Grabette: an open system to record robot-manipulation data
Quick Answer
Grabette is an open-source system for recording robot-manipulation data using a handheld gripper, enabling users to create robot-ready datasets without needing a robot.
Quick Take
It simplifies data collection to two steps: record and process, fostering a collaborative dataset for robot learning.
Key Points
- Grabette captures manipulation tasks using a handheld gripper and two cameras for tracking.
- The system costs approximately €490 for Grabette and €120 for its robotic counterpart, Gripette.
- Data is processed in-browser, allowing quick creation of training-ready datasets.
- Open-source components enable easy assembly from standard parts, promoting accessibility.
- Designed to be robot-agnostic, the data format supports various robotic arms.
DeepSignal Analysis
What happened
Grabette is an open-source system designed to facilitate the recording of robot-manipulation data using a handheld gripper. It allows users to create datasets without needing a robot, simplifying the process to two steps: record and process. The system aims to foster a collaborative dataset for robot learning.
Key evidence
- Grabette is inspired by the Universal Manipulation Interface (UMI) from Stanford, which demonstrated the feasibility of recording manipulation tasks using a handheld gripper and cameras.
- The hardware for Grabette includes a handheld gripper with two cameras and costs approximately 490€, while its robotic counterpart, Gripette, costs around 120€.
- The system allows users to record manipulation tasks and convert them into robot-ready datasets, which can then be uploaded to the Hugging Face Hub for further use.
Why it matters
The development of Grabette addresses a significant challenge in robot learning: the scarcity of diverse, real-world manipulation data. By enabling anyone to record and share manipulation tasks, it democratizes access to valuable datasets, potentially accelerating advancements in robot learning. This collaborative approach could lead to richer datasets that no single lab could compile alone, fostering innovation in the field.
Source Excerpt
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from Hugging Face
See more →
From Hugging Face to Amazon SageMaker Studio in one click
Hugging Face has launched a deep-link integration with Amazon SageMaker Studio, allowing developers to seamlessly transition from model discovery to deployment with a single click. This integration streamlines the process by pre-configuring permissions and providing GPU quota visibility, significantly reducing the time from model selection to experimentation.
