Efficient and Versatile Quadrupedal Skating: Optimal Co-Design Via Reinforcement Learning and Bayesian Optimization
Hanwen Wang, Zhenlong Fang, Josiah Hanna, Xiaobin Xiong
AI summary
Problem
Skating with passive wheels on quadrupeds tightly couples mechanical design and control, making it difficult to achieve stable, efficient, and versatile locomotion without manual tuning. Existing heuristic or decoupled approaches fail to fully exploit this design-control interaction.
Approach
A bilevel optimization framework that uses Bayesian Optimization to search for optimal wheel installation angles while Reinforcement Learning trains a tailored motor policy for each candidate design.
Key results
- Discovers optimal wheel yaw angles that significantly reduce energy consumption compared to human-engineered baselines
- Enables versatile transient behaviors like hockey stops and self-aligning motion through world-frame velocity tracking
- Demonstrates the necessity of hardware-control co-design for achieving efficient skating across multiple directions
- Provides the first systematic study of dynamic skating motion on quadrupedal robots with passive wheels
Why it matters
It offers a practical, automated co-design methodology for hybrid legged-wheeled robots, enabling faster, more energy-efficient, and versatile locomotion for real-world deployment.
Abstract
In this paper, we present a hardware-control co- design approach that enables efficient and versatile roller skating on quadrupedal robots equipped with passive wheels. Passive-wheel skating reduces leg inertia and improves energy efficiency, particularly at high speeds. However, the absence of direct wheel actuation tightly couples mechanical design and control. To unlock the full potential of this modality, we formulate a bilevel optimization framework: an upper-level Bayesian Optimization searches the mechanical design space, while a lower-level Reinforcement Learning trains a motor control policy for each candidate design. The resulting design- policy pairs not only outperform human-engineered baselines, but also exhibit versatile behaviors such as hockey stop (rapid braking by turning sideways to maximize friction) and self- aligning motion (automatic reorientation to improve energy efficiency in the direction of travel), offering the first system- level study of dynamic skating motion on quadrupedal robots.