Reproduce “VFS: Sim-to-Real Virtual Force Sensing of Single and Simultaneous Contacts via a Frozen Torque-to-Force Prior” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
John U. Onyemelukwe, Zhiyang Dou, Ethan Phan, Yuxiang Ma, Wojciech Matusik
VFS estimates contact forces at specified candidate sites by learning a torque-to-force model in simulation and adapting real-robot motor telemetry to its torque inputs. On the OpenManipulator-X end-effector benchmark, it achieves 0.457 ± 0.007 N mean vector force error and 0.77° ± 0.06° replay error over 100 steps. It assumes candidate sites and site flags, not general contact localization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.