---
格式版本: 2
标题: "NebulaVLA: A Dual-Frequency Vision-Language-Action Model With Guide Action for Robotic Manipulation"
原文链接: "https://arxiv.org/abs/2608.16503"
发布日期: "2026-08-17"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:scrape:strict_html_body"
发布时间证据: "citation_date: 2026/08/17"
发布时间校准原因: "规则确认唯一严格发布时间，来源 scrape:strict_html_body"
发布时间校准置信度: "high"
发布时间候选数量: 6
发布时间严格候选数量: 6
发布时间原页读取状态: ""
发布时间未找到原因: ""
发布时间校准时间: "2026-08-18T19:33:01+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-18T19:32:55+08:00"
入库时间: "2026-08-18T11:33:34.104Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=deployment&searchtype=all"
匹配关键词:
  - "deployment"
  - "performance"
  - "AI"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 0
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该arXiv论文为机器人操作VLA模型，与超节点/AI Rack/机柜级AI基础设施等主题完全无关，无任何相关技术或商业信息。"
AI质检模型: "tx-deepseek-v4-flash"
AI质检时间: "2026-08-18T19:33:56+08:00"
AI主题相关性: 0
AI来源权威性: 0
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
AI摘要: "论文提出NebulaVLA，一种异步双频视觉-语言-动作模型，通过解耦高层语义推理与低层动作控制提升计算效率与模块性，并引入GESTURE-7统一动作表示及Guide Action算法保证运动连续性。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:25:15.725Z"
采集批次: "2026年8月18日15点52分05秒"
采集批次ID: "20260818-155205-090"
去重键: "https://arxiv.org/abs/2608.16503"
---

## Computer Science > Robotics

## Title:NebulaVLA: A Dual-Frequency Vision-Language-Action Model With Guide Action for Robotic Manipulation

Authors:[Cong Zhao](https://arxiv.org/search/cs?searchtype=author&query=Zhao,+C), [Shuai Tian](https://arxiv.org/search/cs?searchtype=author&query=Tian,+S), [Xu Zhang](https://arxiv.org/search/cs?searchtype=author&query=Zhang,+X), [Baocheng Ni](https://arxiv.org/search/cs?searchtype=author&query=Ni,+B), [Xinguo Song](https://arxiv.org/search/cs?searchtype=author&query=Song,+X), [Xueying Sun](https://arxiv.org/search/cs?searchtype=author&query=Sun,+X), [Shu Jiang](https://arxiv.org/search/cs?searchtype=author&query=Jiang,+S), [Shouchang Yang](https://arxiv.org/search/cs?searchtype=author&query=Yang,+S), [Bo Tang](https://arxiv.org/search/cs?searchtype=author&query=Tang,+B), [Jin Deng](https://arxiv.org/search/cs?searchtype=author&query=Deng,+J), [Ge Zhu](https://arxiv.org/search/cs?searchtype=author&query=Zhu,+G), [YongCheng Wang](https://arxiv.org/search/cs?searchtype=author&query=Wang,+Y), [Jin Xu](https://arxiv.org/search/cs?searchtype=author&query=Xu,+J), [Ri Yang](https://arxiv.org/search/cs?searchtype=author&query=Yang,+R)

[View PDF](https://arxiv.org/pdf/2608.16503) [HTML (experimental)](https://arxiv.org/html/2608.16503v1)

> Abstract:Real-world deployment of Vision-Language-Action (VLA) models is often bottlenecked by efficiency-performance trade-offs, cross-embodiment generalization, and execution smoothness. We present NebulaVLA, an asynchronous dual-frequency architecture that decouples high-level semantic reasoning from low-level action control, optimizing computational resources and modularity. To bridge semantic gaps across heterogeneous robots, we introduce GESTURE-7, a unified language-grounded action representation. Furthermore, our Guide Action algorithm enforces kinematic continuity via mask-based smoothness constraints. Comprehensive evaluations demonstrate that NebulaVLA significantly outperforms synchronous baselines, achieving an 85.5\\% average success rate on LIBERO-Plus and accelerating action generation by \\textasciitilde 2.7$\\times$. This asynchronous design enables highly efficient and responsive control for practical robotics.

| Comments: |  |
| --- | --- |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| ACM classes: | I.2.9; I.2.10 |
| Cite as: | [arXiv:2608.16503](https://arxiv.org/abs/2608.16503) \[cs.RO\] |
|  | (or [arXiv:2608.16503v1](https://arxiv.org/abs/2608.16503v1) \[cs.RO\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.16503](https://doi.org/10.48550/arXiv.2608.16503) |

## Submission history

From: Cong Zhao \[[view email](https://arxiv.org/show-email/a71af1f0/2608.16503)\]  
**\[v1\]** Mon, 17 Aug 2026 12:40:23 UTC (628 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.16503) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))