Scalable and General Whole-Body Control for Cross-Humanoid Locomotion

TL;DR

This paper presents XHugWBC, a universal humanoid control framework that generalizes across various robot designs through a novel training method, enabling zero-shot transfer to unseen robots.

Contribution

Introduces a cross-embodiment training framework that enables a single policy to control diverse humanoid robots without robot-specific training.

Findings

Successfully controls 12 simulated humanoids and 7 real robots.

Achieves zero-shot transfer to unseen robot morphologies.

Demonstrates robustness and generalization in diverse scenarios.

Abstract

Learning-based whole-body controllers have become a key driver for humanoid robots, yet most existing approaches require robot-specific training. In this paper, we study the problem of cross-embodiment humanoid control and show that a single policy can robustly generalize across a wide range of humanoid robot designs with one-time training. We introduce XHugWBC, a novel cross-embodiment training framework that enables generalist humanoid control through: (1) physics-consistent morphological randomization, (2) semantically aligned observation and action spaces across diverse humanoid robots, and (3) effective policy architectures modeling morphological and dynamical properties. XHugWBC is not tied to any specific robot. Instead, it internalizes a broad distribution of morphological and dynamical characteristics during training. By learning motion priors from diverse randomized embodiments, the policy acquires a strong structural bias that supports zero-shot transfer to previously unseen robots. Experiments on twelve simulated humanoids and seven real-world robots demonstrate the strong generalization and robustness of the resulting universal controller.