Random Matrix Theory For The Uplink In DAS Based On Deficient Sample

In the uplink of the discrete antenna system(DAS),the sample covariance matrix of the received signal is deficient due to the limited signal of the antenna.In addition,because the communication environment is very complex,the model of channel state information(CSI)can not be established accurately.Based on these two questions,if the direct use of the estimated channel state information and the deficient sample covariance matrix,the matual information is calculated the system mutual information accurately and it is very difficult to design an optimal receiver.Increasing the number of samples can only improve the performance of the system to a certain extent,but the computation is also greatly increased.In order to solve the problems mentioned above,to eliminate the influence of deficient sample covariance matrix of the target signal,using fewer sampling number while maintaining performance when the number of sampling system,this paper will optimize the system based on the random matrix theory.Random matrix refers to the elements of the matrix at least one is random.In this paper,we mainly study the large dimension of random matrix.The large dimension of random matrix theory is that the observation dimension of matrix and sample of covariance matrix all tend to infinity,but its ratio is not too small.More applicable to the actual processing scene.Under these conditions,it is more suitable for the actual processing scenarios.We consider modal of linear receiver that conducts minimum mean-square error(MMSE),maximizes signal interference-noise ratio(SINR)and mutual information at the uplink of a DAS with channel fading or interferences.Counteracting the performance degradation of baseband signal caused by deficient sample size regime and keeping the optimality in large sample size have been expressed in this paper studies.However,the linear receiver at base station(BS)and channel capacity(maximum symbol rate)cannot be obtained and has to be estimated in practice since the channel state information is unknown as same as covariance matrix.In this paper,we focus on the need for estimating errors based on linear receiver which is constructed the estimated channel state information and sample support.Eliminating unknown channel state information and covariance matrix,several methods of random matrix theory(RMT)must be used in estimation.In addition,applying the results of Central Limit Theorem(CLT)for spectral statistics and exact separation of Large-dimensional covariance matrix,we proof asymptotic equivalence between deficient sample covariance and precise covariance.Meanwhile,adaptive diagonal loading(DL)expression for sample covariance matrix is calculated at distributed antenna system,obtaining optimal diagonal loading for sample covariance.