arXiv:2606.00277v2 Announce Type: replace Abstract: Fast and low-overhead beam management is essential for the practical deployment of unmanned aerial vehicle (UAV) communication networks operating at millimeter-wave and higher frequencies. In this paper, we propose a radar-assisted beam selection framework for UAV networks that exploits spatial sensing information, including angle-of-departure and distance estimates, to efficiently identify communication beams. Specifically, a digitally beamformed radar array with a relatively small number of antennas provides rapid, coarse UAV localization, while an analog-beamformed communication array, requiring fewer radio-frequency chains, exploits coarse localization to narrow the beam-search region and establish a communication link with the UAV. To provide theoretical insight into the expected overhead, we conduct a probabilistic analysis in which an approximation of the nearest-first beam selection overhead under idealized conditions is proposed and its statistics are derived for a Gaussian error model. The simulation results demonstrate that the proposed radar-assisted beam selection significantly reduces the number of beams that need to be examined, requiring on average at most 1.64 candidate beams even under per-element radar phase errors of up to 100{\deg}.