arXiv cs.AI· Andrea Yaoyun Cui, Hanyu Wang, Xinyi Wang, Bangzheng Li, Lin Zhang, Zhenggang Tang, Shyamal Buch, Dejia Xu, Matthew Tancik·· 3 天前AI 评分34
LIBERO-RIG：面向视觉-语言-动作模型的硬件配置鲁棒性基准
LIBERO-RIG: A Hardware-Setup Robustness Benchmark for Vision-Language-Action Models
AI 导读
LIBERO-RIG 是一个只扰动机器人硬件与配置、保持 LIBERO 任务不变的鲁棒性基准，覆盖相机安装与光学、相机-机械臂标定、驱动增益、控制频率、本体感觉噪声、底座位姿、桌面高度和夹爪垫等 11 个扰动轴，每轴 5 个严重度，在四个 LIBERO 套件上共 210 种条件。
正文 · 原文
Abstract:Robot policies are expected to generalize across deployment environments, including variations in the robot itself: cameras are bumped and remounted, calibrations drift, controllers run at different rates, and end-effector pads wear down. We present LIBERO-RIG, a benchmark that perturbs only the robot's hardware and setup while holding every LIBERO task fixed. The benchmark spans eleven perturbation axes, including camera mounting and optics, camera-to-arm calibration, actuation gain, control rate, proprioceptive noise, base pose, table height, and gripper pads. Each axis has five severity levels, yielding 210 conditions across the four LIBERO suites. We evaluate nine recent vision-language-action models using 500 episodes per condition and substantiate each finding with rigorous statistical tests. Pooled success falls from 0.932 to 0.767 at the middle severity and 0.509 at the highest severity. Performance degradation varies substantially across perturbation axes and action interfaces, and training on visual perturbations does not mitigate faults that leave camera images unchanged. We release the benchmark together with manifests specifying the fixed perturbation parameters for each episode.
| Comments: | 30 pages (4 pages main text plus appendix), 15 figures, 18 tables |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.18843 [cs.RO] |
| (or arXiv:2610.18843v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.18843 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Andrea Yaoyun Cui [view email]
[v1]
Tue, 13 Oct 2026 17:57:18 UTC (6,919 KB)
来源：arXiv cs.AI · arxiv.org