COMPASS: Cross-embodiment Mobility Policy via Residual RL and Skill Synthesis

TL;DR

COMPASS is a unified framework that enables scalable cross-embodiment mobility for robots, using minimal demonstrations and residual reinforcement learning to adapt policies across diverse robot designs with high success rates.

Contribution

The paper introduces COMPASS, a novel approach combining imitation learning, residual RL, and skill synthesis to enable cross-embodiment mobility with minimal data.

Findings

Success rate approximately 5X higher than initial IL policy on unseen embodiments.

Effective zero-shot sim-to-real transfer demonstrated.

Scales across diverse robot platforms with robust generalization.

Abstract

As robots are increasingly deployed in diverse application domains, enabling robust mobility across different embodiments has become a critical challenge. Classical mobility stacks, though effective on specific platforms, require extensive per-robot tuning and do not scale easily to new embodiments. Learning-based approaches, such as imitation learning (IL), offer alternatives, but face significant limitations on the need for high-quality demonstrations for each embodiment. To address these challenges, we introduce COMPASS, a unified framework that enables scalable cross-embodiment mobility using expert demonstrations from only a single embodiment. We first pre-train a mobility policy on a single robot using IL, combining a world model with a policy model. We then apply residual reinforcement learning (RL) to efficiently adapt this policy to diverse embodiments through corrective refinements. Finally, we distill specialist policies into a single generalist policy conditioned on an embodiment embedding vector. This design significantly reduces the burden of collecting data while enabling robust generalization across a wide range of robot designs. Our experiments demonstrate that COMPASS scales effectively across diverse robot platforms while maintaining adaptability to various environment configurations, achieving a generalist policy with a success rate approximately 5X higher than the pre-trained IL policy on unseen embodiments, and further demonstrates zero-shot sim-to-real transfer.