Decoding Method And Performance Analysis Of MRC In Massive MIMO

A Multiple-Input, Multiple-Output (MIMO) technique is a wireless transceiver technology of high efficiency. It can increase spectrum efficiency and improve performance of the system, and it has received extensive attention and research in both academic and industrial circles. However, due to complexity of wireless channel, and interference of some unknown factors, how to decode signals correctly at the receiving terminal plays an important role in the performance of the system. The Maximum-Likelihood (ML) decoder is the optimal decoding method, and it has the minimal probability of error. But its decoding complexity is extremely high, especially in the Massive MIMO system where a base station is with a large scale of antennas. Both Zero-Forcing (ZF) decoder and Minimal Mean-Square Error (MMSE) decoder have low decoding complexity, and only need is to calculate the inverse matrix of the channel matrix. In the Massive MIMO system, ones seek decoders with lower decoding complexity, and hence, the Maximum Ratio Combining (MRC) decoder attracts people's attention.This thesis aims at a single cell MIMO uplink system. Assuming that the base station is equipped with an M receive antennas and K different users at the cell, and each user has only one antenna. Also assuming that the channel is slow flat fading and the base station knows channel state information. Thesis compares the performances of ML?ZF and MMSE decoders. The comparisons show that ML decoder has the best performance, while MMSE decoder is superior to ZF decoder. But their performances are improved with the increase of the antennas. Based on the technique from Euclidean space and sphere, thesis proves some important formulas. These formulas present angle distributions of two vectors and a vector with the space, which will play an important role for us to solve problems from Massive MIMO system.For the above system model, thesis pays attention to the performance of the MRC decoder. Firstly, the MRC decoder is defined exactly for the system. Then, a formula calculating the Pair-Wise Error Probability (PEP) of the system with the decoder is derived. Furthermore, asymptotic analysis of the PEP based on two different situations is given. These analyses reveal the following facts:1) The PEP cannot go to zero even when Signal-to-Noise-Ratio (SNR) goes to infinity if the number of antennas is fixed. Thus, it has an error floor; 2) The PEP goes to zero when the number of antennas at the BS increases to infinity even if SNR is fixed. Finally, power scale law on PEP is discussed. Suppose that the base station equivalent received SNR is scaled asM M-a, and when M goes to infinity, the PEP goes to zero if and only if 0< a< 1. When a= 1, PEP will not go to zero, that is, the error floor occurs. The fact is different from the result shown in literatures. Numerical simulations confirm the above conclusions.