analiza aiqbench
This research introduces a reusable hybrid motion prior that distills human motion imitation skills into a frozen architecture, allowing downstream policies to select gait patterns without retraining. By decoupling locomotion priors from task-specific rewards, the method enables efficient training for complex behaviors like navigation and fall recovery on real-world humanoid robots. The interpretable codebook structure demonstrates how discrete latent variables can effectively modulate available movement strategies in reinforcement learning pipelines.
Learning Reusable Hybrid Motion Priors for Humanoid Locomotion from Motion Imitation
arXiv:2607.24083v1 Announce Type: cross Abstract: Reinforcement learning can produce robust humanoid controllers, but each new task is typically trained as a separate policy with its own reward design and training process. Motion imitation provides an alternative source of motor competence by training policies to track retargeted human motions, yet the resulting controllers remain reference trackers and are not directly usable as task policies. We propose a three-stage pipeline that turns motion-imitation skills into a reusable hybrid motion prior (HMP) for humanoid locomotion. First, an expert policy is trained to imitate retargeted human motion-capture clips. Second, the expert is distilled into a frozen architecture composed of a proprioceptive encoder, a residual vector-quantized (RVQ) codebook, and an action decoder. Third, task-level policies are trained to solve locomotion tasks by selecting discrete codebook entries while the HMP remains frozen. We evaluate the method on velocity tracking, point-goal navigation, and fall-recovery velocity tracking in simulation, and deploy the velocity-tracking policy on a real Unitree G1 robot. The distillation process preserves the tracking behavior of the expert, while the resulting HMP can be reused without retraining as the action interface for different downstream locomotion policies. The learned HMP reveals an interpretable codebook structure in which the number of active RVQ stages modulates the available gait patterns. We further show that training the codebook with the rotation trick improves latent organization and reduces downstream falls compared with a standard straight-through estimator.
Abstract:Reinforcement learning can produce robust humanoid controllers, but each new task is typically trained as a separate policy with its own reward design and training process. Motion imitation provides an alternative source of motor competence by training policies to track retargeted human motions, yet the resulting controllers remain reference trackers and are not directly usable as task policies. We propose a three-stage pipeline that turns motion-imitation skills into a reusable hybrid motion prior (HMP) for humanoid locomotion. First, an expert policy is trained to imitate retargeted human motion-capture clips. Second, the expert is distilled into a frozen architecture composed of a proprioceptive encoder, a residual vector-quantized (RVQ) codebook, and an action decoder. Third, task-level policies are trained to solve locomotion tasks by selecting discrete codebook entries while the HMP remains frozen. We evaluate the method on velocity tracking, point-goal navigation, and fall-recovery velocity tracking in simulation, and deploy the velocity-tracking policy on a real Unitree G1 robot. The distillation process preserves the tracking behavior of the expert, while the resulting HMP can be reused without retraining as the action interface for different downstream locomotion policies. The learned HMP reveals an interpretable codebook structure in which the number of active RVQ stages modulates the available gait patterns. We further show that training the codebook with the rotation trick improves latent organization and reduces downstream falls compared with a standard straight-through estimator.
Machine Learning (cs.LG); Robotics (cs.RO)
arXiv:2607.24083 [cs.LG]
(or arXiv:2607.24083v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.24083 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 326 cuvinte
Subiecte conexe
Unifying Physics-Based Humanoid Interaction with a Context-Conditioned Interaction Prior
arXiv · 9 sept. 2026
Perceptive Behavior Foundation Model: Adapting Human Motion Priors to Robot-Centric Terrain
arXiv · 16 iun. 2026
Perceptive Behavior Foundation Model: Adapting Human Motion Priors to Robot-Centric Terrain
arXiv · 9 iun. 2026
EgoExplorer: A Vision-Motion-Touch Observatory for Human Loco-Manipulation In Situ
arXiv · 14 oct. 2026