Learning Roller-Skating Motions of Humanoid Robots Based on Adversarial Motion Priors

Yunkang Cheng^† Yutong Wu^† Menghan Li Shihe Zhou Mingguo Zhao^*

^† These authors contributed equally. ^* Corresponding author.

Overview Results Methods Acknowledgments

Abstract: Humanoid roller-skating is difficult because the robot must coordinate whole-body balance, rolling contacts, and velocity-dependent posture regulation. This paper presents an adversarial motion prior based reinforcement learning framework for two humanoid roller-skating gaits: Pump Glide skating and Push Glide skating. The two gait datasets are collected independently through motion capture and retargeted to the humanoid robot separately. The retargeted data are then smoothed and resampled into reference motion states for AMP training. The two gaits are learned by independent AMP training pipelines with separate reference datasets, separate policies, and independent reward architectures. Simulation experiments are designed to evaluate gait quality, velocity tracking, turning, and gait-specific reward ablations.

Pump Glide

Medium Friction Response

1x Speed Multiplier

1.5x Speed Multiplier

High Friction Response

1x Speed Multiplier

1.5x Speed Multiplier

Multi-Speed Switching Stability

Trial 1

Trial 2

Push Glide

Medium Friction Response

1x Speed Multiplier

1.5x Speed Multiplier

2x Speed Multiplier

High Friction Response

1x Speed Multiplier

1.5x Speed Multiplier

2x Speed Multiplier

Multi-Speed Switching Stability

Trial 1

Trial 2

Methods

Training: The actor receives partial observations and reconstructs the full state from historical data using an encoder-decoder architecture. The policy is trained with PPO, with rewards from both the environment and a discriminator encoding motion priors, while multiple critics provide value estimates.
Deployment: The real-world robot receives onboard proprioceptive observations, including IMU/gravity, joint states, and command inputs. The trained actor encodes the state history and outputs joint actions for real-time roller-skating control.

Acknowledgments

This research was supported by STI 2030-Major Projects grant number 2021ZD0201402, Beijing Natural Science Foundation (L243004), National Innovation and Entrepreneurship Training Program for College Students (No. 95), Tsinghua University Undergraduate Academic Research Promotion Program, and Spark Program ("Science and Technology Innovation, Spark the Prairie Fire" Tsinghua University Student Innovation Talent Training Program).

We thank Booster Robotics for providing the robot platform, experimental environment, and technical support. We thank SongRui Huang, Zhihan Li, Yiyi Zheng, Yihui Zhang and Lushu Yang for assistance in data collection and experiments.

BibTeX

@misc{cheng2026learning,
  title   = {Learning Roller-Skating Motions of Humanoid Robots Based on Adversarial Motion Priors},
  author  = {Cheng, Yunkang and Wu, Yutong and Li, Menghan and Zhou, Shihe and Zhao, Mingguo},
  year    = {2026},
}