LAUSR

The unmanned aerial vehicle (UAV) channel models are very important and useful for the design and evaluation of UAV communications. In existing works, the terrestrial scatterers are widely stationary or moving all the time, and the UAV is assumed to move at a constant speed. However, these modeling assumptions may not be appropriate in realistic scenarios. In this work, we propose a new 3-dimensional (3-D) dynamic non-wide-sense stationary (non-WSS) cluster-based geometric model for UAV multiple-input multiple-output (MIMO) channels. We assume that a few non-isotropic scatterers are distributed around the terrestrial receiver and have a probabilistic movement behavior. Meanwhile, we consider that the UAV flies at a varying speed, and we design a non-stationary mobility model with acceleration. The proposed channel model performs the birth-death evolutionary process of clusters based on a Markov model, and the time-variant multipath component (MPC) parameters (e.g., angle of arrival/departure (AoA/AoD), power, and delay) are obtained. In addition, we derive the space-time correlation function, mean Doppler spread, and root-mean-square doppler shift spread of the UAV MIMO channels, and provide the quasi-stationary interval. We also investigate the impact of some key model factors on the statistical properties by numerical simulations. Finally, we compare the space-time correlation functions and the second-order statistical properties between the proposed model and the UAV channel measurement results. The close agreements between the numerical simulations and field measurements have validated the proposed model.