Learning Roller-Skating Motions of Humanoid Robots Based on Adversarial Motion Priors
2026-07-12 • Robotics
Robotics
AI summaryⓘ
The authors explore how to help humanoid robots roller-skate by teaching them two different skating styles using a learning method called adversarial motion prior (AMP). They first record human skating movements and adapt these to the robot's body, then train the robots separately on each skating style using those adapted movements. Their approach helps the robot balance, roll, and change direction while skating. They test the robots in simulations to see how well they skate and respond to different challenges.
humanoid robotroller-skatingreinforcement learningadversarial motion priormotion capturegaitpolicy trainingvelocity trackingreward architecturesimulation
Authors
Yunkang Cheng, Yutong Wu, Menghan Li, Shihe Zhou, Mingguo Zhao
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.