Resource Optimization Strategy In Cell-free Massive MIMO

Cell-free massive MIMO eliminates the concept of cellular cells,the macro-space diversity gain is increased by deploying a large number of low-cost and low-power access points,and there is no need to share channel state information,facilitating flexible deployment.The related technologies of cell-free network were proposed not long ago,and it is still in the exploratory stage.This thesis has carried out the following research work on precoding and access point selection in cell-free massive MIMO.In order to reduce the complexity of precoding algorithms in cell-free massive MIMO,two precoding algorithms are proposed.The first is based on the Neumann series precoding algorithm,which uses the Neumann series polynomial expansion to approximate the inverse matrix of the linear precoding algorithm,thus avoiding the matrix inversion process and reducing the complexity of the matrix inversion to a certain extent.The second is the CSM(Cholesky and Sherman-Morrison strategy)-based precoding algorithm,which uses Cholesky decomposition and Sherman-Morrison formula to iteratively solve the inverse matrix of the matrix.So this algorithm can realize the direct calculation of matrix inversion,decrease the performance loss caused by the approximate replacement of Neumann series and reduce the complexity of the system.Then,in view of the problems that still exist in the system after the precoding algorithm is processed,the power minimization power control is introduced to optimize the system power control coefficient and reduce the influence of channel noise.The experimental simulation results show that joining power minimization power control,the precoding algorithm based on Neumann series and the CSM precoding algorithm can keep the spectrum efficiency stable while reducing the complexity of the precoding algorithm.To solve the problems of high backhaul link load and pilot pollution in cell-free massive MIMO,a joint basic genetic algorithm access point selection and greedy pilot allocation scheme is proposed.In this scheme,first,we applies the genetic algorithm based access point selection strategy to the cell-free massive MIMO system.The strategy uses the average rate of system users as a performance indicator,and selects a suitable subset of access points for each user through operations such as selection,crossover,mutation,and iteration,which greatly increases the average rate of system users and reduces the backhaul link load of the system.Then,in the uplink pilot training phase,we use the greedy pilot allocation algorithm which aims at minimizing the pilot pollution effect of the user with the smallest rate and optimize pilot allocation by iterative method,thereby reducing the impact of pilot pollution.The simulation results show that the researched scheme can increase the rate of users polluted by the pilots and the sum rate of the entire system.