Waypoint Planning for Autonomous Aerial Inspection of Large-Scale Solar Farms

Abstract

Solar energy is seen as a sustainable and non-depletable source of energy supply. Worldwide, large-scale solar power infrastructure is being installed every day. Such structures can suffer from many faults and defects that degrade their energy output during their operational life. Detecting such faults and defects requires regular inspection over physically large and distributed solar infrastructure. On-site manual human inspection tends to be impractical, risky and costly. As such, replacing humans with autonomous robotic aerial inspection systems has great potential. In this work, we propose an unmanned aerial vehicle (UAV) waypoint generation system that is specifically designed for aerial inspection of solar infrastructure. Our system takes into consideration the physical structure and the dynamic nature of sun-tracking solar modules and generates waypoints with the right camera viewing pose and drone orientation. Statistical methods are used to generate a randomly selected set of modules as a representation of the entire solar farm. The set is guaranteed to satisfy a user-defined confidence level and margin of error requirements. A path is generated to visit selected modules in an optimal way by deploying the traveling-salesman shortest path algorithm, allowing the vehicle to maximize battery use. Illustrative flights and preliminary inspection results are presented and discussed.