ETD

• autonomous navigation
• GPS-denied environments
• optical flow
• PX4
• robotics
• ROS 2
• sensor fusion
• unmanned aerial vehicle

This thesis presents the design and implementation of a robust autonomous navigation and control framework for unmanned aerial vehicles operating in GPS-denied environments. The proposed system integrates optical flow sensors, the PX4 flight stack, and ROS 2 to enable autonomous flight without relying on satellite-based positioning. A complete offboard control architecture was developed in ROS 2, where control algorithms generate real-time navigation commands and transmit position and velocity setpoints to the PX4 flight controller. The work includes the integration and fusion of optical flow sensor measurements for reliable motion estimation, firmware configuration, software development, and communication between ROS 2 and PX4 through the uXRCE-DDS bridge. The proposed framework was validated through simulation and experimental flight testing to evaluate its performance, robustness, and stability under GPS-denied conditions. The results demonstrate that the combined navigation, sensor fusion, and ROS 2-based offboard control architecture provides a reliable solution for autonomous UAV operation in challenging environments, contributing to the advancement of autonomous aerial systems for indoor and other GPS-denied applications.