Abstract
Millimeter-wave (mmWave) communication offers high data rates and inherent spatial security, but its directional nature poses significant challenges for mobile platforms such as Unmanned Aerial Vehicles (UAVs). In this paper, we develop a low-cost portable Universal Software Radio Peripheral (USRP)-based mmWave Software-Defined Radio (SDR) testbed for UAVs. Our testbed integrates compact USRP-based airborne nodes with Ka-band up/down-conversion front-ends, supports interchangeable antenna configurations including K-band horn antennas and a Sivers EVK02004 beam-steerable patch array, and provides a Python API for real-time beam control. The design enables rapid deployment of UAV-to-UAV as well as UAV-to-ground mmWave links within a portable and extensible aerial research framework. This testbed provides a flexible and open framework for researchers to prototype and evaluate advanced UAV-based mmWave communication strategies, including machine learning-driven beam alignment, neighbor discovery, and secure link acquisition. We present UAV-to-UAV and UAV-to-ground measurements under various scenarios and using different antennas.