Simultaneous User Association and Placement in Multi-UAV Enabled Wireless Networks

In this paper, we are proposing a map-based approach for the optimal placement of multiple UAV-based flying wireless relays in a cellular network. The tackled problem is two-fold, involving a joint UAV-user association problem and 3D placement problem. While related problems were addressed before, the novelty of our approach lies in the fact it builds on a combination of probabilistic and deterministic line-of-sight (LoS) classifiers which exploits the availability of a 3D city map. While the original problem is very challenging in its dimension, we give a low-complexity approach to the placement problem by approximating the optimum UAV positions with a suitably weighted combination of user positions. Our simulations suggest a performance close to that obtained with high complexity exhaustive search for placement.