arXiv：cs.AI· Siddhant Gangapurwala, Luigi Campanaro, Ioannis Havoutis·· 4 小时前AI 评分37
低至 8 Hz 的低频运动控制实现 ANYmal C 四足机器人鲁棒动态行走
Learning Low-Frequency Motion Control for Robust and Dynamic Robot Locomotion
AI 导读
研究团队用低至 8 Hz 的学习型运动控制器，让真实 ANYmal C 四足机器人实现 1.5 m/s 高航向速度、穿越崎岖地形并抵抗外部扰动。对比 5 Hz 至 200 Hz 训练的深度强化学习策略后发现，低频策略对执行延迟和系统动力学变化更不敏感，甚至无需动力学随机化或执行器建模即可完成 sim-to-real 迁移。训练与部署代码已公开。
正文
Abstract:Robotic locomotion is often approached with the goal of maximizing robustness and reactivity by increasing motion control frequency. We challenge this intuitive notion by demonstrating robust and dynamic locomotion with a learned motion controller executing at as low as 8 Hz on a real ANYmal C quadruped. The robot is able to robustly and repeatably achieve a high heading velocity of 1.5 m/s, traverse uneven terrain, and resist unexpected external perturbations. We further present a comparative analysis of deep reinforcement learning (RL) based motion control policies trained and executed at frequencies ranging from 5 Hz to 200 Hz. We show that low-frequency policies are less sensitive to actuation latencies and variations in system dynamics. This is to the extent that a successful sim-to-real transfer can be performed even without any dynamics randomization or actuation modeling. We support this claim through a set of rigorous empirical evaluations. Moreover, to assist reproducibility, we provide the training and deployment code along with an extended analysis at this https URL.
| Comments: | 7 pages, 9 figures and 2 tables |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2209.14887 [cs.RO] |
| (or arXiv:2209.14887v3 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2209.14887 arXiv-issued DOI via DataCite |
|
| Journal reference: | IEEE International Conference on Robotics and Automation (ICRA) 2023 |
Submission history
From: Siddhant Gangapurwala [view email]
[v1]
Thu, 29 Sep 2022 15:55:33 UTC (5,013 KB)
[v2]
Tue, 21 Feb 2023 13:30:56 UTC (4,870 KB)
[v3]
Fri, 2 Oct 2026 17:32:38 UTC (4,870 KB)
来源：arXiv：cs.AI · arxiv.org