Saved in:
Bibliographic Details
Main Authors: Lu, Guanxing, Zhang, Shiyi, Wang, Ziwei, Liu, Changliu, Lu, Jiwen, Tang, Yansong
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.08321
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911960098930688
author Lu, Guanxing
Zhang, Shiyi
Wang, Ziwei
Liu, Changliu
Lu, Jiwen
Tang, Yansong
author_facet Lu, Guanxing
Zhang, Shiyi
Wang, Ziwei
Liu, Changliu
Lu, Jiwen
Tang, Yansong
contents Performing language-conditioned robotic manipulation tasks in unstructured environments is highly demanded for general intelligent robots. Conventional robotic manipulation methods usually learn semantic representation of the observation for action prediction, which ignores the scene-level spatiotemporal dynamics for human goal completion. In this paper, we propose a dynamic Gaussian Splatting method named ManiGaussian for multi-task robotic manipulation, which mines scene dynamics via future scene reconstruction. Specifically, we first formulate the dynamic Gaussian Splatting framework that infers the semantics propagation in the Gaussian embedding space, where the semantic representation is leveraged to predict the optimal robot action. Then, we build a Gaussian world model to parameterize the distribution in our dynamic Gaussian Splatting framework, which provides informative supervision in the interactive environment via future scene reconstruction. We evaluate our ManiGaussian on 10 RLBench tasks with 166 variations, and the results demonstrate our framework can outperform the state-of-the-art methods by 13.1\% in average success rate. Project page: https://guanxinglu.github.io/ManiGaussian/.
format Preprint
id arxiv_https___arxiv_org_abs_2403_08321
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ManiGaussian: Dynamic Gaussian Splatting for Multi-task Robotic Manipulation
Lu, Guanxing
Zhang, Shiyi
Wang, Ziwei
Liu, Changliu
Lu, Jiwen
Tang, Yansong
Robotics
Computer Vision and Pattern Recognition
Performing language-conditioned robotic manipulation tasks in unstructured environments is highly demanded for general intelligent robots. Conventional robotic manipulation methods usually learn semantic representation of the observation for action prediction, which ignores the scene-level spatiotemporal dynamics for human goal completion. In this paper, we propose a dynamic Gaussian Splatting method named ManiGaussian for multi-task robotic manipulation, which mines scene dynamics via future scene reconstruction. Specifically, we first formulate the dynamic Gaussian Splatting framework that infers the semantics propagation in the Gaussian embedding space, where the semantic representation is leveraged to predict the optimal robot action. Then, we build a Gaussian world model to parameterize the distribution in our dynamic Gaussian Splatting framework, which provides informative supervision in the interactive environment via future scene reconstruction. We evaluate our ManiGaussian on 10 RLBench tasks with 166 variations, and the results demonstrate our framework can outperform the state-of-the-art methods by 13.1\% in average success rate. Project page: https://guanxinglu.github.io/ManiGaussian/.
title ManiGaussian: Dynamic Gaussian Splatting for Multi-task Robotic Manipulation
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.08321