arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.02481cs.ROcs.AI
收藏

用于机器人控制的流匹配策略梯度

Flow Policy Gradients for Robot Control

Amazon FAR(亚马逊Far);UC Berkeley(加州大学伯克利分校);Stanford(斯坦福大学);HKU(香港大学);CMU(卡内基梅隆大学)

Brent Yi, Hongsuk Choi, Himanshu Gaurav Singh, Xiaoyu Huang, Takara E. Truong, Carmelo Sferrazza, Yi Ma, Rocky Duan, Pieter Abbeel, Guanya Shi, Karen Liu, Angjoo Kanazawa

AI总结:

本文提出了一种基于流匹配的策略梯度方法,用于训练更复杂的机器人控制策略,实现了在四肢运动、人形运动跟踪和操作任务中的成功,并展示了在仿真到现实迁移中的鲁棒性。

摘要:

基于似然的策略梯度方法是训练机器人控制策略的主要方法,这些方法依赖于可微分的动作似然,这限制了策略输出到简单的分布,如高斯分布。在本工作中,我们展示了如何使流匹配策略梯度——一种最近的框架,可以绕过似然计算——在训练和微调更复杂的策略时变得有效,适用于具有挑战性的机器人控制设置。我们引入了一个改进的目标,使其在四肢运动、人形运动跟踪和操作任务中取得成功,并在两个人形机器人上实现了鲁棒的仿真到现实的迁移。然后我们展示了训练动态的消融分析。结果表明,策略在从头开始训练时如何利用流表示进行探索,以及在微调时相比基线有改进的鲁棒性。

英文摘要:

Likelihood-based policy gradient methods are the dominant approach for training robot control policies from rewards. These methods rely on differentiable action likelihoods, which constrain policy outputs to simple distributions like Gaussians. In this work, we show how flow matching policy gradients, a recent framework that bypasses likelihood computation, can be made effective for training and finetuning more expressive policies in challenging robot control settings. We introduce an improved objective that enables success in legged locomotion, humanoid motion tracking, and manipulation tasks, as well as robust sim-to-real transfer on two humanoid robots. We then present ablations and analysis on training dynamics. Results show how policies can exploit the flow representation for exploration when training from scratch, as well as improved finetuning robustness over baselines.

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。

↑