Think on your feet: Seamless Transition between Human-like Locomotion in Response to Changing Commands

2025-02-26

AI Summary

This work addresses key challenges in humanoid robot locomotion: inaccurate motion tracking under dynamic commands, discontinuous multi-gait transitions, and poor generalization to unseen intermediate motions. To tackle these, we propose a novel imitation learning framework integrating Wasserstein divergence optimization (WGAN-div), hybrid internal model-based state estimation—jointly modeling latent states and velocity—and curiosity-driven exploration. Our approach overcomes bottlenecks in intermediate motion synthesis and cross-domain zero-shot transfer, enabling real-time velocity tracking and seamless multi-task switching. Evaluated in simulation and on diverse real-world terrains, the method achieves high gait fidelity, sub-80-ms command response latency, and successful zero-shot transfer across heterogeneous robot platforms. Results demonstrate significant improvements in dynamic adaptability and generalization capability, particularly for complex, time-varying locomotion tasks.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

While it is relatively easier to train humanoid robots to mimic specific locomotion skills, it is more challenging to learn from various motions and adhere to continuously changing commands. These robots must accurately track motion instructions, seamlessly transition between a variety of movements, and master intermediate motions not present in their reference data. In this work, we propose a novel approach that integrates human-like motion transfer with precise velocity tracking by a series of improvements to classical imitation learning. To enhance generalization, we employ the Wasserstein divergence criterion (WGAN-div). Furthermore, a Hybrid Internal Model provides structured estimates of hidden states and velocity to enhance mobile stability and environment adaptability, while a curiosity bonus fosters exploration. Our comprehensive method promises highly human-like locomotion that adapts to varying velocity requirements, direct generalization to unseen motions and multitasking, as well as zero-shot transfer to the simulator and the real world across different terrains. These advancements are validated through simulations across various robot models and extensive real-world experiments.

Problem

• Seamless transition between human-like locomotion
• Adapt to continuously changing motion commands
• Enhance generalization and environment adaptability

Innovation

• Human-like motion transfer
• Wasserstein divergence criterion
• Hybrid Internal Model