Massive MIMO in Sub-6 GHz and mmWave: Physical, Practical, and Use-Case Differences

Abstract

The use of base stations (BSs) and access points (APs) with a large number of antennas, called Massive MIMO (multiple-input multiple-output), is a key technology for increasing the capacity of 5G networks and beyond. While originally conceived for conventional sub-6 GHz frequencies, Massive MIMO (mMIMO) is ideal also for frequency bands in the range 30-300 GHz, known as millimeter wave (mmWave). Despite conceptual similarities, the way in which mMIMO can be exploited in these bands is radically different, due to their specific propagation behaviors and hardware characteristics. This paper reviews these differences and their implications, while dispelling common misunderstandings. Building on this foundation, we suggest appropriate signal processing schemes and use cases to efficiently exploit mMIMO in both frequency bands. * E. Björnson is with Linköping University, lower frequencies, which is achieved by mMIMO. There are, however, fundamental differences between how mMIMO technology can be designed, implemented, and exploited in sub-6 GHz and mmWave bands. In this paper, we provide a comparative overview, highlighting three main differences:1. The propagation channels build on the same physics, but basic phenomena such as diffraction, attenuation, and Fresnel zones are substantially different.2. The hardware implementation architecture changes with the increasing carrier frequency. More antennas can be integrated into a given area, but the insertion losses, intrinsic power-overhead in radio-frequency (RF) generation, and amplification result in diminishing gains.3. The signal processing algorithms depend on propagation and hardware. Channel estimation is resource-demanding at sub-6 GHz, while beamforming is straightforward. Conversely, mmWave channel estimation and beamforming are theoretically simpler since there are fewer propagation paths, but become challenging if hybrid beamforming is used.In the remainder of this article, we elaborate on these differences, including that they manifest how sub-6 GHz and mmWave bands are to be exploited to target different use-cases in 5G and beyond. Difference I: The propagation channelAn understanding of the electromagnetic propagation is crucial when considering mMIMO systems and frequencies up to mmWave bands. The channels behave fundamentally different from what we are used to in cellular networks, which exposes weaknesses in the channel modeling simplifications conventionally made.

Efficient Angle-Domain Processing for FDD-Based Cell-Free Massive MIMO Systems

“…An attractive alternative to consider is frequency-division duplexing (FDD) based cell-free massive MIMO systems for the following reasons: 1) channel reciprocity in TDD mode might not be accurate due to calibration errors in radio frequency (RF) chains [12], 2) with the lack of downlink training symbols in TDD systems, users may not be able to acquire instantaneous CSI, and thus system performance will deteriorate in detecting and decoding the intended signals, 3) while TDD operation is preferable at sub-6 GHz massive MIMO, in millimeter wave (mmWave) bands FDD may be equally good since the angular parameters of the channel are reciprocal over a wide bandwidth [13], and 4) FDD systems dominate current wireless communications and have arXiv:2001.07438v1 [cs.IT] 21 Jan 2020 many benefits such as lower cost and greater coverage than TDD [14].…”

Deep Learning Design for Joint Antenna Selection and Hybrid Beamforming in Massive MIMO

“…Nearly all of the aforementioned methods strongly rely on perfect CSI knowledge. This is very impractical given the highly dynamic nature of mm-Wave channel [17]. To relax this dependence and obtain robust performance against the imperfections in the estimated channel matrix, we examine a deep learning (DL) approach.…”

“…In [42], manifold optimization or "Manopt" algorithm is suggested to effectively solve the optimization problems in (17) and (25). Note that both of these problems do not require a codebook or a set of array response of transmit and receive arrays [8].…”

“…For the prediction process, we generated J T Monte Carlo experiments where a test data which is separately generated by adding noise on received pilot signal with SNR defined as SNR N−TEST is used. Note that this operation is applied to corrupt input data and test the network against deviations in the input data which can resemble the changes in the channel matrix due to short coherence times in mm-Wave scenario [17].…”

Real-Time Implementation Aspects of Large Intelligent Surfaces

“…Relative to fourthgeneration systems, this has resulted in an order-of-magnitude increase in the system spectral efficiency. Since the inception of massive MIMO in 2010 [5], phenomenal progress has been made in order to understand its theoretical and implementation aspects (see e.g., [2,3,[5][6][7][8][9][10] for a taxonomy). To this end, our understanding of massive MIMO systems has greatly matured over almost a decade.…”