RLRoverLab

Introduction

Welcome to RLRoverLab! This project implements a suite of Reinforcement Learning (RL) agents using Isaac Sim and Isaac Lab. Our suite currently supports a variety of tasks within navigation and manipulation, with ongoing efforts to expand our offerings.

Features

• Navigation and Manipulation Tasks: Implementations of RL agents designed for navigation and manipulation tasks, we are working on integrating more tasks.
• Isaac Sim and ORBIT Integration: Utilizes the advanced simulation environments of Isaac Sim and the ORBIT framework for realistic task scenarios.
• Expandable Framework: Architecture designed for easy extension with new tasks and functionalities.

Getting Started

To get started with RLRoverLab, please refer to our Installation Guide available in the project's wiki. The guide provides comprehensive steps for setting up the suite using Docker as well as instructions for native installation.

Quick Links

• Installation Guide
• Examples and Tutorials
• Adding Custom Robots and Tasks

Train agents in parralel

Parralel.mp4

Parralel_video4_1440p.mp4

Video of trained RL-agent

RL-agent.mp4

Support

If you have questions, suggestions, feel free to contact us.

• Contact Us: For direct inquiries, reach out to Anton Bjørndahl Mortensen at antonbm2008@gmail.com.