Evaluation of Direct Search Algorithms to Trajectory Optimization for a Perpetually Flying Fixed-Wing Solar UAV Providing Network Coverage

Abstract

Recent advances in battery and solar-cell technology have allowed for small-scale solar-powered fixed-wing aircraft to achieve perpetual-endurance flight, where they harvest enough energy to fly for long periods of time without recharging. We consider how this capability can be utilized in a networking context to enable rapid deployment of wireless Internet coverage in areas where it has been disrupted or is otherwise unavailable, providing a relay between portable ground-based devices such as cell phones and the currently developing Very-Low-Earth-Orbit satellite Internet layer. In particular, we investigate the impact of the aircraft trajectory on the level of service provided and the energy balance of the vehicle. We develop a framework for evaluating and optimizing these trajectories in three-dimensions and compare two optimization methods, Particle Swarm Optimization and the Nelder-Mead method, in optimizing the throughput of the trajectory while respecting the energy constraints implicit to providing perpetual-endurance flight.