Experimentation and Simulation with Autonomous Coverage Path Planning for UAVs

UAVs have been widely used in many photography applications. Among photography tasks, covering an area to produce photogrammetry data is a prominent solution for commercial applications. The resources available to automate these processes, despite the many advances, still are somewhat scarce for practical deployment. In this research, the authors intend to bridge this gap by evaluating several simulated and real 3D models. The methods evaluated are also compared with current commercially available planners and the literature state of the art. The research uses seven heuristics methods to generate camera placement, another seven methods to generate offline planning, and seven methods to perform path planning. The tests include synthetic data for statistical significance, real 3D models, and simulations that allow a complete performance overview. Quantitative results will enable the user to visualize each method's performance, while qualitative results will help to understand the results visually. Results are compiled in a path planning library for further research and development. They show that some methods can be 15% more cost-effective while being able to be still computed in a reasonable amount of time.