# LIBERO-VPro基准发布：量化评测机器人基础模型闭环视觉鲁棒性

> 原文链接：https://robotics.hellobike.com/papers/2026-09-21-2609-24350
> 频道：papers | 更新日期：2026-09-21 | 标签：auto, arxiv
> 核心摘要：针对机器人基础模型视觉鲁棒性评估缺失，LIBERO-VPro构建含4维度12类挑战的基准。经19.6万次仿真及200次Franka Research 3真机验证，揭示模型在交互线索中断或空间先验违背时性能骤降，为VLA与WAM架构提供差异化诊断框架。

# LIBERO-VPro基准发布：量化评测机器人基础模型闭环视觉鲁棒性

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[LIBERO-VPro: Benchmarking Closed-Loop Visual Robustness of Robotic Foundation Models](https://arxiv.org/abs/2609.24350)（arXiv 2609.24350）

## 核心内容

针对现有操控基准假设视觉观测完美、难以反映真实部署扰动的问题，研究团队推出LIBERO-VPro基准，系统评估机器人基础模型的闭环视觉鲁棒性。该基准覆盖视觉证据退化、相机数据陈旧、视觉源一致性及任务相关场景变化四大维度，包含12个挑战类别、96种实验设置及3296个任务条件案例。团队对3个视觉-语言-动作（VLA）模型和3个世界-动作（WAM）模型进行了约19.6万次仿真回合测试，并辅以200次Franka Research 3真机实测，填补了动态视觉干扰下模型适应性评估的空白。

评测结果显示，高标称性能掩盖了模型在视觉接地与适应方面的显著弱点。尽管模型对严重物体遮挡具有耐受性，但在局部交互线索被破坏或熟悉空间先验被违背时性能急剧下降；同时对陈旧或缺失观测高度敏感，且在任务前置条件变更需行为适应时表现挣扎。VLA与WAM模型展现出截然不同的鲁棒性特征，表明视觉鲁棒性具有多维度和架构依赖性。该基准为开发能在复杂视觉条件下可靠接地与自适应行动的机器人基础模型提供了系统化诊断工具。

## 关键看点

- **基准规模**：涵盖4大维度12类挑战及96种设置，包含3296个任务条件案例，支撑大规模鲁棒性压力测试
- **验证体系**：完成约19.6万次仿真回合与200次Franka Research 3真机实测，确保评估结果具备Sim-to-Real参考价值
- **失效模式**：模型耐受物体级遮挡，但局部交互线索中断或空间先验违背导致性能骤降，对陈旧观测极度敏感
- **架构差异**：VLA与WAM模型呈现不同鲁棒性画像，证实视觉鲁棒性并非单一指标，而是强依赖于底层模型架构

## 工程与产业观察

该基准揭示了具身大模型从“实验室高分”迈向“工业级可靠”的关键瓶颈：视觉接地能力远比标称成功率重要。对于算法工程师，LIBERO-VPro提供了超越标准Benchmark的压力测试集，有助于定位模型在非结构化环境中的长尾失效点；对于产业决策者，它强调了在部署前必须进行多维度视觉鲁棒性验证，而非仅关注理想条件下的任务完成率，是评估具身智能落地成熟度的重要标尺。

## 英文摘要 (Abstract)

Robotic foundation models achieve impressive performance on standard manipulation benchmarks, yet these evaluations typically assume clean, timely, and consistent visual observations throughout execution. We introduce LIBERO-VPro, a benchmark for systematically evaluating the closed-loop visual robustness of robotic foundation models by perturbing the visual evidence available during execution. LIBERO-VPro covers four complementary dimensions, including Visual Evidence Degradation, Camera Staleness, Visual Source Consistency, and Task-Relevant Scene Variation, spanning 12 challenge categories, 96 experimental settings, and 3,296 task-condition cases. We evaluate three vision-language-action models and three world-action models over approximately 196,000 simulated episodes, complemented by 200 real-world rollouts on a Franka Research 3. Our results reveal that strong nominal performance can mask substantial weaknesses in visual grounding and adaptation. Models often remain successful despite severe object-level occlusion, yet degrade sharply when local interaction cues are disrupted or familiar spatial priors are violated. They are also highly sensitive to stale or missing observations and struggle when changed task preconditions require behavioral adaptation. Finally, VLAs and WAMs exhibit distinct robustness profiles, showing that visual robustness is multi-dimensional and architecture-dependent. LIBERO-VPro provides a systematic diagnostic framework for developing robotic foundation models that can more reliably ground and adapt their actions under challenging visual conditions.

作者：Huiqiong Li, Zhiting Mei, Anirudha Majumdar, Jingjing Chen, Yu-Gang Jiang, Bin Zhu
:::

## 原始参考来源 (Sources)

- [arXiv 2609.24350：LIBERO-VPro: Benchmarking Closed-Loop Visual Robustness of Robotic Foundation Models](https://arxiv.org/abs/2609.24350)