arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2404.02887cs.RO
收藏

通过可微分仿真学习可部署的运动控制

Learning Deployable Locomotion Control via Differentiable Simulation

ETH Zurich(苏黎世联邦理工学院);NVIDIA(NVIDIA公司);RAI Institute(RAI研究所);Georgia Institute of Technology(佐治亚理工学院);Comillas Pontifical University(康普顿斯大学)

Clemens Schwarke, Victor Klemm, Joshua Bagajo, Jean-Pierre Sleiman, Ignat Georgiev, Jesus Tordesillas, Marco Hutter

AI总结:

本文提出一种兼顾物理保真度与信息梯度的可微分接触模型,首次实现完全在可微分仿真器中训练的四足运动策略零样本迁移到真实机器人,验证了可微分仿真用于现实运动控制的可行性。

摘要:

可微分仿真器通过提供系统动力学的解析梯度,有望提高机器人学习的样本效率。然而,它们在运动等接触密集型任务中的应用因接触固有的非光滑特性而变得复杂,阻碍了有效的基于梯度的优化。因此,现有工作通常依赖软接触模型,这些模型提供光滑梯度但缺乏物理准确性,将结果限制在仿真中。为了解决这一局限,我们提出了一种可微分接触模型,旨在提供信息丰富的梯度,同时保持高物理保真度。我们通过在我们的可微分仿真器中利用解析梯度训练四足运动策略,并将学习到的策略零样本成功迁移到现实世界,证明了我们方法的有效性。据我们所知,这代表了完全在可微分仿真器中学习的腿式运动策略首次成功实现从仿真到现实的迁移,确立了使用可微分仿真进行现实世界运动控制的可行性。

英文摘要:

Differentiable simulators promise to improve sample efficiency in robot learning by providing analytic gradients of the system dynamics. Yet, their application to contact-rich tasks like locomotion is complicated by the inherently non-smooth nature of contact, impeding effective gradient-based optimization. Existing works thus often rely on soft contact models that provide smooth gradients but lack physical accuracy, constraining results to simulation. To address this limitation, we propose a differentiable contact model designed to provide informative gradients while maintaining high physical fidelity. We demonstrate the efficacy of our approach by training a quadrupedal locomotion policy within our differentiable simulator leveraging analytic gradients and successfully transferring the learned policy zero-shot to the real world. To the best of our knowledge, this represents the first successful sim-to-real transfer of a legged locomotion policy learned entirely within a differentiable simulator, establishing the feasibility of using differentiable simulation for real-world locomotion control.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑