Motor Babble: Morphology Driven Coordinated Control of Articulated Characters

Overview

Locomotion in humans and animals is highly coordinated. Learning similar coordinated locomotion in virtual characters without reference motion data is challenging due to high dimensionality and redundancy. This paper presents a method for learning locomotion in a low-dimensional latent space defining joint coordination. We introduce "motor babble," where a character interacts with its environment through uncoordinated motor excitations. This generates a motion corpus from which a manifold latent space is extracted. Dimensions of this manifold represent character synergies. Using reinforcement learning, we train the character to locomote in this latent space by selecting appropriate dimensions and learning the corresponding policy.

Key Contributions

• Motor Babble corpus: Generate character-specific motion data via uncoordinated motor excitations (no reference motion needed).
• Synergy (coactivation) latent space: Extract a low-dimensional joint-coordination manifold from the babble corpus (e.g., SVD-based coactivations).
• RL in latent control: Train locomotion by selecting and exciting a small set of synergies instead of controlling all joints directly.
• Less reward engineering: Achieve coordinated gaits with relatively simple task rewards (e.g., target-velocity tracking) thanks to synergy control.
• Morphology generalization + validation: Demonstrate robust coordinated locomotion across diverse articulated morphologies and compare against joint-space baselines/variants.