Research On Low-Complexity Detection Algorithm For Uplink Of Massive MIMO System

Large-scale MIMO technology has become a hot research in the field of wireless communication in recent years,which deploys a large number of antennas at the base station to serve multi-user at the same time to improve the spectrum efficiency and energy efficiency.However,the increase of the number of antennas also increases the complexity of the signal processing at the base station.Therefore,this thesis studies the current algorithm to solve this problem,and proposes an improvement scheme based on Jacobi iterative and approximate message passing algorithm.The main results are summarized as follows.1.This thesis analyzes the widely studied Neumann algorithm and well-performance Richardson algorithm,RI algorithm and Jacobi algorithm.The performance,complexity and iterative coefficients of the algorithm are analyzed by simulation,which provides the entry point for the researching of this thesis.2.The simulation results show that the Jacobi iterative algorithm in the linear algorithm is not affected by the iterative coefficients,and the performance is stable.So the application is strong.But it has the strict convergence condition and the poor convergence rate,so two optimization methods are proposed to accelerate convergence.First,the weighted average method is proposed to accelerate the Jacobian iteration,and the acceleration factor is introduced.Then the conditions for satisfying iterative convergence are deduced,and the value of w is presented when the fastest convergence is reached.Second,combined with the current method of space-based initial solution and the nature of the iterative algorithm,the optimized method is given to make the range of the initial solution more accurate,so as to achieve faster convergence rate.3.The deduction and simplification of approximate message passing algorithm is presented.First,the determination of the statistical characteristics of the message is simplified when the message is updated.In this thesis,the constant v is used to replace the variance in the factor and the variable,then the variance update is a simple arithmetic mean.The second is the simplification of the sequence function in the iterative equation.Assuming that the priori distribution is Gaussian distribution,the sequence function is expressed as a linear estimator.And then the theory in the relevant literature is used to calculate the parameters of the function,and finally the simplified iterative equation is presented.