NVIDIA's USDCraft uses LLM-written programs to build simulation-ready articulated 3D assets
nvidia · hf · 2026-10-09
- NVIDIA proposes USDCraft, framing articulated 3D asset reconstruction for real-to-sim robot manipulation as programmatic modeling grounded in partial geometry: a pretrained LLM writes and revises executable programs without task-specific training.
- Key techniques: source geometry analysis (metric textual mesh descriptions separating observed vs unknown space) and iterative geometric rechecking that maps discrepancies to program edits, plus visual feedback and physical authoring guidance.
- Output: articulated USD assets with explicit physical properties that load into Isaac Sim with no manual adjustment; state-of-the-art articulation recovery on two benchmarks and validated real-to-sim-to-real transfer.
More from Embodied
- Scoble rides Tesla Cybercab: exceeds expectations, plans to buy one to rent out — Scobleizer · 2026-10-09
- Egocentric data collection is all the rage now, says researcher citing an ahead-of-its-time project — gragtah · 2026-10-09
- DreamTrue robot world model cuts interaction defect rate from 48.12% to 6.25%, tops AgiBot challenge — Junyan Li · 2026-10-09
- SuperNav: ZJU's Agent Harness Lets MLLMs Navigate Any Scene Without Fine-tuning — zju3dv · 2026-10-09
- SimpleICL Defines Robot In-Context Learning with a Low-Cost Open Recipe — Minxing Li · 2026-10-09
- UBTECH inks FAW-Volkswagen deal for factory humanoids, targets 10,000 robots a year — CyberRobooo · 2026-10-09