Multi-Gait Learning for Humanoid Robots Using Reinforcement Learning with Selective Adversarial Motion Prior

2026-04-21

AI Summary

This work addresses the challenge of enabling humanoid robots to master diverse gaits within a unified reinforcement learning framework, where conflicting requirements between stability and agility often hinder performance. The authors propose a selective Adversarial Motion Prior (AMP) mechanism that activates AMP only for highly periodic and stability-critical gaits—such as walking, marching, and stair climbing—to accelerate convergence and suppress aberrant behaviors, while deactivating AMP for high-dynamic maneuvers like running and jumping to preserve agility. Trained in simulation using PPO with domain randomization, the policy achieves zero-shot transfer to a 12-degree-of-freedom humanoid platform. Experiments demonstrate superior performance across five gaits compared to uniformly applying AMP: stability-oriented gaits exhibit faster convergence, lower tracking error, and higher success rates, without compromising the agility of dynamic motions.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Intelligent Robots: Motion and Path Planning
• Multiagent Systems: Adversarial Agents

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Learning diverse locomotion skills for humanoid robots in a unified reinforcement learning framework remains challenging due to the conflicting requirements of stability and dynamic expressiveness across different gaits. We present a multi-gait learning approach that enables a humanoid robot to master five distinct gaits -- walking, goose-stepping, running, stair climbing, and jumping -- using a consistent policy structure, action space, and reward formulation. The key contribution is a selective Adversarial Motion Prior (AMP) strategy: AMP is applied to periodic, stability-critical gaits (walking, goose-stepping, stair climbing) where it accelerates convergence and suppresses erratic behavior, while being deliberately omitted for highly dynamic gaits (running, jumping) where its regularization would over-constrain the motion. Policies are trained via PPO with domain randomization in simulation and deployed on a physical 12-DOF humanoid robot through zero-shot sim-to-real transfer. Quantitative comparisons demonstrate that selective AMP outperforms a uniform AMP policy across all five gaits, achieving faster convergence, lower tracking error, and higher success rates on stability-focused gaits without sacrificing the agility required for dynamic ones.

Problem

• humanoid robots
• multi-gait learning
• reinforcement learning
• stability
• dynamic expressiveness

Innovation

• Selective AMP
• multi-gait learning
• humanoid locomotion
• reinforcement learning
• sim-to-real transfer