Analysis of Detection Methods in Massive MIMO Systems

Abstract

Large-scale Multiple Input Multiple Output (MIMO) system is significant in our daily life such as used in hundreds of antennas at the base station work together at the same time. Matrix inversion can be used to solve this problem and calculate the result of formulas in this system, but this method is difficult for some people, and it may easy to make mistakes. In this paper uses some low-complexity signal detection algorithm, which are the Gauss-Seidel (GS) method, successive over relaxation (SOR) method, Minimum mean square error (MMSE) signal detection, in order to avoid the complicated matrix inversion. Can also use a systolic array and show reference FPGA implementation results for various system configurations. Steepest descent algorithm is employed to obtain an efficient searching direction for the following Jacobi iteration to speed up convergence is also a method to solve this problem. Firstly, we have to prove a special property that the MMSE filtering matrix. Then, we prove the convergence of a low-complexity iterative signal detection algorithm. The result shows that the method can reduce the computational complexity from K3 to K2 (K is the number of users). Finally, the result of the experiment prove that the method is more advantage than the recently proposed Neumann series approximation algorithm, and achieved the near-optimal performance of the classical MMSE algorithm.