Duet: Dual-Robot Understanding via Efficient Teaching

2026-06-18

AI Summary

This work addresses the challenge of inefficient data collection that hinders dual-robot collaborative learning in complex tasks. The authors propose an efficient data-driven framework that leverages a unified synchronous VR teleoperation system coupled with a human–human collaboration behavior tracking pipeline to collect high-quality demonstration data from heterogeneous dual robots (Unitree G1 and Dexmate Vega1). By integrating a motion-chunking Transformer architecture, the method pretrains collaborative policies using human prior knowledge and fine-tunes them with only a small number of real-world trajectories. Evaluated on four collaborative tasks, the approach significantly outperforms baselines trained solely on robot-collected data, achieving a 5.4× improvement in data collection efficiency while requiring substantially less human effort.

Abstract

Dual-robot collaboration enables tasks that exceed the reach and payload of a single robot, such as collaboratively transporting objects across environments and executing coordinated handovers. Data acquisition is the primary bottleneck for training these systems. To this end, we introduce DUET, a dual-robot learning framework for mobile manipulation. For efficient data collection, we create a unified dual-embodiment synchronized VR-based teleoperation system for in-domain heterogeneous robot data collection. We further develop a complementary tracking pipeline that records human-human coordination and collaborative mobile manipulation priors. To allow efficient learning, we introduce an Action Chunking Transformer based architecture that first pretrains collaborative policies on efficient human-human demonstrations, before finetuning them on a minimal set of real-robot teleoperation trajectories. We develop a benchmark of four collaborative tasks to evaluate our framework using a Unitree G1 humanoid and a Dexmate Vega1 mobile manipulator. The results demonstrate that harnessing human priors not only yields superior task performance compared to baselines trained only on robot data, but also reduces the total human effort required for data collection. Our human data collection pipeline achieves 5.4x acceleration on average from teleoperation, but we perform equally or better than robot-only data trained policies across all tasks. Our project page is available at https://zhaoy37.github.io/Duet/.

Problem

• dual-robot collaboration
• data acquisition
• mobile manipulation
• human-robot coordination
• teleoperation

Innovation

• dual-robot collaboration
• VR-based teleoperation
• human-human demonstration
• Action Chunking Transformer
• mobile manipulation