Deterministic Equivalent Performance Analysis of Time-Varying Massive MIMO Systems (original) (raw)

2015, IEEE Transactions on Wireless Communications

Delayed channel state information at the transmitter (CSIT) due to time variation of the channel, coming from the users' relative movement with regard to the BS antennas, is an inevitable degrading performance factor in practical systems. Despite its importance, little attention has been paid to the literature of multi-cellular multiple-input massive multiple-output (MIMO) system by investigating only the maximal ratio combining (MRC) receiver and the maximum ratio transmission (MRT) precoder. Hence, the contribution of this work is designated by the performance analysis/comparison of/with more sophisticated linear techniques, i.e., a minimum-mean-square-error (MMSE) detector for the uplink and a regularized zero-forcing (RZF) precoder for the downlink are assessed. In particular, we derive the deterministic equivalents of the signal-to-interference-plus-noise ratios (SINRs), which capture the effect of delayed CSIT, and make the use of lengthy Monte Carlo simulations unnecessary. Furthermore, prediction of the current CSIT after applying a Wiener filter allows to evaluate the mitigation capabilities of MMSE and RZF. Numerical results depict that the proposed achievable SINRs (MMSE/RZF) are more efficient than simpler solutions (MRC/MRT) in delayed CSIT conditions, and yield a higher prediction at no special computational cost due to their deterministic nature. Nevertheless, it is shown that massive MIMO are preferable even in time-varying channel conditions. Index Terms-Massive MIMO, delayed CSIT, channel estimation, channel prediction, linear precoding, linear detection. I. INTRODUCTION Multiuser MIMO (MU-MIMO), applied to next generation systems (e.g., 802.16m [3], LTE-Advanced [4]), is one of the core technologies promising to provide a remarkable increase in data rates. Such systems include several co-channel users communicating with a base station (BS) equipped with multiple antennas. However, the technological transition to 5G systems is expected to demand a thousand-fold higher capacity. Massive MIMO, where the BS includes a very large number of antennas, have emerged as one of the most promising technologies towards this direction because more degrees of freedom and increased power efficiency are achieved by simplifying multiuser processing, reducing transmit power, as