Mobile ALOHA

What if robots could learn complex, bimanual mobile tasks just by watching us?

RecommendedMobile ALOHA offers a significant step forward in low-cost, whole-body teleoperation for bimanual mobile manipulation, providing valuable tools and datasets for the robotics research community.

Mobile ALOHA is a system designed for learning bimanual mobile manipulation tasks through imitation learning. It utilizes a low-cost, whole-body teleoperation system to collect human demonstrations, which are then used to train robots for complex mobile manipulation tasks. The system significantly boosts success rates by co-training with existing static ALOHA datasets.

Visit Mobile ALOHA

Visit Mobile ALOHA

Key Features:

Low-cost whole-body teleoperation system
Bimanual mobile manipulation capabilities
Supervised behavior cloning for task learning
Co-training with existing static ALOHA datasets
Hardware and ML code available on GitHub

Pros

Robotics researchers needing a system for bimanual mobile manipulation data collection
Developers looking to implement imitation learning for complex robot tasks
Academics interested in low-cost teleoperation systems for robotics research

Cons

Requires significant technical expertise to set up and operate
Primarily a research project, not a commercial product
Performance is dependent on the quality and quantity of human demonstrations

Pricing

unknown

Categories:

Research & Science Code & Development

Tags:

robotics imitation learning teleoperation bimanual manipulation mobile robotics open source

Share:

Quick Decision

Try if: You are a robotics researcher or developer interested in advancing bimanual mobile manipulation through imitation learning and open-source hardware/software.

Skip if: You are looking for a commercial, off-the-shelf robotic solution or lack the technical expertise to work with research-level robotics projects.

Not for: Individuals seeking a ready-to-use commercial robot solution; Users without a background in robotics or machine learning

Trust Signals

Founded
2024
Team Size
small
Funding
Boston Dynamics AI Institute and ONR grant N00014-21-1-2685

Tech Details

Platforms

hardware

Open Source

Yes

Support

Company

Name
Stanford University (IRIS Lab & REAL Lab)
Location
Stanford, CA, USA

Mobile ALOHA is a research project developed by researchers at Stanford University, focusing on advancing robotics through imitation learning and teleoperation.

FAQ

What is Mobile ALOHA?

Mobile ALOHA is a research project that introduces a low-cost, whole-body teleoperation system for collecting data to train robots in bimanual mobile manipulation tasks. It uses imitation learning to enable robots to perform complex actions.

Is Mobile ALOHA open source?

Yes, the hardware code and machine learning code for Mobile ALOHA are available on GitHub, indicating it is an open-source project for the robotics community.

What kind of tasks can Mobile ALOHA perform?

Mobile ALOHA can autonomously complete complex mobile manipulation tasks such as sauteing and serving shrimp, opening cabinets, calling and entering an elevator, and rinsing a pan, with high success rates after training.

Use Cases

Robotics researchers needing a system for bimanual mobile manipulation data collection
Developers looking to implement imitation learning for complex robot tasks
Academics interested in low-cost teleoperation systems for robotics research