RPG: Robust Policy Gating for Smooth Multi-Skill Transitions in Humanoid Fighting

TL;DR

This paper introduces RPG, a hybrid expert policy framework that enables humanoid robots to perform smooth, stable, and long-duration multi-skill fighting by improving transition stability and integrating locomotion with combat skills.

Contribution

The work presents a novel policy training method with motion and temporal randomization for seamless skill transitions in humanoid robots, validated in simulation and real-world tests.

Findings

The RPG framework achieves smooth and stable skill transitions in simulation.

The approach enables long-duration humanoid combat with seamless skill switching.

Real-world experiments demonstrate robustness and applicability of the method.

Abstract

Humanoid robots have demonstrated impressive motor skills in a wide range of tasks, yet whole-body control for humanlike long-time, dynamic fighting remains particularly challenging due to the stringent requirements on agility and stability. While imitation learning enables robots to execute human-like fighting skills, existing approaches often rely on switching among multiple single-skill policies or employing a general policy to imitate input reference motions. These strategies suffer from instability when transitioning between skills, as the mismatch of initial and terminal states across skills or reference motions introduces out-of-domain disturbances, resulting in unsmooth or unstable behaviors. In this work, we propose RPG, a hybrid expert policy framework, for smooth and stable humanoid multi-skills transition. Our approach incorporates motion transition randomization and temporal randomization to train a unified policy that generates agile fighting actions with stability and smoothness during skill transitions. Furthermore, we design a control pipeline that integrates walking/running locomotion with fighting skills, allowing humanlike long-time combat of arbitrary duration that can be seamlessly interrupted or transit action policies at any time. Extensive experiments in simulation demonstrate the effectiveness of the proposed framework, and real-world deployment on the Unitree G1 humanoid robot further validates its robustness and applicability.