Multi-Loco: Unifying Multi-Embodiment Legged Locomotion via Reinforcement Learning Augmented Diffusion

TL;DR

Multi-Loco introduces a unified framework combining a morphology-agnostic diffusion model with reinforcement learning to improve generalization and robustness of legged robot locomotion across diverse morphologies in simulation and real-world tests.

Contribution

It presents a novel approach that unifies diffusion models with residual RL policies for cross-embodiment locomotion, enhancing generalization and task performance.

Findings

10.35% average return improvement over standard RL

Gains up to 13.57% in wheeled-biped tasks

Effective in both simulation and real-world experiments

Abstract

Generalizing locomotion policies across diverse legged robots with varying morphologies is a key challenge due to differences in observation/action dimensions and system dynamics. In this work, we propose Multi-Loco, a novel unified framework combining a morphology-agnostic generative diffusion model with a lightweight residual policy optimized via reinforcement learning (RL). The diffusion model captures morphology-invariant locomotion patterns from diverse cross-embodiment datasets, improving generalization and robustness. The residual policy is shared across all embodiments and refines the actions generated by the diffusion model, enhancing task-aware performance and robustness for real-world deployment. We evaluated our method with a rich library of four legged robots in both simulation and real-world experiments. Compared to a standard RL framework with PPO, our approach -- replacing the Gaussian policy with a diffusion model and residual term -- achieves a 10.35% average return improvement, with gains up to 13.57% in wheeled-biped locomotion tasks. These results highlight the benefits of cross-embodiment data and composite generative architectures in learning robust, generalized locomotion skills.