Research On Large Scale MIMO Precoding Algorithms

MIMO technology is becoming mature, and incorporated into emerging wireless broadbandstandards. Basically, the more antennas the transmitter/receiver is equipped with, the moredegrees of freedom the propagation channel can provide, the better the performance is in termsof data rate or link reliability. Thus, very large MIMO systems, also known as massive MIMO orlarge-scale antennas systems turn to be a new research field.Further research is made in this thesis on the improvements of the system performance thatlinear pre-coding methods can bring about, some useful information and numerical results areobtained with the help of computer-aided simulations.Firstly, the article analyses the channel capacity of large scale MIMO system, includinguplink and downlink. And then according to the very feature of the system (large-scale antennas),some linear precoding methods are introduced, such as Zero-Forcing (ZF), Minimum MeanSquare Error (MMSE) and Signal to Leakage and Noise Ratio (SLNR).Through the simulation,it shows that the precoding methods can indeed enhance the system performance and as thenumber of the antenna grows large enough, the three methods mentioned above almost have thesame performance.In addition, under the circumstances of multi-cell, the article focuses on the problem of pilotcontamination due to the increasing number of the antenna, why it can not be eliminated andhow it has a bad effect on the system.The analysis shows that with no negative effect caused byinterference and noise, pilot contamination becomes the primary factor that should beconsidered.Furthermore, the article provides an advanced MMSE precoding method that mitigates thepilot contamination problem.In addition to being linear, this method perform better than the ZFprecoding and even much better when the number of the antenna becomes larger.