Title:Placement Optimization for UAV-Enabled Wireless Networks with Multi-Hop Backhaul

Abstract:Unmanned aerial vehicles (UAVs) have emerged as a promising solution to provide wireless data access for ground users in various applications (e.g., in emergence situations). This paper considers a UAV-enabled wireless network, in which multiple UAVs are deployed as aerial base stations (BSs) to serve users distributed on the ground. Different from prior works that ignore UAV's backhaul connections, we practically consider that these UAVs are connected to the core network through a ground gateway node via rate-limited multi-hop wireless backhauls. We also consider that the air-to-ground (A2G) access links from UAVs to users and the air-to-air (A2A) backhaul links among UAVs are operated over orthogonal frequency bands. Under this setup, we aim to maximize the common (or minimum) throughput among all the ground users in the downlink of this network subject to the flow conservation constraints at the UAVs, by optimizing the UAVs' deployment locations, jointly with the bandwidth and power allocation of both the access and backhaul links. However, the common throughput maximization is a non-convex optimization problem that is difficult to be solved optimally. To tackle this issue, we use the techniques of alternating optimization and successive convex programming (SCP) to obtain a locally optimal solution. Numerical results show that the proposed design significantly improves the common throughput among all ground users as compared to other benchmark schemes.