematically evaluated our method in simulation environments and on a real robot. Our method has shown superior performance over state-of-the-art imitation learning methods in multi-stage manipulation tasks. Furthermore, skills discovered from multi-task demonstrations boost the average task success by 8 percents compared to those discovered from individual tasks.
@article{ral2022-zhu,
author={Yifeng Zhu and Peter Stone and Yuke Zhu},
journal={IEEE Robotics and Automation Letters (RA-L)},
title={Bottom-Up Skill Discovery from Unsegmented Demonstrations for Long-Horizon Robot Manipulation},
year={2022},
doi={10.1109/LRA.2022.3146589},
month="April",
volume="7",issue="2",
pages="4126--33",
abstract={We tackle real-world long-horizon robot manipulation tasks through skill discovery. We present a bottom-up approach to learning a library of reusable skills from unsegmented demonstrations and use these skills to synthesize prolonged robot behaviors. Our method starts with constructing a hierarchical task structure from each demonstration through agglomerative clustering. From the task structures of multi-task demonstrations, we identify skills based on the recurring patterns and train goal-conditioned sensorimotor policies with hierarchical imitation learning. Finally, we train a meta controller to compose these skills to solve long-horizon manipulation tasks. The entire model can be trained on a small set of human demonstrations collected within 30 minutes without further annotations, making it amendable to real-world deployment. We systematically evaluated our method in simulation environments and on a real robot. Our method has shown superior performance over state-of-the-art imitation learning methods in multi-stage manipulation tasks. Furthermore, skills discovered from multi-task demonstrations boost the average task success by 8 percents compared to those discovered from individual tasks.},
wwwnote={<a href="https://ut-austin-rpl.github.io/rpl-BUDS/" target="_blank">Project page</a><br><a href="https://github.com/UT-Austin-RPL/BUDS">Code</a>}
}
Generated by bib2html.pl (written by Patrick Riley ) on Sun Nov 24, 2024 20:24:50