Abstract

In this paper, we consider a NarrowBand-Internet of Things (NB-IoT) network where an Unmanned Aerial Vehicle (UAV) is employed to gather data from IoT devices deployed in a given area. It is well known that UAVs may fly over the terrestrial plane, where and when needed, acting as Unmanned Aerial Base Stations (UABs). In order to serve as many ground IoT devices as possible, a proper trajectory design is fundamental. As we show in the paper, the optimization of the UAV speed and the radio parameters are also essential. Specifically, this paper studies a cluster-based scenario, where IoT devices are deployed according to a Thomas process, and applies a Traveling Salesman Problem approach to design the UAB trajectory. Notably, our model considers the protocol constraints on the number of resource units available on the UAB’s NPUSCH, and the data rate that it can provide to IoT devices. Our results reveal the impact of different design parameters, such as UAB speed and NPRACH periodicity on the network throughput and the number of requests served.