Discrepancy between IK and FK in retargeting human motion to humanoids

LeCAR-Lab / human2humanoid

[IROS 2024] Learning Human-to-Humanoid Real-Time Whole-Body Teleoperation. [CoRL 2024] OmniH2O: Universal and Dexterous Human-to-Humanoid Whole-Body Teleoperation and Learning

https://omni.human2humanoid.com/

289 stars 16 forks source link

Discrepancy between IK and FK in retargeting human motion to humanoids #25

Closed Maxwell-Zhao closed 1 hour ago

Maxwell-Zhao commented 9 hours ago

In the paper, the authors mention that they retarget human motions to the humanoid using inverse kinematics, but in the file grad_fit_h1_shape.py, it appears that forward kinematics is used instead of IK.

QQ_1733317416654

Maxwell-Zhao commented 9 hours ago

This is a FK model in file grad_fit_h1_shape.py. QQ_1733317691605

TairanHe commented 1 hour ago

The IK is implemented based on optimization with FK.