Research On Precoding Algorithms Of Massive MIMO For 5G Mobile Communications

With the dramatic development of mobile communication system from the 1st generation to the 4th generation, mobile communications have brought tremendous convenience for our daily life. Meanwhile, the research on mobile communication system never stops. Studies on the 5th generation mobile communication system are now in process. Massive MIMO has gained huge consideration in 5G due to high achievable data rates, redundancy of the system and lower power consumption. As the antenna number is increasing, precoding algorithm has become hot topic thanks to their simplification of receivers, improvement of spectral efficiency and capacity.This thesis studies massive MIMO precoding algorithms for the 5th generation mobile communication system thoroughly in terms of linear and nonlinear algorithms. For the linear precoding, maximum ratio transmitting(MRT), zero forcing(ZF), minimum mean square error(MMSE), and signal leakage noise ratio(SLNR) are analyzed. Then, bit error rate(BER) and sum rates of them are simulated, which shows that MMSE performs better than ZF and MRT. Later, this thesis compares the signal to interference ratio(SIR) of the linear precoding in multi-cell circumstance, and verifies the theoretical results by simulations. For the nonlinear precoding, dirty paper coding(DPC), Tomlinson-Harashima precoding(THP), and vector precoding(VP) are investigated, also MMSE-THP, ZF-THP and VP are compared by simulations, which demonstrates the merits of nolinear precoding. Lastly, an improved precoding algorithm is proposed, which takes the block error rate(BLER) into consideration. The proposed new algorithm modifies the RBD for the massive MIMO system based on OFDM. Simulations show that the proposed algorithm has lower BLER compared with the traditional linear precoding, also is applicable for multiple resource blocks(RBs).