DroneOS

Open-source framework for autonomous drone control, built on ROS 2 and PX4 Autopilot.

What It Does

DroneOS provides the software layer between you and PX4. Instead of publishing raw DDS topics, you get:

• C++ SDK (drone_core) — High-level drone control as ROS 2 services
• Python API (drone_control.py) — Remote control via rosbridge from anywhere
• CLI (droneos) — Quick operations and AI agent interface

Works in simulation (PX4 SITL + Gazebo) and on real hardware (Pixhawk + Raspberry Pi over 4G/VPN).

Quick Start (Simulation)

This gets you from zero to a drone flying in simulation on a single machine. One simulated drone, one computer, no physical drone hardware required.

1. Clone

# Clone PX4 Autopilot (in your workspace)
git clone https://github.com/PX4/PX4-Autopilot.git

# Clone DroneOS
git clone https://github.com/ortegarod/drone-os.git

2. Start PX4 Simulation

cd PX4-Autopilot
HEADLESS=1 make px4_sitl gz_x500

Wait for Ready for takeoff! in the terminal. This runs PX4 with Gazebo — a full drone simulator.

3. Start DroneOS

In a new terminal:

cd drone-os
docker compose -f docker/dev/docker-compose.dev.yml up -d --build drone_core micro_agent

This starts:

• drone_core — the SDK node that exposes drone control services
• micro_agent — the DDS bridge between PX4 and ROS 2

4. Verify

docker logs -f drone_core_node

You should see services registering: /drone1/arm, /drone1/takeoff, /drone1/set_position, etc. If you see those, you're good.

5. Fly

python3 drone_control.py --drone drone1 --set-offboard
python3 drone_control.py --drone drone1 --arm
python3 drone_control.py --drone drone1 --set-position 0 0 -50

The drone is now 50 meters up. The coordinate system is NED (North-East-Down), so negative Z = altitude. -50 means 50 meters above the takeoff point.

python3 drone_control.py --drone drone1 --set-position 80 -40 -50

Now it's flying 80m north and 40m west at 50m altitude.

python3 drone_control.py --drone drone1 --land

That's it. You just flew a drone.

6. Cleanup

docker compose -f docker/dev/docker-compose.dev.yml down

Then CTRL+C in the PX4 terminal.

Real Hardware Deployment

Same codebase, different Docker Compose file:

# On Raspberry Pi with Pixhawk connected
docker compose -f docker/prod/docker-compose.yml up -d --build drone_core micro_agent

Tech Stack

Layer	Technology
Flight Controller	PX4 Autopilot
Simulation	Gazebo
Middleware	ROS 2 Humble + Micro XRCE-DDS
SDK	C++ (drone_core)
Remote Control	Python (drone_control.py) via rosbridge
Deployment	Docker

Contributing

See CONTRIBUTING.md for how to get started.

Documentation & More Info

For thorough documentation, roadmaps, multi-drone setups, camera configurations, and the full system architecture, please visit the official website:

aerisrobotics.com/docs

License

Open source. See LICENSE for details.