Research On Channel Estimation And Equalization Algorithm In 5G Mobile Communication System

With the general improvement in the quality of human life,there is a further outlook for the future development of science and technology,which also includes mobile communication technology.As candidates for 5G communication systems,UFMC and FBMC have the advantages of saving spectrum utilization and suppressing out-of-band leakage.At the same time,the traditional channel estimation method requires a large amount of pilot information and the reconstruction accuracy is not high.In order to solve this problem,the compressed sensing algorithm came into being and is widely used in various fields.This thesis mainly studies the application of compressed sensing theory in the channel estimation of UFMC and FBMC systems,combined with the channel equalization algorithm,and finally obtains the compensated signal.Compared with the traditional algorithm,the bit error rate of the compensation signal obtained by using the channel estimation algorithm proposed in this paper is lower than that of the original signal.In the third chapter of this thesis,we first introduced the SAMP algorithm,and found that the final selection of the SAMP algorithm's atomic set matrix and the residual signal correlation is not high,so this thesis proposes the SAMP algorithm based on the RSAMP algorithm.based on the SAMP algorithm.This algorithm is to introduce the idea of regularization into the SAMP algorithm,so that each iteration of the atoms has an additional layer of screening.The screening criterion is that the maximum value of the atoms and residuals in the selected atom set cannot exceed twice the minimum value.And to ensure the maximum energy.Such screening can further ensure that the last updated atom set is most relevant to the parameter to be estimated,thereby improving the reconstruction accuracy.In the fourth chapter of this thesis,the complete BCS algorithm is first deduced.During the derivation process,it is found that the traditional BCS algorithm can improve the iteration method when solving excessive parameters,and the recovery of the transition parameters directly affects the reconstruction of the parameters to be estimated.Therefore,this thesis draws on the idea of reserving candidate atoms and variable step size regularization in the greedy algorithm,and proposes two improved BCS algorithms.The BCS algorithm ends with the sparsity K,and the position of the non-zero component after the iteration is the non-zero position of the parameter to be estimated.This approach is not rigorous.In the DBCS algorithm,K positions are reserved as alternatives.After the iteration,K positions can be selected from 2K positions to form a transition parameter.Such multiple screening can improve the reconstruction accuracy;BCS algorithm only Select a position to be set to 1,and the RBCS algorithm proposed in this thesis sets M non-zero positions to 1 at each iteration,then selects L components from the M positions,and updates M after each iteration.This can ensure the selected The location is closer.Finally,the software platform simulation shows that the improved BCS algorithm has an overall performance improvement of 1 to 2dB.