Integrating Connected UAVs into Future Mobile Networks

Abstract

With the increasing number of Unmanned Aerial Vehicles (UAVs) and their applications, such as performing search and rescue or transplant organ delivery, the need for improving the UAV connectivity grows. Currently, User Equipment (UE)s have a range of connectivity options, such as WiFi and Lora. The integration of the UAV as a UE of the mobile network can increase the guarantee of the UAV?s Quality of Service (QoS) and the range of its available connectivity, due to higher reliable and range of mobile networks. This would, in turn, enable wider and more reliable applications of UAVs. In this thesis, we investigate how to improve the QoS of a UAV connected to the mobile network, without requiring changes to the mobile network. The three main contributions of this thesis, towards the integration of UAVs into 5G and beyond are: the identification of the mobility challenges a mobile operator may encounter if a UAV is integrated as a UE of the mobile network; a Reinforcement Learning (RL) approach to optimise the UAV's QoS while adapting UAV's height; and an object detection approach that classify different RATs and extract features from the transmissions.