Co2Gen produces robot task plans for teamwork with humans and robots
Co^2Gen: Collaboratively Generating Task Plans for Decentralized Multi-Robot and Human Collaboration
Robotics
Summary
Robots need to work together with other robots and humans to do complex tasks. The authors developed a way to create example plans for multiple robots working together without needing people to control them directly. Their method creates plans that can be used to teach robot models how to cooperate in a decentralized setting. They found that training robots with these automatically generated plans works just as well as using plans made by humans controlling robots.
What this means in practice
- •For robotics development teams: Generate multi-robot collaborative task plans automatically without human teleoperation to train decentralized robot collaboration models.
- •For industrial automation engineers: Implement decentralized robot teams that coordinate tasks efficiently using models fine-tuned on automatically created collaborative data.
Authors
Yi Loo, Siddhant Shrivastava, Minh Long Pham, Malika Meghjani
Abstract
Visual Language Action (VLA) models are designed to be generalist robot models that can be implemented across embodiments, tasks and environments. However, their ability to collaborate with other robots and humans remains underexplored. In this work, we propose a framework to automate the generation of multi-robot collaborative data for decentralized VLA training without requiring human teleoperation. Our framework first utilizes a decentralized collaborative planning approach to generate multi-robot collaborative robot plans in the form of an executable function. We then utilize the collaborative plan to produce collaborative demonstration data to finetune VLA models in a decentralized manner. We show that our approach produces demonstration data that when used to finetune VLA models on decentralized collaboration tasks matches the task performance of VLA models finetuned on human-teleoperated collaborative data.