Tag
SceneBot is a unified motion tracking framework that conditions a single humanoid policy on both reference motions and per-link contact labels, enabling free-space locomotion, terrain traversal, and whole-body manipulation in contact-rich environments.