arXiv cs.AI· Houlong Xiong, Zhenqi Qiu, Zechen Wang, Suohang Zhang, Yiyu Ren, Wanting Xu, Hongfei Niu, Chengyang He, Ge Sun, Ran Cheng, Qian Zhu·· 9 天前AI 评分33
REACT：面向 VLA 模型机器人控制的滚动去噪与双重解耦框架
REACT: Rolling Denoising and Dual Decoupling for Reactive Robot Control with VLA Models
AI 导读
REACT 是一个滚动去噪框架，让基于 flow 的 VLA 模型在保留长时程上下文的同时提升反应速度：它维护带错开 flow 时间步的持久动作缓冲区，每步用最新观测对完整时域去噪、执行最干净的动作块、将部分精炼的未来块前移并补充新噪声。
正文 · 原文
Authors:Houlong Xiong, Zhenqi Qiu, Zechen Wang, Suohang Zhang, Yiyu Ren, Wanting Xu, Hongfei Niu, Chengyang He, Ge Sun, Ran Cheng, Qian Zhu
Abstract:Flow-based vision-language-action (VLA) models generate action chunks for temporally coherent robot motion, but chunked control creates a fundamental closed-loop trade-off: long chunks provide smooth execution, whereas frequent replanning improves reactivity at the cost of action discontinuities. We introduce REACT, a rolling-denoising framework that makes flow-based VLAs more reactive while preserving long-horizon context. Instead of regenerating entire action chunks from scratch, REACT maintains a persistent action buffer with staggered flow timesteps. At each control step, the full horizon is denoised using the latest observation, the cleanest action block is executed, partially refined future blocks are shifted forward, and fresh noise is appended to the tail. As a result, each executed action block is refined across multiple recent observations before deployment. To support real-time control, we further introduce dual decoupling, which separates sensing, VLM encoding, DiT denoising, and action execution, enabling high-frequency observation updates and action streaming under practical compute constraints. Across the RoboTwin 2.0 simulation benchmark and real-world tasks spanning bimanual manipulation and dynamic control on multiple robot platforms, REACT improves task success and reduces reaction latency while producing smoother trajectories than frequent-replanning and asynchronous baselines.
| Comments: | Accepted to CoRL 2026 as Spotlight. Project page: this https URL |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.12007 [cs.RO] |
| (or arXiv:2610.12007v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.12007 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Houlong Xiong [view email]
[v1]
Thu, 8 Oct 2026 14:09:46 UTC (16,529 KB)
来源：arXiv cs.AI · arxiv.org