Research On Channel Estimation And Equalization In 5G Communication System

The many advantages of Massive MIMO technology make it one of the key technologies of 5G communication.With the increase of the antenna,the dimension of the channel matrix is getting higher and higher,and the channel state infomation will be more complex.This is a higher requirement for the channel estimation and channel equalization.In this paper,we study three pilot based channel estimation methods,in which the Bayesian channel estimation method takes advantage of the statistical characteristics of the channel,such as AOA,and it has good anti pilot pollution capability.Finally,the performance of the LS estimation is compared with the simulation experiment.The experimental results show that the Bayesian channel estimation method has a better anti pilot pollution ability.The channel estimation method based on pilot is simpler,but the pilot pollution will seriously restrict the performance of the system.In order to solve the problem of pilot overhead,a potential choice is to use the sparsity of the wireless channel and use the sparse channel estimation method.Then the compressed sensing theory and recovery algorithms,analyzes characteristics of Massive MIMO system channel beam domain,based on the compressed sensing domain channel estimation method on the beam is put forward,and gives a deterministic pilot matrix design and M-SP algorithm based on ZC sequence.According to the simulation results,the performance of this paper to determine the measurement matrix close to or even better than the random measurement matrix,M-SP algorithm can improve the performance of the SP algorithm to a great extent,improved the disadvantages of SP algorithm to approximate the sparse signal estimation in the performance of the LS channel and beam domain channel estimator proposed in this paper is superior to the traditional the estimated.The Massive MIMO channel equalization technique is divided into linear equilibrium and nonlinear equilibrium.The linear equilibrium algorithm generates an approximate estimate of the transmitted signal by linear conversion of the received vector.The commonly used linear equalization techniques in Massive MIMO systems include ZF,MMSE and OSIC.While not directly to the nonlinear equalization of the received vector of linear transformation,requires a combination of some special treatment,such as SNR sorting and approximate orthogonalization processing etc.,so it has relative high complexity,nonlinear equalization methods are commonly used,interference cancellation(SIC),the equilibrium lattice reduction equilibrium(LR)etc.