Unmanned Aerial Vehicle (UAV) communication is known to suffer from significant interference due to the clearance of the radio paths with ground base stations. Multi-antenna receive combining has the promise of alleviating the impact of interference, translating to improved connectivity performance. In this paper, we evaluate the performance of Conventional Beamforming (CB) and Maximum Ratio Combining (MRC) receivers for UAV communication based on live Long Term Evolution (LTE) networks. Our measurement setup consists of nine Universal Software Radio Peripheral (USRP) boards and a circular antenna array with sixteen elements. The LTE signals are recorded at different UAV flight heights in urban environments, and processed offline. Results show similar Signal-to-Interference-plus-Noise Ratio (SINR) performance by MRC and CB, with CB slightly outperforming MRC provided knowledge of LTE signal structure is used for the beam selection. No significant dependency from the flight height has been observed. The outage probability analysis further emphasizes the benefits of using CB in the studied scenarios.

Cellular Vehicle to Everything (C-V2X) communications with its safety and infotainment services will require a high performance receivers to cope with challenging throughput, latency and reliability requirements. With increasing levels of interference due to cell densification and introduction of the roadside units, single antenna receivers may not be able to provide the required quality of service. In this work we experimentally study the performance of multi antenna receivers based on more than 150 km of data recorded during experiments using a customized software defined radio testbed. The performance of sixteen antennas Maximum Ratio Combiner (MRC) is compared with the receive beamforming technique for the live cellular signals in the 1.8 GHz band. This study is followed by an analysis of the impact of interference and measurement environment on the receiver's performance. The results show that receive beamforming can outperform MRC in low-interfered scenarios with high Line of Sight (LoS) probability, like highways or rural areas, while ensuring comparable performance even in dense urban scenarios where LoS communication cannot be guaranteed.

Cellular networks based on current LTE Advanced or coming 5G technology, constitute an attractive candidate to provide communication links for drones in low level airspace. This paper extends previous findings on path loss exponent and shadow fading variation for an urban environment, by adding new and more extensive empirical evidence. Specifically, by way of the measurement methodology we study also the spatial correlation of the shadow fading, including the lateral spatial autocorrelation (decorrelation distance) and the crosscorrelation in shadowing variability between different heights. Data was obtained with the use of a drone flying at heights up to and above rooftop, using a radio network scanner for measuring live LTE signals at 1800 MHz. The path loss analysis confirmed the validity of previously presented modelling frameworks for the behaviour versus height, as well as the characteristics of the increased interference observed when the drone is above rooftop level. As for the spatial correlation of the shadow fading, the results indicate that decorrelation distance does not change appreciably from ground level up to and above rooftop level, and remains in the order of 10m. The shadow fading variability is generally uncorrelated versus height, although with a trend to be correlated below rooftop level, but not above.

This work evaluates the concept of uplink beamforming for vehicular communications in the sub-6 GHz frequency bands to improve throughput, latency and coverage of the vehicle to Base Station (BS) link. The data recorded in the experimental measurements using live cellular signals are used to study the performance of two direction acquisition methods: the Angle of Arrival (AoA) estimation and downlink-based beam sweep. Next, the feasibility of signal tracking techniques exploiting the location of the vehicle and the BS are investigated to alleviate the need for continuous direction acquisition. The results show that the downlink-based beam sweep leads to higher Signal to Interference and Noise Ratio (SINR) than beamforming based on the estimated AoA. Evaluated tracking techniques are shown to be capable of correctly estimating the beamforming angle for distances in order of hundreds of meters when BS's location is known to the vehicle.