Impact of Microwave and mmWave Channel Models on 5G Systems Performance

Abstract

Channel models and measurements across a wide range of candidate bands for fifth generation wireless networks are considered. Motivated by the different propagation and spatial characteristics between different bands and channel models within a band, we investigate how key channel modeling and spatial parameter differences impact various antenna topologies in terms of sum rate, channel eigenvalue structure, effective degrees of freedom (EDOF), channel connectivity, and massive multiple-input-multiple-output (MIMO) convergence. We show that due to the sparsity of millimeter-wave (mmWave) channels, any variation in spatial parameters can dramatically affect the sum rate. In microwave scenarios, where the probability of line-of-sight (LOS) propagation is low, the structure of the eigenvalues is highly dependent on the richness of scattering. In mmWave bands, where the probability of LOS is higher, the structure of the eigenvalues is largely dependent on the LOS channel model. The uniform linear array is seen to have a superior sum rate and an eigenvalue structure due to the inherently larger interelement spacings and wider azimuth spectra. These observations are seen to affect the sum rate, EDOF, and massive MIMO convergence. Two variations of channel connectivity are then considered, and compared with EDOF, to examine the richness of scattering and channel rank.