Research On Low Complexity Linear Precoding Algorithm For Massive MIMO

Massive MIMO technology can achieve higher capacity,spectrum efficiency and energy efficiency by equipping hundreds of antennas at the base station.These improvements make it one of the key techniques for 5G communication systems.However,performance and reliability are the critical limit.The fact that multi-user share the same spectrum resources lead to the signal interference.It becomes more serious when multiple signals transmit at the same time.In order to solve this problem,we need to use signal processing technology to suppress the interference signal.In Massive MIMO system,using detection technique to suppress the interference is not feasible in downlink system,because of the large number of antennas in base station and the limitations of the users on the signal processing.Thus,utilizing precoding technology at base station to reduce the impact of interference on the system has become the mainstream method in Massive MIMO systems.This thesis focus on the researches of linear precoding algorithms in massive MIMO.The main contents of this thesis are as follows:(1)We analyzed a series of system models and their channel capacities,including the Single-user MIMO and Multi-user MIMO.Moreover,the analysis of the propagation model and channel capacity of Massive MIMO system is the main studying aspect.(2)We studied the existing linear precoding technology in Massive MIMO system,including Matched Filtering(MF)precoding algorithm and Zero Forcing(ZF)precoding algorithm.System capacity of this two precoding algorithms is analyzed and compared.Considering the normalization of precoding matrix,the two different normalization methods are analyzed respectively on the two kinds of precoding systems for the capacity.The result shows that matrix normalization is better for MF while vector normalization is better for ZF.Assuming that the sum rate capacity are equal,the two algorithms are analyzed and compared on the transmission power.The results show that ZF requires smaller transmission power to achieve higher data transmission rate compared to MF.In addition,we also compare and analyze the computation complexity of this two precoding algorithms.The simulation results show that the complexity of ZF precoding algorithm is much higher than MF.(3)We proposed a low complexity precoding algorithm.The number of eliminated inter-users interference is adjusting by the interference cancellation coefficient in this algorithm,which uses orthonormal unit vector and orthogonal vectors to eliminate the inter-users interference.It is effective to enhance the overall system performance and reduce the algorithm complexity.In this paper,we simulate proposed precoding algorithm and compare it with of Matched Filtering(MF)and Zero Forcing(ZF)precoding algorithm on the precoding performance by established experimental simulation platform.The results show that the proposed algorithm can reduce computation complexity compared to ZF.