A Subarray Selection Scheme for Cellular-Connected UAV With Conformal Phased Array Antenna

In this article, the millimeter wave (mmWave) cellular-connected unmanned aerial vehicle (UAV) equipped with conformal phased array antennas in cylindrical shape is investigated, where the UAV acts as an aerial user and served by the base station (BS). Although the cylindrical conformal array (CCA)-enabled UAV can achieve the full spatial coverage and fully exploit the beam gain with directional antenna elements, the curved carrier of CA forces the cellular-connected UAV to activate only a portion of antenna elements rather than activating the entire of them to mitigate the severe aerial-ground interference from other BSs. The subarray selection problem is formulated to maximize signal-to-interference-plus-noise ratio (SINR) with non-linear antenna radiation patterns. To handle the non-linearity of objective problem, we propose a low complexity subarray selection scheme based on genetic algorithm (GA), which utilizes only the information that UAV can acquire in practical scenarios. The simulation results demonstrate that the proposed subarray selection scheme always shows higher SINR and achievable rate compared to the benchmark schemes. Specifically, the proposed subarray selection scheme shows an achievable rate that is about 2% higher than the accelerated particle swarm optimization (APSO) algorithm, and its computational complexity is only about 25%. In addition, the proposed scheme has near optimal performance while having only about 10.4% complexity compared to the optimal subarray obtained through the full search.