Abstract:
The statistical gap between existing correlation-based stochastic channel models (CBSMs) and practical propagation is a longstanding issue, especially for massive multipl...View moreMetadata
Abstract:
The statistical gap between existing correlation-based stochastic channel models (CBSMs) and practical propagation is a longstanding issue, especially for massive multiple-input multiple-output (MIMO) transmission. To address this problem, a pervasively correlated channel model (PCCM) applicable to MIMO channels with arbitrary antenna configurations and scenarios is proposed. Unlike the conventional jointly correlated channel model (JCCM) based on an independently and identically distributed (i.i.d.) random matrix, a new random matrix whose elements are independent and nonidentically distributed (i.n.d.) generalized Gamma complex Gaussian mixture (GGCGM) variables with correlated envelopes is used to approximate the real channel statistics better. Moreover, the PCCM can be simplified via the Rayleigh fading assumption for rapid evaluation of channel performance and supports backward compatibility with existing CBSMs under specific assumptions. Demonstrative numerical experiments for MIMO channels are conducted based on the 3GPP TR 38.901 model considering various scenarios, frequencies, and antenna configurations. The channel capacity distributions are obtained based on random samples generated using the geometry-based stochastic channel model (GBSM), the JCCM, and the proposed PCCM. The numerical results show that the PCCM is more flexible than the JCCM with respect to high-order statistics, enabling more accurate estimation of massive MIMO transmission channel performance.
Published in: IEEE Transactions on Communications ( Volume: 72, Issue: 4, April 2024)