Analysis Of Precoding Technologies For Massive MIMO Systems

Massive Multiple-input Multiple-output (MIMO) technology is one of the novel technologies in wireless communications systems. By deploying tens to hundreds of antennas at the base station (BS), massive MIMO technology can achieve promising capacity and spectral efficiency. However, there are lots of data streams in the system using massive MIMO technology (i.e., massive MIMO system), the interference in the massive MIMO system is tremendous. Hence, applying certain signal processing technologies to such systems to sup-press the interference is necessary. Moreover, since massive MIMO systems are with large scale, it is not feasible to put the signal processing technologies at each users. As a result, precoding technology becomes the key technology to suppress the interference.Based on the circumstances, this thesis is focusing on the researches on the precoding technologies in massive MIMO systems. The main contributions of this thesis are as follow.Firstly, this thesis gives an intuitional and insightful explanation to Energy..Penalty problem of linear precoding technologies in massive MIMO systems. Then, the thesis extends the derivations into more practical scenarios, such as, more kinds of precoding technologies, and more practical channel models. Specifically, the thesis calculates the asymptotic expressions of downlink in massive MIMO system under three different precoding technologies, namely, Zero-forcing (ZF) precoding, Regularized ZF (RZF) precoding, and Singu-lar Value Decomposition (SVD) based precoding. And it is also pointed out that the sum-rate of the three precoding technologies decreases with increasing number of served users when the transmit power is limited.Secondly, this thesis proposed an off-line algorithm to determine the op-timal number of users served in the downlink of massive MIMO system. The sum-rate in the downlink of massive MIMO system is maximized when serv- ing such number of users. Additionally, assuming M and N are the number of antennas at BS and all users respectively, making the system fully loaded (M=K) is not the optimal strategy when using linear precoding technologies. A considerable sum-rate gain can be obtained by adding a few extra antennas at BS when the system is fully loaded.Finally, taking the bad performance of the linear precoding technologies when the system is full loaded, a novel non-linear precoding technology is pro-posed in this thesis. The proposed non-linear precoding technology is based on vector precoding, termed as Oriented Reactive Tabu Search (ORTS) precod-ing. The proposed achieves a good tradeoff between performance and com-putational overheads, and sharing almost the same diversity gain with Lattice Reduction (LR) precoding.