Research On Massive MIMO Signal Detection Algorithm Based On Group Sparse And Low Rank Tensor Decomposition

Massive multiple input multiple output(MIMO)system can significantly improve the energy efficiency and spectrum utilization of the system by equipping a large number of antennas in a base station(BS).However,with the increase of ultra-dense cells and a large number of users,eliminating inter-cell interference and multi-user interference has become a research hotspot of massive MIMO systems.Tensor decomposition as a learning algorithm for processing multidimensional data has been widely used in wireless communication systems.In this paper,the following research is conducted on the interference problem in multi-user massive MIMO and multi-cell multi-user massive MIMO systems.Firstly,in order to solve the problem of inter-user channel interference in multi-user massive MIMO systems,this paper constructs a multi-dimensional tensor model for the received signal at the user end,and proposes a channel estimation algorithm based on group sparse and low rank tensor decomposition.The proposed algorithm divides the received signal tensor into multiple sub-tensor blocks,and uses the multi-linear signal classification algorithm to estimate the initial channel for each sub-tensor.According to the estimated channel characteristic parameters,the channels from the same source are clustered into groups by the K-means algorithm based on the channel multipath component distance,and the channel estimation model based on group sparse and low rank tensor decomposition is used to perform inter-user channel.Channel interference cancellation,resulting in an optimized estimated channel.The simulation results show that compared with the traditional estimation algorithm,the proposed algorithm can significantly improve the accuracy of channel estimation and the system combining rate.Secondly,in order to improve the signal detection accuracy of multi-user massive MIMO system,this paper constructs the base station-side received signal into a third-order tensor model,and proposes a large-scale MIMO system signal detectionalgorithm based on group sparse and low rank tensor decomposition.The algorithm uses the symbol matrix,channel tensor and precoding tensor sent by the user as three add-ons in the TUCKER-2 decomposition model,and uses the Kronecker product-based least squares method to utilize the kronecker product between the symbol matrix and the channel matrix.The structure can be combined with the channel matrix to obtain an estimated symbol matrix.The simulation results show that compared with other traditional detection algorithms,the proposed algorithm can greatly improve the accuracy of signal detection and reduce the bit error rate.Good system performance is maintained in different modulation modes,and the effect is more obvious as the number of antennas increases.Finally,for the inter-cell interference and inter-user interference existing in the multi-cell multi-user massive MIMO system,the multi-dimensional tensor modeling of the signal received by each cell base station is proposed,and the group sparse and low rank tensor decomposition is proposed.Multi-cell large-scale MIMO signal detection algorithm.The algorithm uses the least squares method and the generalized TUCKER-(2,3)decomposition model to identify and estimate the channel parameter matrix in a closed form.User sources from different cells are clustered into groups based on estimated channel parameters.On this basis,the data symbols are estimated using the low rank tensor TUCKER-(2,3)decomposition algorithm.The simulation results show that the proposed algorithm exhibits good system performance under different modulation modes and antenna configurations,and has better system performance than the traditional interference cancellation algorithm.