ModularSensorSystem

Table of Contents

Main Features

• Plugin-based architecture
  • Sources
  • Data Processors
• Simultaneous data recording
• Data visualization
• "Clipping" feature
• Real-time data processing
• Efficient and organized UI/UX

Architecture Overview

For my own sanity and anyone else who might want to contribute, I am going to give a brief overview of the architecture of this project.

• The frontend built with Qt 6.9.0 and C++20, providing a modern and responsive user interface.
• The backend is implemented in C++20, ensuring high performance and efficiency.
  • I am attempting to use the Qt framework as much as possible (ex: QString instead of std::string), but there are some areas where I have to use standard C/C++ libraries.
• Dependencies are managed with a vcpkg manifest.
  • This allows for easy installation and management of libraries across different platforms; specifically, Windows and Linux.
• The system is built around a plugin-based architecture, allowing for easy addition of new data sources and processors.
  • Plugins are dynamically loaded at runtime, enabling users to extend the system's capabilities without modifying the core codebase.
  • Currently, 2 features are "pluggable": sources and processors (more on them in a bit)
  • To create a new plugin, you can follow the Plugin Development Guide (unfinished).
• A "Source" is a class that provides data to the system.
  • A "Source" could be a physical device (video, audio, GPS, etc.), a software component, an API, or any other data provider.
  • All sources implement a common interface, allowing the system to handle them uniformly.
• A "Processor" is a class that processes data from sources.
  • This could include filtering, transformation, or any other data manipulation.
  • Primarily (for my use case), processors will be used for real-time object detection, motion detection, and tracking.

Supported Source Types

• Analog Sensors
• Video Input
• Audio Input
• GPS NMEA