This repository provides the codebase for TidyBot, a personalized robot assistant designed for household cleanup tasks. It leverages Large Language Models (LLMs) to infer and generalize user preferences for object placement, enabling robots to adapt quickly to individual needs and achieve high accuracy in tidying rooms. The project is targeted at researchers and developers in robotics and AI interested in human-robot interaction and personalized assistance.

How It Works

TidyBot combines language-based planning and perception with the few-shot summarization capabilities of LLMs. It learns user preferences from minimal examples, inferring generalized rules for where objects should be placed. This approach allows for rapid adaptation to individual users, overcoming the challenge of diverse personal tastes in object organization.

Quick Start & Requirements

• Installation: Recommended via Conda environments. Requires separate tidybot and vild environments.
• Prerequisites: Tested on Ubuntu 20.04.6 LTS. Requires GPU workstation for server code. Specific hardware for mobile manipulator (TidyBot) is implied.
• Setup: Detailed setup instructions are available in the server and robot directories.
• Resources: Official project page: https://tidybot.github.io/

Highlighted Details

• Achieves 91.2% accuracy on unseen objects in benchmark datasets.
• Successfully tidies 85.0% of objects in real-world test scenarios with the TidyBot mobile manipulator.
• Offers both teleoperation mode for manual control and fully autonomous mode for end-to-end tidying.
• Includes a benchmark dataset for evaluating personalization capabilities.

Maintenance & Community

The project is associated with the Autonomous Robots (AuRo) lab and the authors from the 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). No specific community channels (e.g., Discord, Slack) are mentioned in the README.

Licensing & Compatibility

The README does not explicitly state a license. It is presented as a code release accompanying a research publication, implying potential research-use restrictions. Commercial use or linking with closed-source projects would require clarification.

Limitations & Caveats

The setup is tested on Ubuntu 20.04.6 LTS, and specific hardware configurations for the robot and server are implied but not fully detailed. Troubleshooting sections address potential issues with mobile base and arm accuracy, camera errors, and server port conflicts.