Reinforcement Learning-Based Control for an Inline Skating Humanoid Robot

2026-06-30

AI Summary

This work addresses the underactuated, highly dynamic, and unstable control challenges faced by humanoid robots equipped with passive inline roller skates. The authors propose a reinforcement learning approach that requires no human motion priors or imitation learning, enabling the robot to autonomously discover energy-efficient, dynamic skating gaits through carefully designed reward functions. Key innovations include a heterogeneous geometric wheel model (spherical and ellipsoidal), a success-rate-based curriculum for command training, and a dedicated rolling reward mechanism, collectively achieving precise six-degree-of-freedom control. The learned policy is successfully deployed zero-shot on the real-world Booster T1 robot, reducing energy consumption by 50% compared to conventional walking while demonstrating robustness to disturbances and agile capabilities such as high-speed turning.

Abstract

As humanoid robots become increasingly dynamic, coupling them with reinforcement learning offers a promising approach to solving the complex, underactuated mechanics of passive inline skating. Equipping a humanoid robot with passive inline skating wheels presents an opportunity to combine the versatile agility of humanoids with the high-speed, energy-efficient locomotion strategies utilized by human skaters. In this paper, we train and deploy a reinforcement learning control policy that enables novel locomotion strategies for a humanoid robot modified to equip consumer inline skates instead of conventional feet. Unlike previous work limited to quadrupedal robots or actively driven wheels, our system allows for precise 6-DoF control of the skates to execute dynamic, edge-driven propulsion strategies. Our skating strategies emerge entirely from our reward structure, without reliance on human motion data, imitation learning, or kinematic priors. We overcome the inherent instability of passive wheels and simulation contact artifacts by utilizing different geometric wheel models (spherical and ellipsoidal) during training and validation, along with a custom success-based command curriculum and a specialized rolling reward. Consequently, our policy demonstrates up to a 50% reduction in Cost of Transport (CoT) compared to standard walking gaits. The resulting policy successfully transfers zero-shot to the physical Booster T1 hardware. Real-world deployments demonstrate dynamic balance, the ability to reject active physical perturbations, and agile locomotion strategies capable of turning at speed. A video of our results can be found at https://www.youtube.com/watch?v=-_APcOS7uFo.

Problem

• humanoid robot
• inline skating
• reinforcement learning
• underactuated locomotion
• dynamic balance

Innovation

• reinforcement learning
• humanoid robot
• passive inline skating
• zero-shot transfer
• edge-driven propulsion