Researchers have introduced OSMA-Bench, a new benchmark designed to evaluate how well household robots can adapt their navigation behavior based on the type and proximity of obstacles. The benchmark utilizes spatially matched scenarios that vary only in obstacle type to test for semantic safety distinctions. Findings indicate that current vision-language models (VLMs) and vision-language-action (VLA) policies often fail to translate their understanding of obstacle safety into appropriate motion adjustments, despite correctly identifying the obstacles. AI
IMPACT Highlights a critical gap in translating semantic understanding to safe, adaptive robot motion, impacting future household robotics development.
RANK_REASON The cluster contains a research paper introducing a new benchmark and metric for evaluating robot navigation safety. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- Gotit.pub
- Hugging Face
- OSMA-Bench
- ScienceCast
- Semantic Sensitivity Index
- Vision-language-action model
- vision-language model
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
