PX4

Last reviewed: March 2026 px4.io ↗

What It Is

PX4 is a professional-grade, open-source autopilot system for drones, UAVs, and other unmanned vehicles. It is a project of the Dronecode Foundation, which operates under the Linux Foundation — the same organization that stewards Linux, Kubernetes, and other critical open-source infrastructure. PX4 is completely free, open source (BSD-3-Clause license), and supported by a global community of developers, researchers, and commercial companies.

PX4 runs on dedicated flight controller hardware — primarily the Pixhawk family of boards (Pixhawk 6X, 6C, Cube Orange+, Holybro boards). It handles everything an autopilot must do: sensor fusion (accelerometers, gyroscopes, magnetometers, barometers, GPS), state estimation (Extended Kalman Filter), flight control (PID and cascaded controllers), mission execution, failsafe management, and communication with ground stations via MAVLink protocol.

What distinguishes PX4 from ArduPilot (the other major open-source autopilot) is its architecture. PX4 uses a publish-subscribe messaging system (uORB) and runs on NuttX RTOS, making it modular and suitable for commercial certification efforts. Companies like Auterion, senseFly (AgEagle), and Wingtra build commercial products on PX4. It is the autopilot behind many of the most advanced research and commercial drone platforms in the world.

Aerospace Applications

Research UAV Platforms

University research labs worldwide use PX4 as the autopilot for custom research drones. Its modular architecture allows researchers to swap in custom control algorithms, sensor drivers, and mission logic without modifying the core codebase. Labs at ETH Zurich, TU Delft, Georgia Tech, and MIT run PX4 on platforms ranging from 250mm micro-quads to fixed-wing survey aircraft.

Commercial Drone Products

Auterion (founded by PX4's original creator, Lorenz Meier) builds its enterprise drone operating system on PX4. Wingtra uses PX4 for its WingtraOne VTOL survey drone, which dominates the professional mapping market. senseFly (now AgEagle) used PX4 derivatives for its eBee agricultural mapping drones. These are production aircraft flying daily commercial missions.

Autonomous Flight Research

PX4 integrates with ROS 2 (Robot Operating System 2) through its XRCE-DDS bridge, enabling advanced autonomy research: GPS-denied navigation, visual SLAM, multi-vehicle coordination, and AI-based decision making. This PX4 + ROS 2 stack is the standard platform for academic papers in autonomous aerial systems.

Hardware-in-the-Loop Simulation

PX4 includes built-in support for Gazebo, jMAVSim, and AirSim simulators. You can run the real PX4 firmware in a software-in-the-loop (SITL) simulation on your laptop — no hardware required. This means you can develop, test, and debug flight control code without risking a physical aircraft. SITL simulation is how professional drone developers work before flight testing.

Urban Air Mobility (UAM)

eVTOL companies designing air taxis and cargo drones evaluate PX4 as a starting point for their flight control systems. While production eVTOL aircraft require certified firmware, PX4's architecture — particularly its EKF2 state estimator and control allocation framework — informs the design of commercial systems.

Getting Started

High School

Start with simulation — no hardware needed. PX4 runs entirely in software on your computer.

  • Install PX4 SITL (Software-In-The-Loop) following the PX4 Developer Guide
  • Fly a simulated quadcopter in Gazebo using QGroundControl (free ground station software)
  • Create and execute waypoint missions in the simulator
  • Learn the basics of MAVLink protocol — how the autopilot communicates with ground stations
  • If you have the budget (~$200–400), build a small quadcopter with a Pixhawk 6C Mini and fly it with PX4

Undergraduate

Move from flying to developing. This is where PX4 becomes a serious engineering tool.

  • Study the PX4 architecture: uORB messaging, EKF2 state estimation, commander module, navigator
  • Write a custom PX4 module in C++ — even something simple like a data logger or a custom failsafe
  • Integrate PX4 with ROS 2 using the XRCE-DDS bridge for offboard control
  • Build a custom airframe (fixed-wing, VTOL, hexacopter) and configure PX4 for it
  • Use MAVSDK (Python or C++) to write autonomous mission scripts
  • Contribute to the PX4 open-source project on GitHub — start with documentation or small bug fixes

Advanced / Graduate

Research-level work: custom control algorithms, novel state estimation, multi-vehicle systems.

  • Implement custom flight control algorithms (adaptive control, model predictive control, neural network controllers)
  • Develop visual-inertial odometry (VIO) pipelines for GPS-denied navigation
  • Build multi-vehicle swarm coordination using PX4 + ROS 2
  • Conduct formal verification of flight control software — critical for certification
  • Publish at ICRA, IROS, or AIAA conferences — PX4-based research is widely accepted

Career Connection

RoleHow PX4 Is UsedTypical EmployersSalary Range
Flight Controls EngineerDesign, tune, and validate autopilot control laws for UAVs and eVTOL aircraftAuterion, Wingtra, Joby Aviation, Archer Aviation$110K–$170K
Robotics Software EngineerDevelop autonomy stacks (PX4 + ROS 2) for commercial and defense dronesShield AI, Skydio, Anduril, Zipline$130K–$200K
GNC Engineer (Guidance, Navigation, Control)State estimation, sensor fusion, trajectory planning for unmanned systemsNorthrop Grumman, General Atomics, L3Harris, NASA JPL$100K–$160K
Drone Systems EngineerFull-stack integration — autopilot, payloads, communications, ground stationAuterion, senseFly/AgEagle, DJI Enterprise partners$90K–$140K
Test & Validation EngineerHardware-in-the-loop testing, flight test campaigns, certification evidenceJoby Aviation, Wisk, Archer Aviation, Reliable Robotics$100K–$155K
Embedded Systems EngineerFirmware development on NuttX RTOS, sensor driver integration, hardware bring-upAuterion, Holybro, CubePilot, Hex/ProfiCNC$105K–$165K
Verified March 2026