Get Started
Research questionHow can human motion be retargeted across different humanoid morphologies without hand-crafted mappings or losing pose detail?Human and humanoid bodies differ in morphology, degrees of freedom, joint ranges, and kinematic constraints, making direct motion transfer unreliable. Sparse, manually chosen correspondences provide limited guidance for reproducing detailed poses and interactions across robot embodiments.
Computer Vision
Machine Learning
Robotics
Latest papersRecent research connected to this question, newest first.Unified Motion Retargeting for Humanoids with Learned Point Cloud CorrespondenceThe source studies Unified Motion Retargeting, which uses learned dense correspondence between exterior human and robot point clouds for constrained point-cloud matching. Evidence covers heterogeneous motion sources and humanoid embodiments, with downstream scenarios including locomotion and interaction; reported outcomes concern motion fidelity and plausibility.research paper · Sep 7, 2026
Related questions
How can humanoid tracking policies learn pose transitions from unordered data without ordered motion clips?How can humanoid morphology and whole-body control be co-designed to reproduce human movement?How can language-driven humanoid control follow instructions while staying physically plausible and stable over long horizons?How can humanoid lower-limb joint offsets be calibrated without external tracking when pitch-chain errors are weakly observable?
Home
Topics
Search
Library