Polybot: Training One Policy Across Robots While Embracing Variability

TL;DR

Polybot introduces a unified policy training approach that effectively transfers manipulation skills across diverse robotic platforms by aligning observation spaces and internal representations, improving success rates and efficiency.

Contribution

The paper presents a novel framework for training a single policy applicable to multiple robots by combining observation/action space alignment and contrastive learning of internal representations.

Findings

Significant success rate improvements across different robots.

Enhanced sample efficiency with new task data.

Effective domain shift bridging through contrastive learning.

Abstract

Reusing large datasets is crucial to scale vision-based robotic manipulators to everyday scenarios due to the high cost of collecting robotic datasets. However, robotic platforms possess varying control schemes, camera viewpoints, kinematic configurations, and end-effector morphologies, posing significant challenges when transferring manipulation skills from one platform to another. To tackle this problem, we propose a set of key design decisions to train a single policy for deployment on multiple robotic platforms. Our framework first aligns the observation and action spaces of our policy across embodiments via utilizing wrist cameras and a unified, but modular codebase. To bridge the remaining domain shift, we align our policy's internal representations across embodiments through contrastive learning. We evaluate our method on a dataset collected over 60 hours spanning 6 tasks and 3 robots with varying joint configurations and sizes: the WidowX 250S, the Franka Emika Panda, and the Sawyer. Our results demonstrate significant improvements in success rate and sample efficiency for our policy when using new task data collected on a different robot, validating our proposed design decisions. More details and videos can be found on our anonymized project website: https://sites.google.com/view/polybot-multirobot