On the Association of Small Cell Base Stations with UAVs Using Unsupervised Learning

Abstract

Small cell networks (SCNs) offer a cost-effective coverage solution to wireless applications demanding high data rates. However in SCNs, a challenging problem is the proper management of backhaul links to small cell base stations (SCBSs). To make a good backhaul link, perfect line-of-sight (LoS) communication between the SCBSs and the core network plays a vital role. In this study, we use the idea of employing unmanned aerial vehicles (UAVs) to provide connectivity between SCBSs and the core network. We focus on the association of SCBSs with UAVs by considering multiple communication-related factors including data rate limit and available bandwidth resources of the backhaul. In particular, we address the optimum placement of UAVs to serve a maximum number of SCBSs while considering available resources using unsupervised \textit{k}- means algorithm. Numerical results show that the proposed approach outperforms the conventional approach in terms of associated SCBSs, bandwidth consumption, available link utilization, and sum- rate maximization.