Reproduce “ImpedanceMimic: Learned Impedance as Action Space for Force-Accurate Motion Imitation” 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
Matthias Jammot, Pembe Gizem Özdil
ImpedanceMimic adapts joint stiffness, damping, and position targets to track recorded motion while matching ground-reaction forces. Measured forces supervise training, while the policy uses motion and a frozen kinematics-to-GRF estimate at inference. On held-out clips from seen subjects, it reaches 96.9% survival, 2.28° joint-angle RMSE, and 0.80 vertical-force correlation, versus Fixed-PD’s 93.0%, 2.91°, and 0.73. Future-motion input, up to 400 ms, limits live use.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.