UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling
UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling
This research presents UniT, a framework that enables human-to-humanoid policy learning and world modeling by creating a unified physical language. The method uses a tri-branch cross-reconstruction mechanism and a shared latent space to achieve state-of-the-art data efficiency and zero-shot task transfer, validated on both simulation and real-world deployments.
This paper introduces UniT, a framework that bridges the gap between human and humanoid robot policy learning by creating a unified physical language. The approach leverages cross-reconstruction mechanisms and a shared latent space to enable zero-shot task transfer and improved world modeling.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org