analiza aiqbench
SimWeaver advances zero-shot visual-language-action policies for deformable object manipulation by leveraging a measurement-backed simulator and single-image asset generation. This approach achieves high real-world success rates across diverse tasks like plastic-bag handling without requiring teleoperation or per-task calibration, significantly reducing training costs compared to data-driven baselines.
SimWeaver: Zero-Shot RGB Sim-to-Real for Deformable Manipulation
arXiv:2606.15338v1 Announce Type: new Abstract: RGB sim-to-real for deformable manipulation has remained largely unsolved without real-world fine-tuning. We present SimWeaver, which trains zero-shot RGB VLA policies on 200 simulated demonstrations per task, reaching above 80% per-task and 91% average real-world success across 5 diverse deformable tasks including plastic-bag manipulation, without teleoperation or per-task calibration. SimWeaver combines a reliable measurement-backed simulator (SimWeaver-Sim) with an extensible asset framework supporting single-image generation(SimWeaver-Asset), a deterministic topology-aware trajectory synthesizer (SimWeaver-Syn), and a sim-to-real protocol with ISP-aware photometric augmentation (SimWeaver-Real). On silk grasping, the sim-trained policy reaches 100% under visual distribution shifts where real-data baselines drop to 9-70%, at two orders of magnitude lower per-trajectory cost. We will release SimWeaver and a representative asset subset. Project page: https://simweaver.github.io/
Abstract:RGB sim-to-real for deformable manipulation has remained largely unsolved without real-world fine-tuning. We present SimWeaver, which trains zero-shot RGB VLA policies on 200 simulated demonstrations per task, reaching above 80% per-task and 91% average real-world success across 5 diverse deformable tasks including plastic-bag manipulation, without teleoperation or per-task calibration. SimWeaver combines a reliable measurement-backed simulator (SimWeaver-Sim) with an extensible asset framework supporting single-image generation(SimWeaver-Asset), a deterministic topology-aware trajectory synthesizer (SimWeaver-Syn), and a sim-to-real protocol with ISP-aware photometric augmentation (SimWeaver-Real). On silk grasping, the sim-trained policy reaches 100% under visual distribution shifts where real-data baselines drop to 9-70%, at two orders of magnitude lower per-trajectory cost. We will release SimWeaver and a representative asset subset. Project page: this https URL
Robotics (cs.RO)
arXiv:2606.15338 [cs.RO]
(or arXiv:2606.15338v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.15338 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 136 cuvinte
Subiecte conexe
SoftVTBench: A Safety-Aware Visuo-Tactile Benchmark for Physically Constrained Robotic Manipulation of Deformable Objects (Early Version)
arXiv · 21 aug. 2026
FolDeX: A Physical-World Benchmark for Long-Horizon Robotic Manipulation of Deformable Objects
arXiv · 10 sept. 2026
DefFlow: Goal-Conditioned Deformable Object Manipulation using 3D Point Flows
arXiv · 19 oct. 2026
The Language of Deformation: Semantic Failure Analysis as a Prerequisite for Dexterous Manipulation of Deformable Objects
arXiv · 27 oct. 2026