CEI: A Unified Interface for Cross-Embodiment Visuomotor Policy Learning in 3D Space

TL;DR

This paper introduces CEI, a framework enabling the transfer of visuomotor policies across different robot embodiments by aligning trajectories based on functional similarity, demonstrated through extensive simulation and real-world experiments.

Contribution

The paper presents a novel cross-embodiment learning framework that aligns and transfers policies between diverse robot morphologies using gradient-based optimization of functional similarity.

Findings

Successfully transferred policies across 16 robot embodiments in simulation.

Achieved an 82.4% average transfer ratio in real-world tasks.

Extended CEI with spatial generalization and multimodal motion generation.

Abstract

Robotic foundation models trained on large-scale manipulation datasets have shown promise in learning generalist policies, but they often overfit to specific viewpoints, robot arms, and especially parallel-jaw grippers due to dataset biases. To address this limitation, we propose Cross-Embodiment Interface (\CEI), a framework for cross-embodiment learning that enables the transfer of demonstrations across different robot arm and end-effector morphologies. \CEI introduces the concept of \textit{functional similarity}, which is quantified using Directional Chamfer Distance. Then it aligns robot trajectories through gradient-based optimization, followed by synthesizing observations and actions for unseen robot arms and end-effectors. In experiments, \CEI transfers data and policies from a Franka Panda robot to \textbf{16} different embodiments across \textbf{3} tasks in simulation, and supports bidirectional transfer between a UR5+AG95 gripper robot and a UR5+Xhand robot across \textbf{6} real-world tasks, achieving an average transfer ratio of 82.4\%. Finally, we demonstrate that \CEI can also be extended with spatial generalization and multimodal motion generation capabilities using our proposed techniques. Project website: https://cross-embodiment-interface.github.io/