Over the past few years, there has been a growing interest in using unmanned aerial vehicles (UAVs) for high-rate wireless communication systems due to their highly flexible deployment and maneuverability. The aim of this article is to propose a 3-D multi-UAV deployment approach to provide Quality-of-Service (QoS) requirements for different types of user distributions in the presence of co-channel interference by maximizing the minimum achievable system throughput for all of the ground users. The proposed approach is divided into two separate algorithms. In the first algorithm, by using the mean-shift technique and prior knowledge of users’ positions provided by the global positioning system (GPS), it has been shown that one can simultaneously find $xy$ coordinates of UAVs, which are associated with the maximum of users’ density and schedule users to UAVs. Once the $xy$ -Cartesian coordinates of UAVs are determined, UAVs’ altitudes and transmit powers are separately optimized. Since these problems are nonconvex optimizations, the successive convex optimization technique has been applied to approximate their nonconvex constraints. In the second algorithm, the block coordinate descent technique is leveraged to jointly optimize UAVs altitudes and transmit powers by tightening the bounds obtained for approximations. It is then proven that the suggested algorithm is guaranteed to converge. The computational complexity of the proposed placement approach is derived. Numerical experiments are carried out to evaluate the performance of our technique and show its superiority to conventional benchmarks.
Read full abstract