CoorDex: Coordinating Body and Hand Priors for Continuous Dexterous Humanoid Loco-Manipulation

Humanoid loco-manipulation is often simplified as walking to an object, stopping to manipulate it, and then resuming locomotion. It also commonly relies on low-degree-of-freedom end effectors whose behavior is close to an open-close grasp primitive. We introduce CoorDex, a learning pipeline that converts high-dimensional body and dexterous hand control into coordinated latent residual control, enabling high-DoF dexterous loco-manipulation. Starting from simulated whole-body and hand demonstrations, CoorDex trains privileged motion-tracking teachers for the humanoid body and dexterous hand, distills them into proprioception-conditioned latent priors, and uses the frozen priors as the action space for downstream residual reinforcement learning. A Coordinated Latent Residual Policy composes these priors through shared task context and separate body-hand residual heads, preserving natural whole-body motion while improving finger-level contact accuracy. CoorDex enables a Unitree G1 humanoid with a 20-DoF Wuji hand to perform several high-DoF dexterous loco-manipulation skills, including non-stop bottle grasping and carrying, fridge-door opening, and cube pick-and-turn. Controlled ablations on the walk-grasp task show that direct joint-space PPO and unstructured latent prediction fail under the same reward budget, while the proposed latent-prior interface and coordinated residual structure make high-dimensional contact-rich loco-manipulation trainable.