Hierarchical World Models as Visual Whole-Body Humanoid Controllers

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hansen, Nicklas, S V, Jyothir, Sobal, Vlad, LeCun, Yann, Wang, Xiaolong, Su, Hao
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913837465206784
author Hansen, Nicklas
S V, Jyothir
Sobal, Vlad
LeCun, Yann
Wang, Xiaolong
Su, Hao
author_facet Hansen, Nicklas
S V, Jyothir
Sobal, Vlad
LeCun, Yann
Wang, Xiaolong
Su, Hao
contents Whole-body control for humanoids is challenging due to the high-dimensional nature of the problem, coupled with the inherent instability of a bipedal morphology. Learning from visual observations further exacerbates this difficulty. In this work, we explore highly data-driven approaches to visual whole-body humanoid control based on reinforcement learning, without any simplifying assumptions, reward design, or skill primitives. Specifically, we propose a hierarchical world model in which a high-level agent generates commands based on visual observations for a low-level agent to execute, both of which are trained with rewards. Our approach produces highly performant control policies in 8 tasks with a simulated 56-DoF humanoid, while synthesizing motions that are broadly preferred by humans.
format Preprint
id arxiv_https___arxiv_org_abs_2405_18418
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Hierarchical World Models as Visual Whole-Body Humanoid Controllers
Hansen, Nicklas
S V, Jyothir
Sobal, Vlad
LeCun, Yann
Wang, Xiaolong
Su, Hao
Machine Learning
Computer Vision and Pattern Recognition
Robotics
Whole-body control for humanoids is challenging due to the high-dimensional nature of the problem, coupled with the inherent instability of a bipedal morphology. Learning from visual observations further exacerbates this difficulty. In this work, we explore highly data-driven approaches to visual whole-body humanoid control based on reinforcement learning, without any simplifying assumptions, reward design, or skill primitives. Specifically, we propose a hierarchical world model in which a high-level agent generates commands based on visual observations for a low-level agent to execute, both of which are trained with rewards. Our approach produces highly performant control policies in 8 tasks with a simulated 56-DoF humanoid, while synthesizing motions that are broadly preferred by humans.
title Hierarchical World Models as Visual Whole-Body Humanoid Controllers
topic Machine Learning
Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2405.18418