arXiv cs.AI· Yusuke Takagi, Motonari Kambara, Daichi Yashima, Koki Seno, Kento Tokura, Komei Sugiura·· 3 天前AI 评分29
AnolisVLA：面向移动操作的轻量级视觉-语言-动作模型，采用深度状态空间模型
AnolisVLA: Lightweight Vision-Language-Action Model with Deep State Space Models for Mobile Manipulation
AI 导读
AnolisVLA 是一款采用深度状态空间模型处理多模态序列的轻量级视觉-语言-动作模型，用于语言引导的机器人操作。在同等参数规模下，其任务成功率、推理速度和 GPU 显存占用均优于 SmolVLA；在移动操作中，任务成功率比 30 亿参数的 π₀.₅ 高出 21 个百分点，推理速度约为其 3 倍。
正文 · 原文
Abstract:In this study, we address the problem of language-guided robotic manipulation, where a robot is required to manipulate a wide range of objects based on visual observations and natural language instructions. This task is essential for service robots that operate in human environments, and requires safety, efficiency, and task-level generality. Although Vision-Language-Action models (VLAs) have demonstrated strong performance for this task, their deployment in resource-constrained environments remains challenging because of the computational cost of standard transformer backbones. To overcome this limitation, we propose AnolisVLA, a lightweight VLA that uses a deep state space model to process multimodal sequences efficiently. The model uses lightweight, fast sequential state modeling to process visual and textual inputs. This design enables efficient trajectory generation. We evaluated the proposed method in simulation and on two real robot platforms. Compared with SmolVLA at a comparable parameter scale, AnolisVLA achieved higher task success rates, faster inference, and comparable GPU memory usage. In mobile manipulation, it outperformed the larger, 3-billion-parameter $\pi_{0.5}$ model by 21 percentage points in task success rate while achieving approximately three times faster inference.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2603.15046 [cs.RO] |
| (or arXiv:2603.15046v2 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2603.15046 arXiv-issued DOI via DataCite |
Submission history
From: Yusuke Takagi [view email]
[v1]
Mon, 16 Mar 2026 09:57:45 UTC (13,269 KB)
[v2]
Mon, 12 Oct 2026 14:11:09 UTC (19,049 KB)
来源：arXiv cs.AI · arxiv.org