UAV Simulation for Semantic Mapping and Human Detection

UAV Simulation for Semantic Mapping and Human Detection

Introduction

  The project focuses on the development of an Unmanned Aerial Vehicle (UAV) application designed for manual keyboard-controlled navigation, real-time Simultaneous Localization and Mapping (SLAM) utilizing RTAB-Map (Real-Time Appearance-Based Mapping), and advanced computer vision-based human detection using YOLO.

 

Purpose

  By leveraging custom CAD digital modeling, modern robotics frameworks, and artificial intelligence, this system establishes a scalable platform suitable for complex, large-scale missions such as surveillance, mapping, and search-and-rescue operations.

 

Development

Component Technology / Tool Used
Operating System Ubuntu 24.04 LTS (Noble Numbat)
Robotics Framework ROS 2 (Jazzy Jalisco)
UAV Modeling SolidWorks with URDF Exporter plugin (URDF & STL generation)
Simulation Environment Gazebo Harmonic
Flight Control Stack PX4 Autopilot Stack
SLAM Package RTAB-Map (Real-Time Appearance-Based Mapping)
Computer Vision Packages OpenCV, YOLO (You Only Look Once)
Ground Control & Telemetry QGroundControl (Real-time flight operation management)
Sensor Data Visualization RViz
Programming Languages C++ (Low-level performance/development) Python (High-level script orchestration)
Repository
 

Images