Altitude-Dependent Cellular Spectrum Occupancy: From Measurements to Stochastic Geometry Models

The growing demand for autonomous aerial vehicle (AAV)-based aerial connectivity across urban and rural environments highlights the need for a deeper understanding of spectrum occupancy in aerial corridors. In particular, understanding the altitude-dependent behavior of spectrum occupancy in cellular networks is critical, as such understanding could be used to enable beyond visual line of sight (BVLOS) UAV connectivity in the future. While there are existing models for characterizing altitude-dependent interference in the literature, they are not validated by data and need to be compared with real-world measurements. To address these gaps, in this article, we conduct cellular spectrum occupancy measurements at various sub-6 GHz bands for altitudes up to 300 m, in both urban and rural environments. To characterize the real-world spectrum occupancy behavior, we introduce two different approaches: 1) a new theoretical model utilizing stochastic geometry with altitude-dependent factors (SOSGAD), and 2) a ray-tracing model tailored to site-specific line of sight (LOS) and non-LOS scenarios. We also derive the asymptotic behavior of the SOSGAD model as the UAV altitude increases. Through comparative analysis, we evaluate how well the SOSGAD and ray-tracing models characterize altitude-dependent spectrum occupancy in both uplink and downlink scenarios.