Skip to content

About AirStack

AirStack is a comprehensive, modular autonomy stack for aerial robotics developed by the AirLab at Carnegie Mellon University's Robotics Institute.

Project Mission

AirStack aims to provide a complete, field-tested autonomy framework that enables researchers and developers to:

  • Deploy autonomous aerial systems in simulation and real-world environments
  • Rapidly prototype new algorithms and approaches
  • Integrate custom modules through standardized interfaces
  • Scale from single to multi-robot operations
  • Transition from simulation to hardware with minimal changes

Team

AirStack is developed and maintained by the AirLab at CMU's Robotics Institute, with contributions from students, researchers, and collaborators.

Core Contributors: Andrew Jong, John Keller, John Liu, et al. See GitHub Contributors for more.

Principal Investigator: Sebastian Scherer

Research

AirStack builds upon years of research in autonomous aerial systems:

  • Obstacle avoidance in cluttered environments
  • Multi-robot coordination and exploration
  • Vision-based state estimation
  • Field robotics for search and rescue

Related Publications: See AirLab Publications

Status

Alpha Release

AirStack is currently in ALPHA status and intended for internal AirLab use.

  • API and functionality may change
  • Requires AirLab account for access to resources
  • Not yet ready for public release

We welcome feedback from early users to improve the system for eventual public release.

Architecture

AirStack features:

  • Modular Design: Swap components easily through ROS 2 interfaces
  • Layered Autonomy: Interface, Sensors, Perception, Local Planning, Global Planning, Behavior
  • Docker-Based: Consistent development and deployment environments
  • Simulation Support: NVIDIA Isaac Sim with Pegasus extension
  • Multi-Robot: Independent robots with coordinated operation

See: System Architecture

Supported Platforms

Simulation

  • NVIDIA Isaac Sim (primary)
  • Gazebo (planned)

Hardware

  • NVIDIA Jetson (Orin, Xavier NX, TX2)
  • ModalAI VOXL (VOXL 2, VOXL Flight)
  • x86-64 desktop/laptop (development)

Software

  • ROS 2 Jazzy
  • Ubuntu 22.04/24.04
  • Docker with NVIDIA Container Toolkit

License

License Pending

License to be determined. Likely Apache 2.0 or MIT for open source components.

For now, access requires AirLab account and approval.

FAQ

General Questions

Who can use AirStack?

Currently, AirStack is in alpha and limited to AirLab members and collaborators. We plan to release it publicly once it reaches a stable state.

What hardware do I need?

For development: Ubuntu 22.04 machine with NVIDIA GPU (RTX 3070+ recommended), 16GB+ RAM, and 100GB+ free storage.

For deployment: NVIDIA Jetson or ModalAI VOXL onboard computer.

Do I need a real drone to use AirStack?

No! You can develop and test entirely in Isaac Sim simulation before deploying to hardware.

Technical Questions

What's the difference between local and global planning?

Local planning handles obstacle avoidance and trajectory generation at high frequency (10+ Hz) for immediate safety.

Global planning generates waypoint paths at lower frequency (1 Hz) for exploration and goal navigation.

See: Autonomy Modes Tutorial

Can I use my own planner/controller?

Yes! AirStack is designed for easy module swapping. Create a ROS 2 package with the correct interface, integrate it into the bringup layer, and you're ready.

See: Integration Checklist

How do I add a new sensor?

Add sensor integration to the sensors layer, ensuring data is published to standard topics (with correct frame IDs and timestamps).

See: Sensors Overview

Does AirStack support multi-robot missions?

Yes! Each robot gets a unique ROS_DOMAIN_ID and namespace. All run independently while coordinating through the Ground Control Station.

See: Multi-Robot Simulation Tutorial

Can I run AirStack without Docker?

While technically possible, we strongly recommend using Docker. It ensures consistent environments, simplifies dependency management, and matches how we test and deploy.

Development Questions

How do I debug a module?

Build with debug symbols, use ROS 2 tools (ros2 topic echo, ros2 node info), and leverage the debug-module skill from the AI Agent Guide.

See: Development Environment

Where should I add my custom module?

Place it in the appropriate layer under robot/ros_ws/src/<layer>/<type>/<your_module>/. For example, a new planner goes in robot/ros_ws/src/local/planners/my_planner/.

See: AI Agent Guide

How do I test in simulation?

Launch AirStack with Isaac Sim, run your module, and verify behavior. Record ROS bags for offline analysis.

See: Testing Guide

Contact and Support

Acknowledgments

AirStack development is supported by:

  • Carnegie Mellon University Robotics Institute
  • AirLab research sponsors and collaborators
  • Open source ROS 2 community
  • NVIDIA Isaac Sim and Pegasus simulator teams

Special thanks to all contributors and early adopters who provide valuable feedback.

Contributing

We welcome contributions from AirLab members and collaborators!

See: Contributing Guide

Areas where we especially need help:

  • Documentation improvements
  • Bug reports and fixes
  • New module implementations
  • Testing and validation
  • Hardware platform support

Roadmap

Planned features and improvements:

  • Public Release: Stable API and public documentation
  • Additional Simulators: Gazebo support
  • More Hardware Platforms: Broader Jetson family, additional flight controllers
  • Enhanced Multi-Robot: Improved coordination and communication
  • Expanded Tutorials: More examples and use cases
  • Performance Optimization: Faster planning and perception
  • Better Testing: Comprehensive test suite and CI/CD

Check GitHub Issues for current priorities.