Robust Quadrupedal Locomotion via Risk-Averse Policy Learning

Jiyuan Shi, Chenjia Bai, Haoran He, Lei Han, Dong Wang, Bin Zhao, Mingguo Zhao, Xiu Li, Xuelong Li

科研成果: 书/报告/会议事项章节会议稿件同行评审

6 引用 (Scopus)

摘要

The robustness of legged locomotion is crucial for quadrupedal robots in challenging terrains. Recently, Reinforcement Learning (RL) has shown promising results in legged locomotion and various methods try to integrate privileged distillation, scene modeling, and external sensors to improve the generalization and robustness of locomotion policies. However, these methods are hard to handle uncertain scenarios such as abrupt terrain changes or unexpected external forces. In this paper, we consider a novel risk-sensitive perspective to enhance the robustness of legged locomotion. Specifically, we employ a distributional value function learned by quantile regression to model the aleatoric uncertainty of environments, and perform risk-averse policy learning by optimizing the worst-case scenarios via a risk distortion measure. Extensive experiments in both simulation environments and a real Aliengo robot demonstrate that our method is efficient in handling various external disturbances, and the resulting policy exhibits improved robustness in harsh and uncertain situations in legged locomotion.

源语言英语
主期刊名2024 IEEE International Conference on Robotics and Automation, ICRA 2024
出版商Institute of Electrical and Electronics Engineers Inc.
11459-11466
页数8
ISBN(电子版)9798350384574
DOI
出版状态已出版 - 2024
活动2024 IEEE International Conference on Robotics and Automation, ICRA 2024 - Yokohama, 日本
期限: 13 5月 202417 5月 2024

出版系列

姓名Proceedings - IEEE International Conference on Robotics and Automation
ISSN(印刷版)1050-4729

会议

会议2024 IEEE International Conference on Robotics and Automation, ICRA 2024
国家/地区日本
Yokohama
时期13/05/2417/05/24

指纹

探究 'Robust Quadrupedal Locomotion via Risk-Averse Policy Learning' 的科研主题。它们共同构成独一无二的指纹。

引用此