FACT: A Multimodal Force-Action-Track Foundation Model for Human-to-Robot Dexterity
아직 한국어판이 없어 영어 원문으로 표시.
초록 발췌
Human demonstrations offer a scalable alternative to costly robot data collection for dexterous manipulation, but most human datasets lack force measurements. Dexterous manipulation requires not only visual and proprioceptive understanding but also the ability to anticipate and regulate interaction forces.
초록에서 가져왔습니다. 요약을 준비 중입니다.