Closed-form Analysis of RZF in Multicell Massive MIMO Over Correlated Rician Channel

Correlation exists in all practical channels, and each channel's correlation level varies. Determining correlation matrices while estimating the channel and precoding is vital in providing spatial directivity to the signals. This paper considers the performance analysis of the downlink multi-cell multiuser Massive Multiple Input Multiple Output (mMIMO) system with user-specific correlation matrices for the Rician Fading channel. Using the Random Matrix Theory (RMT) results, the closed-form for the asymptotic analysis of ergodic downlink rate per user terminal (UT) is derived for the spatially correlated system with Regularized Zero-Forcing (RZF) scheme. A realistic multipath environment consisting of a dominant line-of-sight (LOS) path along the non-deterministic non-line-of-sight (NLOS) component with imperfect channel information and pilot contamination is considered for deriving the expression. The system is evaluated for the carrier frequency of 30GHz in the mmWave range with different parameters using Monte Carlo simulations to validate the derived expression. The numerical results show that the derived expression for ergodic user rate provides a tight approximation for the large-dimensional system and accurate results for small values. As for the Rician Factor, K = 0.5, transmit power of 10dB and training power of 6dB, the average user rate obtained for M = 100 and K = 50 is 1.1233 bits/sec/Hz/User/cell for simulated value and 1.1243 bits/sec/Hz/User/cell for theoretical value. The proposed correlated multicellular mMIMO system is shifted from the mmWave range to sub - 6GHz to compare it with the existing system in (Sanguinetti et al. in IEEE Transactions on Communications 67:1939-1955, 2018). The derived results show the improvement in system throughput for the proposed system.