Latency aware 3d placement and user association in drone-assisted heterogeneous networks with fso-based backhaul

Drone Base Station (DBS) is expected to play an important role in next generation cellular networks due to its maneuverability in establishing air-to-ground Line of Sight (LoS) links. Working as a mobile relay node between the Macro Base Station (MBS) and the ground users, the DBS can significantly reduce the latency of ground users by offloading traffic loads from the MBS to the DBS since a better channel condition can be provided. In this paper, we focus on how to determine the 3D location of the DBS, the user association and the bandwidth allocation policy between the MBS and the DBS in order to minimize the total average latency ratio of all users with the constraint of each user's QoS requirement and total available bandwidth. The formulated problem is a mixed integer nonconvex optimization problem, a very challenging and difficult problem. We propose a cyclic iterative algorithm to efficiently solve it by decomposing the primal problem into two subproblems, i.e., the user association and bandwidth allocation problem and the 3D DBS placement problem. In each iteration, the two sub-problems are alternatively optimized by using the output of one as the input for the other. Numerical simulation results demonstrate the significant latency ratio reduction of our proposed algorithm as compared to other baseline schemes.