ModSkill: Physical Character Skill Modularization
This work addresses the problem of scaling imitation learning to larger motion datasets for researchers and developers in robotics and animation, representing an incremental advancement over previous methods.
The paper tackles the challenge of generalizing motor skills for simulated characters by introducing ModSkill, a framework that decouples full-body skills into modular components for independent body parts, resulting in improved performance in precise full-body motion tracking and reusable skill embeddings for diverse tasks.
Human motion is highly diverse and dynamic, posing challenges for imitation learning algorithms that aim to generalize motor skills for controlling simulated characters. Previous methods typically rely on a universal full-body controller for tracking reference motion (tracking-based model) or a unified full-body skill embedding space (skill embedding). However, these approaches often struggle to generalize and scale to larger motion datasets. In this work, we introduce a novel skill learning framework, ModSkill, that decouples complex full-body skills into compositional, modular skills for independent body parts. Our framework features a skill modularization attention layer that processes policy observations into modular skill embeddings that guide low-level controllers for each body part. We also propose an Active Skill Learning approach with Generative Adaptive Sampling, using large motion generation models to adaptively enhance policy learning in challenging tracking scenarios. Our results show that this modularized skill learning framework, enhanced by generative sampling, outperforms existing methods in precise full-body motion tracking and enables reusable skill embeddings for diverse goal-driven tasks.