A Complete Coverage Algorithm for 3D Structural Inspection using an Autonomous Unmanned Aerial Vehicle

This thesis presents a novel algorithm for complete coverage of three-dimensional structures to address the problem of autonomous structural inspection using an Unmanned Aerial Vehicle (UAV). The proposed approach uses a technique of cellular decomposition based on Morse decomposition to decompose the 3D target structure into 2D coverable faces that are subsequently connected using a graph-based representation. We then use graph traversal techniques such as the Traveling Salesman Problem (TSP) to generate a flight coverage path through the decomposed faces for a UAV to completely cover the target structure, while reducing the coverage time and distance. To test the validity of our proposed approach, we have performed a series of experiments using a simulated AscTec Firefly UAV in different environments with 3D structures of different sizes and geometries, within the Robot Operating System (ROS) Gazebo simulator. Our results show that our approach guarantees complete coverage of the target structure. Comparison of our coverage strategy with other strategies shows that our proposed TSP-based coverage strategy performs up to 50% better in reducing the flight path with an average of 30% fewer turns and 12% less coverage duration than a largest-area-first approach