Welcome to OmniDrones’s documentation!

OmniDrones is an open-source platform designed for reinforcement learning research on multi-rotor drone systems. Built on Nvidia Isaac Sim, OmniDrones features highly efficient and flxeible simulation that can be adopted for various research purposes.

The platform, as released with our paper, currently focus on end-to-end learning of agile controllers for drones. It offers a suite of benchmark tasks and algorithm baselines to provide preliminary results for subsequent works.

An overview of the platform is shown below:

If you use OmniDrones in your research, please cite the paper with the following bibtex entry:

@misc{xu2023omnidrones,
   title={OmniDrones: An Efficient and Flexible Platform for Reinforcement Learning in Drone Control},
   author={Botian Xu and Feng Gao and Chao Yu and Ruize Zhang and Yi Wu and Yu Wang},
   year={2023},
   eprint={2309.12825},
   archivePrefix={arXiv},
   primaryClass={cs.RO}
}

Getting Started

• Workstation Installation
  • Developer Guide: Working with VSCode
• Running the RL Scripts
• Troubleshooting and FAQs
  • Q: Can I use Python versions other than what is shipped with Omniverse?
  • GPU Driver-related Issues
  • Slow Startup

Demos

• Downwash between Multiple Drones
• Deployment on Crazyflie 2.1
  • Results
  • Quick Start
• Navigation in Cluttered Environment
  • Task Description
  • Visualization
  • Quick Start

Tutorials

• UAV Models and Usage
  • Using the UAV Models
  • Importing/Creating New UAVs
  • Customized Dynamics
  • Building New Configurations
• Environment
  • Specification
  • Interaction and Stepping Logic
  • Data Collection
  • Creating New Tasks
  • Environment Transforms
• Controllers and Usage
  • Using the Controllers
  • Adding Controllers
  • RL Performance

Tasks

• Single-Agent Tasks
• Multi-Agent Tasks
  • Quick Start

Misc.

• Development Roadmap

Indices and tables

• Index

• Module Index

• Search Page