ArduPilot — Open-Source Autopilot for Drones, Rovers and Submarines

Introduction

ArduPilot is the most trusted open-source autopilot platform, powering everything from hobbyist quadcopters to commercial survey drones and research submarines. Its codebase has been in active development since 2010 and runs on a wide range of flight controllers.

What ArduPilot Does

• Controls multirotor, fixed-wing, helicopter, rover, boat, and submarine vehicles
• Executes autonomous missions with GPS waypoints, geofences, and return-to-launch
• Streams real-time telemetry to ground control stations over radio, Wi-Fi, or cellular links
• Supports companion computers for onboard image processing and obstacle avoidance
• Integrates with MAVLink protocol for interoperability with ground stations and payloads

Architecture Overview

ArduPilot runs a real-time control loop on STM32-based flight controllers (Pixhawk family and others). A hardware abstraction layer (AP_HAL) separates vehicle logic from board-specific drivers. The SITL (Software in the Loop) simulator lets developers test on a desktop without hardware. Vehicle types share common libraries for EKF navigation, PID control, and MAVLink communication.

Self-Hosting & Configuration

• Flash firmware via Mission Planner (Windows), QGroundControl (cross-platform), or waf CLI build
• SITL simulator runs entirely on Linux or macOS for testing without hardware
• Parameters are tuned through ground station software or MAVLink commands
• Supports Lua scripting for custom on-board logic and automation
• Log files (.bin) can be analyzed with MAVExplorer or the web-based UAV Log Viewer

Key Features

• Supports over 100 flight controller boards from various manufacturers
• Extended Kalman Filter (EKF3) fuses GPS, IMU, barometer, and compass data for accurate positioning
• Autonomous mission execution with conditional waypoints, do-commands, and spline paths
• Geofencing and failsafe actions (return to launch, land, continue mission)
• SITL and HITL simulation for safe development and regression testing

Comparison with Similar Tools

• PX4 — alternative open-source autopilot with Dronecode backing; ArduPilot supports more vehicle types and has a larger hobbyist community
• Betaflight — focused on FPV racing quads with low-latency control; ArduPilot targets autonomous missions and GPS navigation
• DJI SDK — proprietary; ArduPilot is fully open source and hardware-agnostic
• iNav — navigation firmware for fixed-wing and multirotor; ArduPilot offers richer mission planning and companion computer integration
• Mission Planner / QGroundControl — ground station software that pairs with ArduPilot firmware, not competitors

FAQ

Q: What hardware do I need to get started? A: A Pixhawk-compatible flight controller (starting around $50), a GPS module, and a radio telemetry link. For testing, SITL requires only a Linux or macOS machine.

Q: Can ArduPilot be used commercially? A: Yes. ArduPilot is licensed under GPLv3. Many commercial drone companies build products on ArduPilot firmware.

Q: How do I contribute code? A: Fork the repository, develop on a feature branch, test in SITL, and submit a pull request. The project has detailed developer documentation and an active Discord community.

Q: Does ArduPilot support obstacle avoidance? A: Yes, via companion computers running ROS or simple rangefinder-based avoidance. The BendyRuler and Dijkstra path planners are built in.

Sources

• https://github.com/ArduPilot/ardupilot
• https://ardupilot.org/dev/