Learning to Walk in Costume: Adversarial Motion Priors for Aesthetically Constrained Humanoids
This work addresses the challenge of balancing aesthetic appeal with functional performance in entertainment robots, representing an incremental advancement in applying learning-based methods to domain-specific constraints.
The paper tackled the problem of enabling a custom-built humanoid robot with aesthetic-driven design constraints, such as a disproportionately large head and restricted movement, to learn natural-looking locomotion using Reinforcement Learning with Adversarial Motion Priors, resulting in stable standing and walking behaviors.
We present a Reinforcement Learning (RL)-based locomotion system for Cosmo, a custom-built humanoid robot designed for entertainment applications. Unlike traditional humanoids, entertainment robots present unique challenges due to aesthetic-driven design choices. Cosmo embodies these with a disproportionately large head (16% of total mass), limited sensing, and protective shells that considerably restrict movement. To address these challenges, we apply Adversarial Motion Priors (AMP) to enable the robot to learn natural-looking movements while maintaining physical stability. We develop tailored domain randomization techniques and specialized reward structures to ensure safe sim-to-real, protecting valuable hardware components during deployment. Our experiments demonstrate that AMP generates stable standing and walking behaviors despite Cosmo's extreme mass distribution and movement constraints. These results establish a promising direction for robots that balance aesthetic appeal with functional performance, suggesting that learning-based methods can effectively adapt to aesthetic-driven design constraints.