The Role of Domain Randomization in Training Diffusion Policies for Whole-Body Humanoid Control

TL;DR

This paper explores how dataset diversity and size, enhanced through domain randomization, impact the training of diffusion policies for humanoid whole-body control, revealing larger datasets are necessary for effective locomotion policies.

Contribution

It demonstrates the importance of dataset size and diversity, via domain randomization, for training diffusion policies in humanoid locomotion, an area less explored compared to manipulation.

Findings

Diffusion policies can achieve stable walking in simulation.

Larger and more diverse datasets are needed for successful humanoid locomotion.

Domain randomization improves dataset robustness for policy training.

Abstract

Humanoids have the potential to be the ideal embodiment in environments designed for humans. Thanks to the structural similarity to the human body, they benefit from rich sources of demonstration data, e.g., collected via teleoperation, motion capture, or even using videos of humans performing tasks. However, distilling a policy from demonstrations is still a challenging problem. While Diffusion Policies (DPs) have shown impressive results in robotic manipulation, their applicability to locomotion and humanoid control remains underexplored. In this paper, we investigate how dataset diversity and size affect the performance of DPs for humanoid whole-body control. In a simulated IsaacGym environment, we generate synthetic demonstrations by training Adversarial Motion Prior (AMP) agents under various Domain Randomization (DR) conditions, and we compare DPs fitted to datasets of different size and diversity. Our findings show that, although DPs can achieve stable walking behavior, successful training of locomotion policies requires significantly larger and more diverse datasets compared to manipulation tasks, even in simple scenarios.