Learning Impact-Rich Rotational Maneuvers via Centroidal Velocity Rewards and Sim-to-Real Techniques: A One-Leg Hopper Flip Case Study

TL;DR

This paper introduces a framework for learning impact-rich rotational behaviors in robots using centroidal velocity rewards and sim-to-real techniques, demonstrated through a successful one-leg hopper flip.

Contribution

It proposes a novel centroidal angular velocity reward and actuator-aware sim-to-real methods, enabling the first hardware realization of a full front flip in a robotic hopper.

Findings

Centroidal velocity rewards improve rotation accuracy.

Modeling motor regions enhances sim-to-real transfer.

Successful hardware flip demonstrates method effectiveness.

Abstract

Dynamic rotational maneuvers, such as front flips, inherently involve large angular momentum generation and intense impact forces, presenting major challenges for reinforcement learning and sim-to-real transfer. In this work, we propose a general framework for learning and deploying impact-rich, rotation-intensive behaviors through centroidal velocity-based rewards and actuator-aware sim-to-real techniques. We identify that conventional link-level reward formulations fail to induce true whole-body rotation and introduce a centroidal angular velocity reward that accurately captures system-wide rotational dynamics. To bridge the sim-to-real gap under extreme conditions, we model motor operating regions (MOR) and apply transmission load regularization to ensure realistic torque commands and mechanical robustness. Using the one-leg hopper front flip as a representative case study, we demonstrate the first successful hardware realization of a full front flip. Our results highlight that incorporating centroidal dynamics and actuator constraints is critical for reliably executing highly dynamic motions. A supplementary video is available at: https://youtu.be/atMAVI4s1RY