Discrete Policy: Learning Disentangled Action Space for Multi-Task Robotic Manipulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wu, Kun, Zhu, Yichen, Li, Jinming, Wen, Junjie, Liu, Ning, Xu, Zhiyuan, Tang, Jian
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913749047181312
author Wu, Kun
Zhu, Yichen
Li, Jinming
Wen, Junjie
Liu, Ning
Xu, Zhiyuan
Tang, Jian
author_facet Wu, Kun
Zhu, Yichen
Li, Jinming
Wen, Junjie
Liu, Ning
Xu, Zhiyuan
Tang, Jian
contents Learning visuomotor policy for multi-task robotic manipulation has been a long-standing challenge for the robotics community. The difficulty lies in the diversity of action space: typically, a goal can be accomplished in multiple ways, resulting in a multimodal action distribution for a single task. The complexity of action distribution escalates as the number of tasks increases. In this work, we propose \textbf{Discrete Policy}, a robot learning method for training universal agents capable of multi-task manipulation skills. Discrete Policy employs vector quantization to map action sequences into a discrete latent space, facilitating the learning of task-specific codes. These codes are then reconstructed into the action space conditioned on observations and language instruction. We evaluate our method on both simulation and multiple real-world embodiments, including both single-arm and bimanual robot settings. We demonstrate that our proposed Discrete Policy outperforms a well-established Diffusion Policy baseline and many state-of-the-art approaches, including ACT, Octo, and OpenVLA. For example, in a real-world multi-task training setting with five tasks, Discrete Policy achieves an average success rate that is 26\% higher than Diffusion Policy and 15\% higher than OpenVLA. As the number of tasks increases to 12, the performance gap between Discrete Policy and Diffusion Policy widens to 32.5\%, further showcasing the advantages of our approach. Our work empirically demonstrates that learning multi-task policies within the latent space is a vital step toward achieving general-purpose agents.
format Preprint
id arxiv_https___arxiv_org_abs_2409_18707
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Discrete Policy: Learning Disentangled Action Space for Multi-Task Robotic Manipulation
Wu, Kun
Zhu, Yichen
Li, Jinming
Wen, Junjie
Liu, Ning
Xu, Zhiyuan
Tang, Jian
Robotics
Learning visuomotor policy for multi-task robotic manipulation has been a long-standing challenge for the robotics community. The difficulty lies in the diversity of action space: typically, a goal can be accomplished in multiple ways, resulting in a multimodal action distribution for a single task. The complexity of action distribution escalates as the number of tasks increases. In this work, we propose \textbf{Discrete Policy}, a robot learning method for training universal agents capable of multi-task manipulation skills. Discrete Policy employs vector quantization to map action sequences into a discrete latent space, facilitating the learning of task-specific codes. These codes are then reconstructed into the action space conditioned on observations and language instruction. We evaluate our method on both simulation and multiple real-world embodiments, including both single-arm and bimanual robot settings. We demonstrate that our proposed Discrete Policy outperforms a well-established Diffusion Policy baseline and many state-of-the-art approaches, including ACT, Octo, and OpenVLA. For example, in a real-world multi-task training setting with five tasks, Discrete Policy achieves an average success rate that is 26\% higher than Diffusion Policy and 15\% higher than OpenVLA. As the number of tasks increases to 12, the performance gap between Discrete Policy and Diffusion Policy widens to 32.5\%, further showcasing the advantages of our approach. Our work empirically demonstrates that learning multi-task policies within the latent space is a vital step toward achieving general-purpose agents.
title Discrete Policy: Learning Disentangled Action Space for Multi-Task Robotic Manipulation
topic Robotics
url https://arxiv.org/abs/2409.18707