Metraipolis › archive › dc_2abfaee1d619

Blackout vs. Freeze: Analyzing Physical Failure Modes of VLAs under Camera Faults — arXiv:2609.39145

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.39145
License (as archived)
CC0-1.0
Archived
2026-10-07T12:44:59.819Z
SHA-256 of the archived text
272908bb18ea1e42a485d3182b5f222df7c3fcae0c30ff9be0f75700e8790363
Stable URL of this page
https://metraipolis.com/documento/dc_2abfaee1d619

How to cite

Blackout vs. Freeze: Analyzing Physical Failure Modes of VLAs under Camera Faults — arXiv:2609.39145 — arXiv (en), https://arxiv.org/abs/2609.39145 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2abfaee1d619 (Metraipolis archive dc_2abfaee1d619, sha256 272908bb18ea1e42a485d3182b5f222df7c3fcae0c30ff9be0f75700e8790363)
BibTeX
@misc{dc2abfaee1d619,
  title = {Blackout vs. Freeze: Analyzing Physical Failure Modes of VLAs under Camera Faults — arXiv:2609.39145},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.39145},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2abfaee1d619; sha256 272908bb18ea1e42a485d3182b5f222df7c3fcae0c30ff9be0f75700e8790363},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39145. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Unreliable visual inputs can harm task performance and cause potential physical safety risks for vision-language-action (VLA) models. We analyze how $\pi 0.5$ and GR00T models act under input faults such as image blackouts and freezing. We find that blackout and freezing produce distinct physical failure modes even when task-success rates are similarly low: freezing causes more extreme joint behavior, whereas blackout after gripper closure can cause more object drops, most markedly without proprioception. Selective intervention studies reveal that proprioception (current robot state) partly compensates for the removed robot depictions and reduces non-target contact. However, it cannot sufficiently restore task success when wrist-view object information is removed, even when aided by the remaining scene view. We then evaluate two mitigation approaches: camera-blackout training and training-free replacement of faulty visual embeddings. Both improve task success in selected conditions, but can increase unintended contact or disturbance to surrounding objects. Real-robot trials further show that successful execution under camera faults can still involve unintended physical interactions. These findings motivate designing VLA policies that use the robot and object information still available under camera faults to limit hazardous motion.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.AI).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.39145 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T12:14:14.029Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).