# SimVLA实现纯仿真数据训练移动操作VLA零样本真机迁移

> 原文链接：https://robotics.hellobike.com/papers/2026-10-08-2610-11248
> 频道：papers | 更新日期：2026-10-08 | 标签：auto, arxiv
> 核心摘要：SimVLA框架仅用合成数据训练移动操作VLA，无需遥操数据即实现零样本真机迁移。在补货、倒水等任务中性能超越50条真实演示策略，验证了多模态仿真监督对具身模型泛化的有效性。

# SimVLA实现纯仿真数据训练移动操作VLA零样本真机迁移

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[SimVLA: Zero-Shot Sim-to-Real VLA Learning for Mobile Manipulation](https://arxiv.org/abs/2610.11248)（arXiv 2610.11248）

## 核心内容

针对移动操作VLA真机数据采集成本高的问题，研究提出SimVLA端到端框架，完全基于合成仿真数据训练且无需遥操作。该模型在补货、倾倒及清洁等任务上实现零样本真机迁移，包括真实家庭环境，其性能显著优于基于50条域内真实演示训练的策略，证实纯仿真路径在移动操作领域的可扩展性。

技术层面采用两阶段训练：预训练结合SimAction（35类任务原子技能组合）与SimVQA（利用仿真特权状态提供空间几何及子任务级视觉语言监督）；后训练混合SimAction与SimDeploy（策略 rollout 数据）。这种多重互补监督机制有效弥补了Sim-to-Real鸿沟，表明结构化仿真信号是替代昂贵真机标注的关键要素。

## 关键看点

- **纯仿真训练**：完全依赖合成数据训练移动操作VLA，无需任何真机遥操作数据即可实现零样本迁移至真实家庭环境
- **双重数据集**：构建SimAction（35种任务原子技能组合）与SimVQA（利用仿真特权状态提供空间几何及子任务级VL监督）
- **性能基准**：在补货、倒水等移动操作任务中，零样本SimVLA性能超越基于50条域内真实演示微调的策略
- **后训练策略**：混合SimAction与SimDeploy（策略rollout数据）进行后训练，强化模型在多样化仿真环境中的决策鲁棒性

## 工程与产业观察

SimVLA证明了通过原子技能组合与特权状态监督构建的高质量仿真数据，可替代小规模真机演示训练移动操作大模型。这为具身智能研发提供了低成本数据飞轮范式，提示工程团队应重视仿真器内部状态的结构化利用，而非单纯追求视觉逼真度，以加速VLA从Sim到Real的闭环迭代。

## 英文摘要 (Abstract)

Large-scale, diverse datasets have driven the success of LLMs and VLMs. But VLAs for robotics remain limited by the cost and complexity of real-world data collection. While simulation offers a scalable alternative, its potential for sim-to-real VLA learning in mobile manipulation remains largely underexplored. We introduce SimVLA, an end-to-end framework that trains VLAs entirely on synthetic simulation data without teleoperation for mobile manipulation. SimVLA is first pre-trained on two complementary simulation-derived datasets: SimAction, a large-scale robot action dataset spanning 35 diverse mobile manipulation tasks, generated by composing atomic skills, and SimVQA, which leverages privileged simulator state to provide spatial, geometric, and subtask-level visual-language supervision. We further post-train SimVLA on a mixture of SimAction and SimDeploy, a dataset collected from policy rollouts across diverse simulated environments. We evaluate SimVLA on tasks including restocking, pouring, and cleaning, and show zero-shot transfer to real-world mobile manipulation, including real home environments. SimVLA outperforms policies trained on 50 in-domain real-world demonstrations, suggesting that simulation can enable scalable sim-to-real mobile manipulation. We further demonstrate the value of multiple complementary forms of supervision for effectively leveraging simulation in VLA training.

作者：Kyoungin Baik, Youngwoon Lee
:::

## 原始参考来源 (Sources)

- [arXiv 2610.11248：SimVLA: Zero-Shot Sim-to-Real VLA Learning for Mobile Manipulation](https://arxiv.org/abs/2610.11248)