QGroundControl (QGC) is a professional-grade, cross-platform ground control station for MAVLink-enabled unmanned aerial vehicles (UAVs), supporting both PX4 and ArduPilot autopilots. It provides full flight control, real-time telemetry monitoring, mission planning with complex survey/scan patterns, video streaming with hardware-accelerated decoding (D3D, Vulkan, DMABuf), log analysis (ULog, DataFlash), and vehicle configuration wizards. The architecture is a monolithic Qt/C++17 application with a QML-based UI, organized around a plugin system for firmware-specific features, a custom state machine framework for async workflows, and a type-safe Fact system for parameter management. It also includes extensive CI/CD infrastructure, a comprehensive test suite, and support for Android, iOS, macOS, Windows, and Linux. The repository is hybrid in that it contains documentation (vitepress), Python tooling for CI/code generation, and data-oriented log parsers, but its dominant character is a complex desktop/mobile application.
Flix is an open-source ESP32-based quadcopter project built from scratch. The repository contains flight controller firmware, a Python control/telemetry library, C++ utilities for log conversion, simulation assets, and extensive documentation. It is primarily an application repo, though it also includes CAD models and data-conversion tools that serve as supporting assets. At runtime, the ESP32 firmware reads IMU data, estimates attitude, and drives motors, communicating via MAVLink and ESP-NOW for telemetry and control. The Python library (tools/pyflix/flix.py) provides an observable API for interacting with the vehicle, and CLI tools convert flight logs from CSV to MCAP and ULog formats for analysis in standard robotics tooling. A Gazebo simulation and a Python test harness validate behavior without hardware. Architecture is a multi-part monorepo: firmware, Python tools, C++ tools, docs, and 3D-printable mechanical parts. External dependencies include the Arduino ESP32 core, MAVLink, ESP-NOW, Gazebo, and custom math classes. The code is generally well-structured, but some data flow relies on global variables and hardware-specific assumptions, limiting reuse outside the ESP32 ecosystem. The project appears actively maintained with CI workflows, but dependency pinning and tight hardware coupling create moderate drift risk.