X-Morph: Human Motion Priors for Scalable Robot Learning Across Morphologies

2026-06-29

AI Summary

This work addresses the challenge of transferring human motion priors to non-anthropomorphic legged robots, which typically lack large-scale locomotion datasets for direct behavioral learning. The authors propose X-Morph, a framework that enables, for the first time, generalizable behavior prior transfer from human motion to diverse non-humanoid legged platforms—including quadrupeds, hexapods, and arm-equipped quadrupeds. X-Morph integrates cross-morphology motion retargeting to generate kinematically feasible reference trajectories and leverages privileged reinforcement learning combined with causal policy distillation for efficient policy transfer. Experiments demonstrate that X-Morph successfully deploys a wide range of locomotion skills across three distinct robot morphologies, generalizes to unseen human motions, and supports downstream applications such as video-based teleoperation, behavior-prior-guided control, and text-driven motion generation.

Abstract

Recent progress in humanoid behavior models has been driven in large part by abundant human motion data, but comparable motion data is scarce for non-humanoid legged robots such as quadrupeds, hexapods, and quadruped manipulators. A promising alternative is to repurpose human motion across embodiments; however, direct retargeting often produces motions that are visually plausible yet physically inconsistent or difficult to track under robot dynamics. We present X-Morph, a human-motion-to-robot-behavior pipeline that converts human motion into deployable locomotion and loco-manipulation policies for diverse non-humanoid legged morphologies. A cross-morphology retargeting stage converts human motions into kinematically plausible, intent-preserving robot references, which are then tracked by a privileged RL policy and distilled into a causal student policy. We evaluate X-Morph on three morphologically distinct platforms: a quadruped, a hexapod, and a quadruped equipped with a manipulator. The resulting policies track diverse retargeted motions, generalize to unseen human motions, and support downstream use cases including video-based teleoperation, behavior-prior control, and text-conditioned motion generation. These results suggest that large-scale human motion can serve as a substrate for learning broad, reusable behavior priors beyond humanoid robots. Project page: https://maker-rat.github.io/morph/

Problem

• human motion priors
• non-humanoid robots
• motion retargeting
• legged locomotion
• behavior transfer

Innovation

• cross-morphology retargeting
• human motion priors
• privileged reinforcement learning
• policy distillation
• loco-manipulation