PX4

PX4
Name	PX4
Developer	Dronecode Foundation, Auterion, Lorenz Meier
Initial release	2011
Programming language	C++, Python, NuttX
Operating system	NuttX, Linux
License	BSD-3-Clause

PX4 is an open-source flight control software stack for unmanned aerial vehicles and other robotic vehicles. It provides guidance, navigation, and control algorithms, sensor drivers, middleware, and toolchains used by manufacturers, research institutions, and hobbyists. PX4 integrates with autopilot hardware, ground control stations, and simulation environments to support development, testing, and deployment of aerial and terrestrial systems.

Overview

PX4 is maintained within an ecosystem that includes the Dronecode Foundation, Auterion, academic labs, and commercial vendors. The project supplies a modular autopilot architecture combining real-time kernels such as NuttX and user-space components running on Linux for companion computers. PX4 supports a range of vehicle classes and emphasizes portability, safety, and extensibility to interoperate with standards from organizations like RTCA and collaborations with industry partners such as Intel Corporation and Qualcomm. The stack is commonly paired with ground control software such as QGroundControl and integrates with simulation tools including Gazebo (software), JSBSim, and FlightGear.

History and Development

Development began in the early 2010s by contributors from the Swiss Federal Institute of Technology in Zurich and independent developers; key contributors include individuals who later co-founded companies such as Auterion. The project evolved through community governance under the Dronecode Foundation, an umbrella created by the Linux Foundation to foster open-source unmanned systems. PX4 has undergone major rewrites and modularization milestones, shifting from monolithic firmware toward middleware-driven designs compatible with companion computers like NVIDIA Jetson and Raspberry Pi. Over time PX4 incorporated features driven by research from institutions such as ETH Zurich and industry partners including Parrot SA and EHang.

Architecture and Components

PX4’s architecture separates low-level flight control, middleware, and high-level autonomy. The real-time flight stack runs on autopilot hardware using the NuttX real-time operating system, while higher-level modules run on Linux companion computers employing middleware like ROS and communication bridges. Core components include sensor drivers for devices by Bosch Sensortec, InvenSense, and STMicroelectronics; state estimation modules leveraging algorithms from the Kalman filter family; a mixer system mapping actuator outputs; and a command/control API compatible with MAVLink. PX4 utilizes a publish/subscribe messaging layer and firmware modules for attitude control, position control, and mission execution. Developers interact with the stack via build systems such as CMake and toolchains targeting processors from STM32 families.

Supported Vehicles and Hardware

PX4 supports multicopters, fixed-wing, VTOL (vertical takeoff and landing), and ground robots. Officially supported autopilot hardware includes flight controllers based on STM32F4, STM32F7, and newer STM32H7 microcontrollers produced by companies like Pixhawk Project partners. Reference hardware platforms include flight controllers from vendors such as Holybro, CUAV, Hex Technology, and 3D Robotics legacy models. PX4 also interoperates with companion computers like NVIDIA Jetson Xavier, Raspberry Pi, and Intel NUC devices, and peripherals including GPS modules from u-blox, airspeed sensors such as PicoVGA, and radio links like SiK and Ubiquiti Networks solutions.

Flight Control Features and Modes

PX4 implements stabilization, altitude hold, position hold, and autonomous navigation including missions, geofencing, and return-to-launch. Advanced control capabilities include sensor fusion-based state estimation, model-predictive control (MPC) modules, and trajectory generation suited for VTOL transitions and fixed-wing flight envelopes. Flight modes interoperate with ground control systems supporting commands from MAVLink-compatible controllers and payload control for gimbals such as products by Gremsy and DJI-compatible peripherals. PX4 includes fail-safe behaviors, battery management integrations, and dynamic parameter tuning used by research groups at institutions like MIT and Stanford University studying aerial autonomy.

Software Ecosystem and Tools

The PX4 software ecosystem comprises build and CI infrastructure, simulation, developer tools, and ground control applications. Commonly used tools include QGroundControl for mission planning, the PX4 toolchain for cross-compilation, and simulation integrations with Gazebo (software) and RotorS. Continuous integration and testing engage platforms such as GitHub Actions and vendor-sponsored testbeds. The ecosystem fosters plugins and SDKs enabling integration with cloud services from providers like Amazon Web Services and Microsoft Azure for beyond-visual-line-of-sight (BVLOS) operations, and SDKs used by companies including Auterion and third-party integrators.

Safety, Certification, and Compliance

PX4’s development emphasizes traceability, testing, and safety standards compliance to facilitate certification work with aviation authorities and industry bodies such as RTCA and national civil aviation authorities. The project provides documentation, unit tests, hardware-in-the-loop (HIL) test frameworks, and model-based verification workflows employed by manufacturers seeking compliance with standards like DO-178C and DO-254 in parts of their product stacks. Collaborations with commercial integrators aim to bridge open-source innovation and regulatory certification pathways used by enterprise customers and research institutions.

Category:Open-source robotics