Unverified50% confidenceFactExact time
Humans and robots are inconsistent in kinematic structure, joint ranges, and link lengths, so directly mapping human trajectories to the robot produces residual pose errors, leading to viewpoint mismatch.
1
Sources
50%
Confidence
Long-term
Relevance
11/3/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedExisting robot datasets suffer from imprecise calibration: errors in camera calibration or kinematic parameters distort the correspondence between actions and visual outcomes.78% similarUnverifiedEmbodiment gap refers to the structural differences in physical form between different agents (such as humans and robots), manifested in aspects such as kinematic structure, joint range of motion, link lengths and mass distribution, and sensor configuration.74% similarUnverifiedDistortion of physical properties in 3D assets (e.g., drawer friction coefficients, object center-of-mass offsets) causes policies learned by robots in simulation to be difficult to transfer to real environments, a phenomenon known as the sim-to-real gap.71% similarUnverifiedPixel's human-aligned kinematic design aims to narrow the morphology gap between human demonstration data and robot execution, reducing retargeting distortion.69% similarUnverifiedEgoExplorer's contributions include: placing vision, motion, touch, and finger details into the same wearable collection framework; using a calibration-plus-constraint method to demonstrate that sparse sensing can also yield relatively high-quality motion reconstruction; and directly linking the dataset with humanoid retargeting, shortening the path from human demonstration to robot policy.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1130376API
curl https://kongchang.com/api/v1/knowledge/claims/1130376MCP
get_claim(id=1130376)