Research On Low Complexity Signal Detection Algorithms In Massive MIMO System

Massive Multiple Input Multiple Output(Massive MIMO)technology is one of the key technologies of the 5th-generation mobile communication system.Compared with traditional MIMO technology,it adds more antennas and improves the system capacity,spectrum efficiency,data transmission rate and transmission reliability.However,due to more and more antennas and terminal equipment join in the system,the computational complexity of the signal detection is relatively high,and it is difficult to implement,which is one of the difficult problems in the study.The thesis mainly focuses on the low-complexity algorithm of signal detection in Massive MIMO system.The thesis first studies the signal linear detection algorithm of the Massive MIMO system.Pointing at the Minimum Mean Square Error(MMSE)algorithm in the linear detection algorithm,which has near-optimal detection performance,but the shortcomings of the algorithm's high complexity,the thesis researches and proposes a Jacobi iterative signal detection algorithm based on dynamic weighting factors,which avoids the process of high-dimensional matrix inversion by transforming the signal detection problem into solving the linear equations problem,and sets a weighting factor according to the optimization theory dynamically update the weighting factor.The simulation results show that the algorithm can converge with a small number of iterations,and the Bit Error Rate(BER)performance can approach the direct inversion MMSE algorithm under the premise that the algorithm complexity is kept at O(K~2).Next,in view of the relatively high computational complexity of the Jacobi iterative signal detection algorithm based on dynamic weighting factors in the signal detection algorithm based on solving linear equations,the thesis researches and proposes a Kaczmarz iterative signal detection algorithm based on the residual homogeneous ordering.This algorithm processes the filter matrix through a residual homogenization sorting method,so that the algorithm preferentially selects the hyperplane equations with greater correlation in the iterative process,and uses the channel hardening characteristics of the massive MIMO system to perform the initial value reasonable estimation and selection of appropriate relaxation parameters are provided,so as to further improve the convergence rate of the algorithm while reducing the complexity and speed up the convergence process.The simulation results show that the computational complexity of the algorithm is lower,and it can converge in a small number of iterations.At the same time,it has a BER performance close to that of the MMSE algorithm.Finally,in the scenario where the number of antennas of the Massive MIMO system is very large or even tends to infinity,a theoretical solution is given to the influence brought by the geometric approximation principle of the Kaczmarz algorithm,and the filter matrix is sampled reasonably by the probability random sampling theory to speed up convergence rate.