MimicAll enables motion tracking across different humanoid robots
MimicAll: Structured Embodiment Conditioning for Cross-Humanoid Whole-Body Motion Tracking
Robotics
Summary
Humanoid robots come in many shapes and sizes, making it hard to use one method to control all of them. The authors created MimicAll, a system that helps a single learning model track and reproduce motions across many different humanoid bodies. It uses special information about each robot's body to better match movements and adapt to new robots it hasn’t seen before. Tests showed that MimicAll works much better than models trained for just one robot at a time.
What this means in practice
- •For robotics engineers: Train a single control policy that tracks whole-body motions across various humanoid robot designs without retraining for each new body.
- •For animation studios: Use a shared motion tracking system to animate characters with different body types from the same motion data more efficiently.
Authors
Jia Li, Lingyun Xu, Yitang Li, Hao Zhao
Abstract
Humanoid motion trackers reproduce diverse behaviors, but sharing a policy across bodies requires accounting for differences in structure, morphology and dynamics. We present MimicAll, a reinforcement-learning framework for shared whole-body motion tracking across heterogeneous humanoids. A structured embodiment descriptor is encoded into global and joint-aligned conditions, which are combined with observation history and body-matched motion references to predict joint-target offsets. Physics-consistent embodiment randomization, body-calibrated retargeting and heterogeneous simulation support training across 4,096 variants from 16 humanoid families. We introduce a benchmark of 464 bodies across five distribution strata to evaluate motion generalization, within-family embodiment shifts and unseen-family transfer. Controlled ablations show that embodiment information improves motion completion with simple concatenation, while self-attention yields the lowest global position errors. With equal training transitions, shared training achieves 99.4% 30-second success on randomized G1 bodies in 100STYLE, versus 42.9% for a single-body specialist. With Flat-MLP encoding and concatenation, continued training on a motion subset excluding obstacle categories achieves 98.3% and 93.5% 30-second success on 100STYLE for sampled training bodies and out-of-distribution variants within seen families, respectively. On an eight-motion LAFAN1 subset excluding ground and get-up sequences across 464 bodies, mismatched body conditions reduce self-attention's 30-second success from 61.7% to 47.0%, demonstrating a closed-loop benefit from embodiment information.