Spatial Correlation Models of Large-Scale Antenna Topologies Using Maximum Power of Offset Distribution and its Application

In this paper, we present an approximate expression for spatial correlation of cylindrical and uniform rectangular arrays regarding the maximum power of offset distributions useful for evaluating geometrybased stochastic channel models. The underlying concept of the proposed spatial correlation expression is that the maximum power of arrival varies relative to the offset distribution in the azimuth and zenith domains. Verification is accomplished with the help of computer simulation, where perfect agreement between theoretical and Monte Carlo simulation results is found. Toward the implementation of large-scale MIMO transmitters in constrained physical spaces to accomplish the demands for the next generation of wireless systems, we examine the effect of the proposed spatial correlation model on the downlink (DL) performance of massive multi-input-multi-output (MIMO) system. Results illustrate that the DL capacity does not demonstrate any sensitivity to the effect of correlation by increasing the number of elements in a realistic massive MIMO channel. Surprisingly, while the spatial correlation restricts the DL capacity at a reduced number of transmit antennas, the restriction becomes insignificant as the number of transmit antennas increases. This supports existing developments that user channels decorrelate when the number of antennas at the base station increases. Finally, experimental results show that separation of more than half wavelength is inadequate to decorrelate the antenna elements.