Measurement based performance evaluation of drone self-localization using AoA of cellular signals

This paper focus on drone self-localization based on received signals from stationary base stations with known locations without having access to global navigation satellite system (GNSS) signals. In the considered method, the drone first estimates the angle of arrival (AoA) based on the downlink signal strength measurements by rotating non-isotropic antenna. Once AoA is estimated from several base stations, the drone can be localized on a 2D plane by using the least squares method. This method was selected due to its simplicity and thus ease of implementation in low cost drones. To better understand the achievable performance of the studied AoA estimation method, we performed several measurements using a prototype system to characterize the antenna rotation and AoA estimation errors. The error models were then plugged into simulator to analyze the achievable performance in typical macro cellular setting. Our measurement results indicate, that the AoA could be estimated with 6.02° accuracy in case of a hovering drone. Our simulation results indicate, that in typical macro cellular setup with line-of-sight propagation, the median localization error was less than 45 m for a hovering drone.