From exploration to surveillance: autonomous UAV systems for indoor and outdoor missions

Unmanned Aerial Vehicles (UAVs) are increasingly used in applications such as industrial inspection, environmental monitoring, and security operations. In these scenarios, autonomous capabilities are essential to ensure safe and reliable operation in complex environments, especially when Global Navigation Satellite System (GNSS) signals are unavailable or continuous human supervision is impractical. This thesis investigates the design and implementation of UAV-based systems for autonomous exploration and industrial surveillance through two complementary case studies. The first project, developed within the Leonardo Drone Contest, focuses on an autonomous UAV capable of operating in GPS-denied indoor environments. The system integrates perception, navigation, and control within a robotic architecture based on the PX4 autopilot and ROS 2 middleware. A stereo vision pipeline reconstructs the surrounding environment and generates point clouds for obstacle detection and mapping. Autonomous exploration strategies allow the UAV to navigate unknown environments, while path planning and trajectory tracking algorithms ensure safe and feasible flight. The second project concerns the design of an industrial UAV surveillance system developed during an internship at FieldRobotics. The focus shifts from onboard autonomy to the development of a high-level supervision infrastructure for industrial monitoring. The system integrates UAVs equipped with thermal sensors and automatic charging stations with a web-based control interface that enables operators to supervise missions, monitor telemetry data, and access real-time video streams. The proposed solutions were validated through simulation and experimental tests, demonstrating the effectiveness of the architectures for autonomous navigation and industrial monitoring tasks.