Efficient and Versatile Quadrupedal Skating: Optimal Co-design via Reinforcement Learning and Bayesian Optimization

2026-03-18

AI Summary

This work addresses the challenge of achieving efficient and agile gliding in passive wheeled quadrupedal robots, whose mechanical design and motion control are tightly coupled due to the absence of actuation at the wheels. To overcome this limitation, the authors propose a hardware-control co-design approach formulated as a two-layer optimization framework: the upper layer employs Bayesian optimization to search over mechanical designs, while the lower layer uses reinforcement learning to train a dedicated control policy for each candidate morphology. This methodology enables, for the first time at the system level, dynamic gliding behaviors in quadrupedal robots, successfully demonstrating complex maneuvers such as hockey-stop braking and self-alignment. The resulting designs significantly outperform handcrafted baselines, achieving substantial improvements in both energy efficiency and locomotion performance during high-speed gliding.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Search and Optimization: Sampling/Simulation-based Search
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

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.

Problem

• quadrupedal skating
• passive wheels
• hardware-control co-design
• dynamic locomotion
• mechanical design and control coupling

Innovation

• hardware-control co-design
• reinforcement learning
• Bayesian optimization
• passive-wheel skating
• quadrupedal locomotion