arXiv cs.AI· Jonathan C. Kao, Jason Chan, Andy Wang·· 13 小时前AI 评分38
Flow Control：用键盘等简单实时输入操控 VLA 模型
Flow Control: Steering Vision-Language-Action Models with Simple Real-Time Inputs
AI 导读
研究者提出 Flow Control，可将键盘等通用输入写入 VLA 模型 flow matching 动作专家的初始条件，无需重训练或微调即可实时操控机器人动作。该方法对次优输入具有鲁棒性，能显著提升任务成功率并加快完成速度，用其轨迹微调还能改进自主策略。论文已被 CoRL 2026 接收。
正文 · 原文
Abstract:We introduce flow control of vision-language-action (VLA) models, a simple and effective way to steer VLA actions in real time through generic inputs, such as a keyboard. This method can be used out-of-the-box and does not require retraining or fine-tuning VLAs. It enables relatively crude user inputs to steer a VLA to align with user intent. Flow control writes the user input into the initial condition of the VLA's flow matching action expert. Because the flow is deterministic, integration retains information about the user input while moving it toward the action distribution learned during training, so that the generated actions reflect the user's intent (high fidelity) and stay close to the VLA's own unsteered actions (high conformity). We demonstrate that flow control has many desirable properties: (1) flow control accurately and responsively steers robot actions with user inputs, (2) it is robust to suboptimal user inputs, (3) it enables users to steer VLAs to achieve significantly higher success rates and faster task completion, and (4) fine-tuning a VLA on flow control trajectories improves the autonomous policy. Together, these results provide a simple and intuitive way for users to help steer VLA actions, increasing task performance. Project page and videos: this https URL.
| Comments: | Accepted to the 10th Conference on Robot Learning (CoRL 2026). 29 pages. Project page: this https URL |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC) |
| Cite as: | arXiv:2606.10180 [cs.RO] |
| (or arXiv:2606.10180v2 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2606.10180 arXiv-issued DOI via DataCite |
Submission history
From: Jonathan Kao [view email]
[v1]
Mon, 8 Jun 2026 21:16:37 UTC (16,711 KB)
[v2]
Mon, 19 Oct 2026 20:37:04 UTC (5,322 KB)
来源：arXiv cs.AI · arxiv.org