Hugging Face library for real-world robotics

Overview

This project began as research project for automated the recycle process in warehouse. Hugging Face’s LeRobot project is an open source project on machine learning models for real-world robotics, especially within domains like imitation and reinforcement learning. It is a wrapper on top of Stanford paper ALOHA. Stanford ALOHA refers to a series of robotics systems, with the primary goal of creating low-cost, open-source hardware for robotics research. ALOHA initially started as a bimanual teleoperation system but has since evolved into Mobile ALOHA, which combines mobility with advanced whole-body manipulation capabilities. Mobile ALOHA is designed to perform complex real-world tasks like cooking, cleaning, and navigating environments, making it useful for research in areas such as imitation learning and human-robot interaction.

The deliverable from this project is to pick and place recycled phones.

Implementation

The robot arms are built based on koch 1.1 version.

After the data collection and training the final result is