T-GMP: Terrain-conditioned Generative Motion Priors for Versatile and Natural Humanoid Locomotion
For humanoid robot locomotion, this work addresses the challenge of adapting motion priors to varying terrains while maintaining naturalness.
T-GMP introduces a terrain-conditioned generative motion prior that enables humanoid robots to achieve both natural locomotion and robust terrain traversal, outperforming baselines in success rate and motion smoothness.
Achieving both anthropomorphic naturalness and robust terrain traversal remains a fundamental challenge in humanoid locomotion. Existing Reinforcement Learning (RL) approaches typically rely on fixed motion priors, limiting their adaptability to varying environments. We propose Terrain-conditioned Generative Motion Priors (T-GMP), a module that captures a terrain-conditioned latent motion manifold from a few expert state-terrain demonstrations using a Conditional Variational Autoencoder (CVAE). The learned priors enable smooth style transitions, facilitating a unified policy that adapts to terrain variations. We integrate T-GMP into an adversarial learning pipeline with our proposed Foothold Penalty, where a discriminator dynamically modulates naturalness constraints conditioned on local terrain features, guiding the generation of versatile and human-like motions. Experimental results demonstrate that our method outperforms existing baselines in traversal success rate and motion smoothness, while preserving biomimetically natural and physically coordinated motions.