Low complexity precoding schemes for massive MIMO systems

Abstract

In order to deal with the challenges of the exponentially growing communication traffic and spectrum bands with wider bandwidth, massive MIMO technology was been proposed, which employs an unprecedented number of base station antennas simultaneously to serve a smaller number of user terminals in the same channel. Although the very large antenna arrays for massive multiple-input multiple-output (MIMO) systems lead to unprecedented data throughputs and beamforming gains to meet these data traffic demands, they also lead to prohibitively high energy consumption and hardware complexity. In terms of precoding schemes, the conventional linear precoding entirely processes the complex signals in the digital domain and then upconverts to the carrier frequency after passing through radio frequency (RF) chains, which can achieve near-optimal performance with the large antenna arrays. However, it is infeasible because with fully digital precoding, every antenna element needs to be coupled with one RF chain, including the digital-toanalog convertors, mixers and filters, which is accountable for excessively high hardware cost and power consumption. This thesis focuses on the design and analysis of low complexity precoding schemes. The novel contributions in this thesis are presented in three sections. First, a low complexity hybrid precoding scheme is proposed for the downlink transmission of massive multi-user MIMO systems with a finite dimensional channel model. By analysing the structure of the channel model, the beamsteering codebooks are combined with extracting the phase of the conjugate transpose of the fast fading matrix to design the RF precoder, which thereby harvests the large array gain achieved by an unprecedented number of base station antennas. Then a baseband precoder is designed based on the equivalent channel with zero forcing (ZF) precoding. In addition, a tight upper bound on the spectral efficiency is derived and the performance of hybrid precoding is investigated. Second, based on successive refinement, a new iterative hybrid precoding scheme is proposed with a sub-connected architecture for mmWave MIMO systems.In each iteration, the first step is to design the RF precoder and the second step is to design the baseband precoder. The RF precoder is regarded as an input to update the baseband precoder until the stopping criterion is triggered. Phase extraction is used to obtain the RF precoder and then the baseband precoder is optimized by the orthogonal property. This algorithm effectively optimizes the hybrid precoders and reduces the hardware complexity with sub-connected architecture. A closed-form expression of upper bound for the spectral efficiency is derived and the energy efficiency and the complexity of the proposed hybrid precoding scheme are analyzed. Finally, the use of low-resolution digital-to-analog converters (DACs) for each antenna and RF chain is considered. Moreover, in a more practical scenario, the hardware mismatch between the uplink and the downlink for the channel matrix is a focus, where the downlink is not the transpose of the uplink in time-division duplex mode. The impact of one-bit DACs on linear precoding is studied for the massive MIMO systems with hardware mismatch. Using the Bussgang theorem and random matrix theorem, a closed-form expression for the signal to quantization, interference and noise ratio with consideration of hardware mismatch and one-bit ZF precoding is derived, which can be used to derive the achiev- able rate. Then a performance approximation is also derived in the high signal-to-noise ratio (SNR) region, which is related to the ratio of the number of base station antennas and the number of mobile users , and the statistics of the circuit gains at the base station. In conclusion, analytical and numerical results show that the proposed techniques are able to achieve close-to-optimal performances with low hardware complexity, thus the low complexity precoding schemes can be valid candidates for practical implementations of modern communication systems.