Pegasus Simulator

Overview

Pegasus Simulator is a framework built on top of NVIDIA Omniverse and Isaac Sim. It is designed to provide an easy yet powerful way of simulating the dynamics of multirotors vehicles. It provides a simulation interface for PX4 integration as well as custom python control interface. At the moment, only multirotor vehicles are supported, with support for other vehicle topologies planned for future versions.

If you find Pegasus Simulator useful in your academic work, please cite the paper below. It is also available here.

@INPROCEEDINGS{10556959,
   author={Jacinto, Marcelo and Pinto, João and Patrikar, Jay and Keller, John and Cunha, Rita and Scherer, Sebastian and Pascoal, António},
   booktitle={2024 International Conference on Unmanned Aircraft Systems (ICUAS)},
   title={Pegasus Simulator: An Isaac Sim Framework for Multiple Aerial Vehicles Simulation},
   year={2024},
   volume={},
   number={},
   pages={917-922},
   keywords={Simulation;Robot sensing systems;Real-time systems;Sensor systems;Sensors;Task analysis},
   doi={10.1109/ICUAS60882.2024.10556959}}

Latest Updates

• 2024-11-01: Pegasus Simulator v4.2.0 is released for Isaac 4.2.0. This version is NOT compatible with older versions of Isaac Sim. This version includes a new experimental interface for Ardupilot integration, provided by open-source contributor Tomer Tiplitsky.

• 2024-08-02: Pegasus Simulator v4.1.0 is released for Isaac 4.1.0. This version is NOT compatible with older versions of Isaac Sim.

Guidance, Control and Navigation Project

In parallel to this project, the Pegasus (GNC) guidance, control, and navigation project serves as the foundation control code for performing real-world experiments for my Ph.D. More information can be found at this link: Pegasus GNC.

Developer Team

This simulation framework is an open-source effort, started by me, Marcelo Jacinto in January/2023. It is a tool that was created with the original purpose of serving my Ph.D. workplan for the next 4 years, which means that you can expect this repository to be mantained, hopefully at least until 2027.

• Project Founder

  • Marcelo Jacinto, under the supervision of Prof. Rita Cunha and Prof. Antonio Pascoal (IST/ISR-Lisbon)

• Architecture

  • Marcelo Jacinto

  • João Pinto

• Multirotor Dynamic Simulation and Control

  • Marcelo Jacinto

• Example Applications

  • Marcelo Jacinto

  • João Pinto

• Ardupilot Integration (Experimental)

  • Tomer Tiplitsky

Getting Started

• Installation
  • Installing NVIDIA Isaac Sim
  • Configuring the environment variables
  • Installing the Pegasus Simulator
  • Installing PX4-Autopilot
• Development
  • Code structure

Tutorials

• Run in Extension Mode (GUI)
• Create a Standalone Application
• Your First Simulation
• Create a Custom Controller
• Simulation with Moving People

Features

• Environments
  • List of provided simulation environments
  • Setting the Map Global Coordinates
• Vehicles
• PX4 Integration
• ArduPilot (Experimental)
  • Installing Ardupilot (Arducopter)
  • Running a Simulation with Ardupilot (GUI Mode)
  • Ardupilot Integration Architecture

Source API

• API Reference
  • Sensors
  • Graphical Sensors
  • Graphs
  • Dynamics
  • Thruster
  • Control Backends
  • Vehicle
  • People
  • Pegasus Interace
  • Default Params

References

• Contributing
• Known Issues
• Changelog
• Roadmap
• License
• Bibliography

Other Simulation Frameworks

In this section, we acknowledge the nobel work of those who came before us and inspired this work:

• [KH04] Gazebo simulator

• [FBAS16] RotorS simulation plugin for gazebo

• [MHP15] PX4-SITL simulation plugin for gazebo

• [SDLK18] Microsoft Airsim project for Unreal Engine

• [SNK+21] Flightmare simulator for Unity

• [Bab13] jMAVSim java simulator

“If I have seen further than others, it is by standing upon the shoulders of giants.”, Sir Isaac Newton

Project Sponsors

• Dynamics Systems and Ocean Robotics (DSOR) group of the Institute for Systems and Robotics (ISR), a research unit of the Laboratory of Robotics and Engineering Systems (LARSyS).

• Instituto Superior Técnico, Universidade de Lisboa

The work developed by Marcelo Jacinto and João Pinto was supported by Ph.D. grants funded by Fundação para as Ciências e Tecnologias (FCT).