Researchers have developed OTRetarget, a novel method for transferring human motion to humanoid robots, including the manipulation of objects. This approach uses entropic optimal transport to maintain consistency in interactions with the ground and objects, even when body proportions differ. OTRetarget jointly optimizes robot and object poses, achieving a high Jaccard score for robot-object interaction and low depth error in simulations. The method has been successfully demonstrated on a physical G1 humanoid robot for tasks like pick-and-place operations. AI
IMPACT Enables more realistic and adaptable robot behaviors in complex manipulation tasks.
RANK_REASON The cluster contains an academic paper detailing a new method for motion retargeting in robotics. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →