Salvato in:
Dettagli Bibliografici
Autori principali: Han, Lei, Zhu, Qingxu, Sheng, Jiapeng, Zhang, Chong, Li, Tingguang, Zhang, Yizheng, Zhang, He, Liu, Yuzhen, Zhou, Cheng, Zhao, Rui, Li, Jie, Zhang, Yufeng, Wang, Rui, Chi, Wanchao, Li, Xiong, Zhu, Yonghui, Xiang, Lingzhu, Teng, Xiao, Zhang, Zhengyou
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:https://arxiv.org/abs/2308.15143
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917713644879872
author Han, Lei
Zhu, Qingxu
Sheng, Jiapeng
Zhang, Chong
Li, Tingguang
Zhang, Yizheng
Zhang, He
Liu, Yuzhen
Zhou, Cheng
Zhao, Rui
Li, Jie
Zhang, Yufeng
Wang, Rui
Chi, Wanchao
Li, Xiong
Zhu, Yonghui
Xiang, Lingzhu
Teng, Xiao
Zhang, Zhengyou
author_facet Han, Lei
Zhu, Qingxu
Sheng, Jiapeng
Zhang, Chong
Li, Tingguang
Zhang, Yizheng
Zhang, He
Liu, Yuzhen
Zhou, Cheng
Zhao, Rui
Li, Jie
Zhang, Yufeng
Wang, Rui
Chi, Wanchao
Li, Xiong
Zhu, Yonghui
Xiang, Lingzhu
Teng, Xiao
Zhang, Zhengyou
contents Knowledge from animals and humans inspires robotic innovations. Numerous efforts have been made to achieve agile locomotion in quadrupedal robots through classical controllers or reinforcement learning approaches. These methods usually rely on physical models or handcrafted rewards to accurately describe the specific system, rather than on a generalized understanding like animals do. Here we propose a hierarchical framework to construct primitive-, environmental- and strategic-level knowledge that are all pre-trainable, reusable and enrichable for legged robots. The primitive module summarizes knowledge from animal motion data, where, inspired by large pre-trained models in language and image understanding, we introduce deep generative models to produce motor control signals stimulating legged robots to act like real animals. Then, we shape various traversing capabilities at a higher level to align with the environment by reusing the primitive module. Finally, a strategic module is trained focusing on complex downstream tasks by reusing the knowledge from previous levels. We apply the trained hierarchical controllers to the MAX robot, a quadrupedal robot developed in-house, to mimic animals, traverse complex obstacles and play in a designed challenging multi-agent chase tag game, where lifelike agility and strategy emerge in the robots.
format Preprint
id arxiv_https___arxiv_org_abs_2308_15143
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Lifelike Agility and Play in Quadrupedal Robots using Reinforcement Learning and Generative Pre-trained Models
Han, Lei
Zhu, Qingxu
Sheng, Jiapeng
Zhang, Chong
Li, Tingguang
Zhang, Yizheng
Zhang, He
Liu, Yuzhen
Zhou, Cheng
Zhao, Rui
Li, Jie
Zhang, Yufeng
Wang, Rui
Chi, Wanchao
Li, Xiong
Zhu, Yonghui
Xiang, Lingzhu
Teng, Xiao
Zhang, Zhengyou
Robotics
Artificial Intelligence
Knowledge from animals and humans inspires robotic innovations. Numerous efforts have been made to achieve agile locomotion in quadrupedal robots through classical controllers or reinforcement learning approaches. These methods usually rely on physical models or handcrafted rewards to accurately describe the specific system, rather than on a generalized understanding like animals do. Here we propose a hierarchical framework to construct primitive-, environmental- and strategic-level knowledge that are all pre-trainable, reusable and enrichable for legged robots. The primitive module summarizes knowledge from animal motion data, where, inspired by large pre-trained models in language and image understanding, we introduce deep generative models to produce motor control signals stimulating legged robots to act like real animals. Then, we shape various traversing capabilities at a higher level to align with the environment by reusing the primitive module. Finally, a strategic module is trained focusing on complex downstream tasks by reusing the knowledge from previous levels. We apply the trained hierarchical controllers to the MAX robot, a quadrupedal robot developed in-house, to mimic animals, traverse complex obstacles and play in a designed challenging multi-agent chase tag game, where lifelike agility and strategy emerge in the robots.
title Lifelike Agility and Play in Quadrupedal Robots using Reinforcement Learning and Generative Pre-trained Models
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2308.15143