A drone’s flight height can be estimated from radio signals of base stations. Simulations yield an accuracy of about two meters. Ongoing flight experiments aim to validate the technique in real 5G networks.
Written by Christian Bettstetter. Illustration created using Adobe Firefly.
Most drones rely on satellite-based navigation systems like GPS for positioning and navigation. However, these satellite signals are sometimes unavailable due to high buildings — or they are blocked, jammed, or spoofed — making complementary positioning technologies essential. While many alternatives exist, most are designed for ground users and provide only two-dimensional information. Drones, by contrast, also require accurate height information, which is typically obtained from dedicated onboard sensors.
A different solution for height estimation has been developed by Enrique Caballero, Aymen Fakhreddine, and Christian Bettstetter from the University of Klagenfurt. Instead of relying on dedicated sensors, their method uses the wireless transceiver that many drones carry to connect to cellular networks. Doctoral student Enrique Caballero explains: “Drones experience radio channels that differ fundamentally from those of ground users. When a drone gains altitude, it moves through different antenna sidelobes and establishes line-of-sight links to many base stations. These effects create distinctive signal strength patterns, which can be learned and interpreted by machine learning algorithms.”
To evaluate the approach, the researchers simulated a suburban area around the University of Klagenfurt using real base station locations and aerial propagation models. More than six million signal strength samples were generated and used to train different machine-learning models. Among the tested methods, an ensemble decision tree achieved the best accuracy, achieving a root mean square error of less than two meters.
The project is now entering its next phase. In collaboration with Lakeside Labs, experiments are conducted with drones connected to 5G networks. These tests will assess how well the technique performs under real-world conditions and how antenna configurations and environmental characteristics affect its accuracy. If successful, radio-based height estimation could become a valuable complement to GPS-based navigation, providing an additional source of information and increasing the resilience of drone operations when GPS signals are unavailable or unreliable.
Publication
Enrique Caballero, Aymen Fakhreddine, and Christian Bettstetter. Drone Height Estimation from 5G Signals. Manuscript accepted for Proc. European Wireless, Rimini, Italy, June 17–18, 2026. Download author’s preprint.
This research was funded in whole or in part by the Austrian Science Fund (FWF) (grant ESPRIT-54). Some text passages in this blog article stem from the above-mentioned publication, mainly in modified form. Suggestions for language improvement were generated using ChatGPT.