Introduction
As the industry moves toward Beyond Visual Line of Sight (BVLOS) drone operations for medical deliveries and urban surveillance, reliable wireless connectivity is essential. However, existing cellular networks are optimized for ground users. Unmanned Aerial Vehicles (UAVs) face unique challenges: antenna misalignment, where drones fly in the "nulls" of downward-tilted base station antennas, and increased interference due to unobstructed Line-of-Sight (LoS) to multiple distant towers. This study investigates whether a two-tier network—integrating urban small cells with traditional macrocells—can mitigate these issues in a dense urban environment.
Experimental Methodology
The research team conducted a measurement campaign in the Dublin Smart Docklands. The setup utilized a DJI Matrice M300 UAV equipped with a Google Pixel 3 handset. The network operated on the 3.6GHz band (Band 42) with two distinct tiers: Tier 1 consisted of rooftop Massive MIMO macrocells, and Tier 2 comprised small cells mounted on lampposts and traffic lights.
The drone followed a back-and-forth flight path at altitudes ranging from 30m to 120m. Data was collected in both Idle Mode (signal monitoring) and Dedicated Mode (active data transfer) to evaluate Reference Signal Received Power (RSRP), Signal to Interference and Noise Ratio (SINR), and throughput.
Key Findings and Results
The study revealed a "Massive MIMO Paradox": the benefits of Massive MIMO decrease at higher altitudes because the clear LoS reduces the multipath diversity that these systems rely on. In contrast, small cells acted as a reliable "aerial anchor." While macrocell signal quality in the Grand Canal Quay (GCQ) area dropped significantly as the drone ascended, the small cells in the North Wall Quay (NWQ) maintained a remarkably stable performance profile.
Specifically, downlink throughput in the macrocell-heavy GCQ area dropped by over 30% (from ~70 Mbps to ~40 Mbps) as height increased to 120m. In the small-cell-prioritized NWQ area, throughput remained consistent at 30-35 Mbps. Furthermore, the hierarchical network successfully managed handovers, maintaining a low rate of approximately one every 30 seconds, preventing the "handover storms" often feared in aerial networking.
Conclusion
The experimental results demonstrate that while traditional macrocells struggle to serve high-altitude UAVs due to antenna patterns and interference, a secondary tier of urban small cells provides the stability necessary for reliable BVLOS operations. Future research will explore these dynamics at mmWave frequencies and investigate uplink Quality of Service (QoS) for high-bandwidth applications like real-time 4K video streaming.
Summary updated