5G-Based Localization for Unmanned Aerial Vehicles

Unmanned Aerial Vehicles (UAVs) have shown great promise for applications such as delivery, disaster response, mapping, infrastructure inspection, and passenger transport. However, the reliance on Global Navigation Satellite Systems (GNSS) poses a significant limitation in urban and suburban environments, where signal reliability is often compromised. The loss of GNSS signals poses a major safety risk for autonomous urban air mobility systems. 5G New Radio-based positioning methods offer a promising alternative using 5G radio signals from existing infrastructure, from 5G network towers using triangulation, providing reliable localization in dense environments. Using the existing ubiquitous infrastructure and readily available radio signals, 5G New Radio enables easily integrable structure for localization, potentially reducing the reliance on additional sensors like IMUs, LiDAR, or depth camera, which are susceptible to failure or degraded performance in challenging conditions such as low lighting or unstructured environments. While much existing research on 5G PRS has been simulation-based, and hardware testbeds for UAV platforms remain limited, our work introduces a hardware testbed to localize the UAV by calculating time difference of arrival (TDOA). We mount a software-defined radio (SDR) on the UAV to capture and store 5G PRS signals for post-processing, establish a synchronized transmission setup from ground stations, and develop a signal processing pipeline for TDOA-based localization. To validate our localization method, we characterize localization errors by comparing 5G PRS-based position estimates with those obtained from Motion Capture (MoCap) systems. We have achieved a mean error of approximately 4.6 meters in three experiments conducted in a controlled indoor environment.