AI RESEARCH

SCDP: Learning Humanoid Locomotion from Partial Observations via Mixed-Observation Distillation

arXiv CS.LG

ArXi:2603.09574v1 Announce Type: cross Distilling humanoid locomotion control from offline datasets into deployable policies remains a challenge, as existing methods rely on privileged full-body states that require complex and often unreliable state estimation. We present Sensor-Conditioned Diffusion Policies (SCDP) that enables humanoid locomotion using only onboard sensors, eliminating the need for explicit state estimation. SCDP decouples sensing from supervision through mixed-observation.