Abstract:Modern robotic policies increasingly rely on action chunking to execute complex tasks in the physical world. While action chunking improves temporal consistency at moderate action frequencies, it becomes insufficient when the action frequency is further increased (e.g., to 60~Hz). At such high frequencies, policies often fail to generate actions that are both temporally smooth and spatially consistent. We address this challenge by shifting high-frequency action learning from the action space to a latent space with variational autoencoder (VAE). This formulation significantly improves both temporal and spatial consistency of high-frequency control. To enable smooth real-time execution, we further introduce Reuse-then-Refine, a chunk-level refine strategy that improves continuity between adjacent action chunks under asynchronous inference. As a result, robots controlled by our policy can execute complex contact-rich tasks continuously, with less pauses and jerky motions. Experiments on three real-world contact-rich robotic tasks show that our approach consistently completes tasks with smooth motions. Our code and data are available at this https URL.
| Comments: | 17 pages, 10 figures |
| Subjects: | Robotics (cs.RO) |
| Cite as: | arXiv:2605.24931 [cs.RO] |
| (or arXiv:2605.24931v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2605.24931 arXiv-issued DOI via DataCite (pending registration) |
From: Kunyun Wang [view email]
[v1]
Sun, 24 May 2026 08:22:14 UTC (1,605 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。