V-MAO: Generative Modeling for Multi-Arm Manipulation of Articulated Objects
Abstract
Manipulating articulated objects requires multiple robot arms in general. It is challenging to enable multiple robot arms to collaboratively complete manipulation tasks on articulated objects. In this paper, we present V-MAO, a framework for learning multi-arm manipulation of articulated objects. Our framework includes a variational generative model that learns contact point distribution over object rigid parts for each robot arm. The training signal is obtained from interaction with the simulation environment which is enabled by planning and object-centric control. We deploy our framework in a customized MuJoCo simulation environment and demonstrate that our framework achieves a high success rate on six different objects and two different robots. We also show that generative modeling can effectively learn the contact point distribution on articulated objects.
Links
Paper with reviews: Openreview
Arxiv: Arxiv
Contact: xingyul3 {at} cs {at} cmu {dot} edu for more information
Poster
Video
Bibtex
@inproceedings{liu2021v-mao,
title = {V-MAO: Generative Modeling for Multi-Arm Manipulation of Articulated Objects},
author = {Xingyu Liu and Shun Iwase and Kris M. Kitani},
booktitle = {The Conference on Robot Learning (CoRL)},
year = {2021},
}