Learning to communicate in UAV-aided wireless networks: Map-based approaches

We consider the scenario of a UAV-mounted flying base station providing data communication services to a number of radio nodes spread over the ground. We focus on the problem of resource-constrained UAV trajectory design with (i) optimal parameter learning and (ii) optimal data throughput as key objectives, respectively. While the problem of throughputoptimized trajectories has been addressed in prior works, the formulation of an optimized path to efficiently discover propagation parameters has not yet been addressed. When it comes to the data communication phase, the advantage of this work comes from the exploitation of a 3D city map. While the optimization of a flying path directly based on the raw map data leads to an intractable non-differentiable cost minimization problem, we introduce a novel map compression method allowing us to tackle the problem with standard tools. The path optimization is then combined with a node scheduling algorithm. The advantages of both the learning path optimization and the map compression method for data communication trajectory design are illustrated in an urban IoT setting.