Low Complexity Detection Algorithms In Massive MIMO Systems

Massive MIMO(Massive Multiple-input Multiple-output, Massive MIMO) is one of the evolution directions of MIMO Technology. The number of antennas at the base station is extended to tens or even hundreds to provide a higher diversity gain and multiplexing gain, which can greatly improve the channel capacity and spectrum utilization. However, with the increase of the number of antennas, the original receiver signal detection algorithms can not meet the communication standard because of its high complexity and poor performance. This thesis mainly introduces several detection algorithms with good performance and low complexity in Massive MIMO system.Firstly, the common detection algorithms in MIMO is introduced into the Massive MIMO system, and it is found that the conventional linear detection algorithm can achieve very good performance when the number of the antennas is much larger than the number of users. Then,this thesis introducs the basic principle and the realization method of two Massive MIMO detection algorithms, the likelihood ascent search(LAS) algorithm and the reactive tabu search(RTS) algorithm and analyses the effect of the initial solutions, the number of antennas and the number of iterations on the performance of the algorithms.Then, a variety of LAS and RTS optimization schemes are introduced,and the simulation results and analysis of the bit error rate(BER) performance and complexity of the above algorithms are also presented in the thesis.A LAS optimization scheme which is based on random list is proposed.This algorithm has better simulation performance and lower complexity than other optimization algorithms.As the problem of poor performance when the two algorithms appplyed to high order modulation, the QP algorithm is brought into the Massive MIMO detection.2QP and BB(L,M) algorithms are proposed which can obtain better performance when adopting the high order modulation. According to the the signal detection simulation results of the two algorithms,they have good BER performance and low complexity. Finally, several feasible suggestions are put forward for the practical applications.