Dobb·E

About Dobb·E

Dobb·E is an open-source framework designed for teaching robots household tasks efficiently. It utilizes a demonstration collection tool, The Stick, to gather data and train robots through imitation learning. With impressive success rates, Dobb·E is aimed at enhancing home robotics for everyday users.

Dobb·E offers access to an open-source software stack and models, which can be utilized freely. Users can tap into valuable datasets and innovative tools, allowing for effective robotic task learning. The platform encourages experimentation and collaboration, making robotics accessible and user-friendly.

The user interface of Dobb·E is designed for simplicity and efficiency, ensuring a smooth experience for users. With intuitive navigation and quick access to tools and resources, Dobb·E allows users to focus on teaching robots without overwhelming complexity. It promotes engagement and ease of use.

How Dobb·E works

To interact with Dobb·E, users begin by onboarding and familiarizing themselves with The Stick tool for demonstration collection. Users demonstrate tasks for five minutes to gather data, which Dobb·E then uses to train a robot to replicate the task. With just 20 minutes, users can effectively teach robots household tasks, significantly easing the learning process.

Key Features for Dobb·E

Rapid Learning Capability

One of Dobb·E's standout features is its rapid learning capability, enabling robots to learn new household tasks in just 20 minutes. This innovative approach makes Dobb·E a valuable tool for owners seeking efficient robotic assistance in their homes, revolutionizing home robotics.

Open-Source Accessibility

Dobb·E's open-source nature allows users access to software, models, and datasets at no cost. This feature encourages community collaboration and individual exploration in robotic learning, positioning Dobb·E as a pivotal resource for innovators and hobbyists in home robotics.

The Stick Demonstration Tool

The Stick is a unique tool developed by Dobb·E for collecting robot training demonstrations. Made from affordable materials, it simplifies the process, making it easy for anyone to teach robots multiple tasks efficiently, thus expanding the reach and usability of home robotics.