Multiuser Massive MIMO Uplink Channel Estimation Based On Independent Column-independent Factorized

Massive multiple-input multiple-output(MIMO)technology has the potential to significantly improve spectral efficiency and energy efficiency,and has become a research hotspot in the next generation of wireless communication technology.To fully explore the advantages of massive MIMO communication systems,it is necessary to obtain real-time and accurate channel state information(CSI).In time division duplex(TDD)mode,the reciprocity of channels means that the downlink channel from the base station to each user can be obtained by estimating the uplink channel.Therefore,multi user massive MIMO uplink channel estimation has attracted widespread discussion among scholars at home and abroad.Numerous studies have shown that due to limited local scattering effects in communication environments,massive MIMO channels have sparsity in the angle domain,which can greatly improve the performance of multi user massive MIMO channel estimation.However,utilizing sparsity for multi user massive MIMO uplink channel estimation is extremely challenging,as the coupling effect between the pilot matrix and the sparse signal makes it impossible to separate the sparse forms of each user channel.The least squares(LS)estimator is one of the commonly used methods to remove the coupling effect of the pilot matrix,but its performance heavily depends on the selection of the pilot.Under non orthogonal pilots,the output signal-to-noise ratio of the LS estimator will be significantly reduced,which will bring significant performance losses to the sparse signal recovery problem corresponding to each user.Another common decoupling method is to vectorization the received signal,but it will greatly increase the dimension of the dictionary matrix,resulting in a huge amount of computation.Therefore,the current multi user massive MIMO uplink channel estimation problem requires a more effective decoupling method to ensure the estimation accuracy of the uplink channel.In addition,the current multi user massive MIMO uplink channel estimation methods have limited applicability and generally consider Gaussian noise.However,actual communication systems are often affected by impulse noise,and ignoring the impact interference caused by impulse noise can seriously reduce the performance of multi user massive MIMO uplink channel estimation.In addition,in actual communication systems,arbitrary planar arrays are usually deployed on the base station side.Compared with uniform linear array(ULA),the array flow pattern of any planar array not only includes azimuth angle but also elevation angle,so a two-dimensional grid is needed to cover the entire angle domain.In order to maintain the same grid gap as the one-dimensional grid,the required grid points for the two-dimensional grid are much larger than those for the onedimensional grid,which brings about a problem of high computational complexity.At the same time,the highly correlated basis vectors in the dictionary matrix bring about serious pseudo peak problems.In response to the above issues,the main research content of this article is as follows:This paper proposes a multi user massive MIMO uplink channel estimation method based on independent column decomposition,utilizing the sparsity of massive MIMO channels to address the performance loss caused by decoupling pilots in LS estimators.By introducing the variable Bayesian inference(VBI)of independent column decomposition into the sparse Bayesian learning(SBL)framework,the method adaptively decouples the pilot matrix and sparse matrix,thus avoiding the performance loss caused by the use of LS estimator,and effectively recovering sparse signals under any pilot.The proposed method does not increase the dimension of dictionary matrix,so the computational complexity is far lower than that of the uplink channel estimation method using vectorization operation.The traditional SBL method assumes that the angle of arrival(Ao A)of the signal is located at grid points.However,in practice,the actual Ao A is random,and even dense grids make it difficult to make Ao A located at grid points,which leads to angle mismatch problems.This paper adopts an off network model that approximates the real Ao A through Linear approximation,which effectively alleviates this problem and significantly improves the channel estimation performance.For the limited applicability of current multi user uplink channel estimation methods,this paper proposes a multi user massive MIMO robust uplink channel estimation method suitable for any planar array under the influence of impulse noise.This article proposes a novel SBL framework to combat impact noise by utilizing channel sparsity and sparse representation of impact noise.This framework utilizes the sparsity of impact noise for modeling,and uses the VBI method of independent column decomposition to separate the sparse signals and the sparsity of impact noise,significantly reducing the impact of impact noise and significantly improving the estimation performance of massive MIMO uplink channels for multiple users.On the other hand,this article further expands the research scope from the deployed ULA on the base station side to any planar array antenna.The proposed method proposes a coarse non-uniformly sampled twodimensional grid to address the computational complexity and serious pseudo peak issues associated with traditional two-dimensional grids.The coarse grid greatly reduces the number of grid points,avoiding the huge computational burden of traditional two-dimensional grids.The non-uniformly sampled two-dimensional grid reduces the correlation of the basis vectors in the dictionary matrix and avoids serious pseudo peak problems.The simulation experimental results show that the channel estimation accuracy of the proposed method is superior to existing methods.