Channel Estimation in Millimeter Wave MIMO Systems: The Tensor-based Methods

This dissertation presents two novel tensor-based methods for solving channel estimation (CE) problems in Millimeter Wave (mmWave) multiple-input multiple-output (MIMO) wireless communication systems. First, we proposed a method of tensor rank regularization with bias compensation for CE in a hybrid mmWave MIMO system. We modified the CANDECOMP/PARAFAC(CP) decomposition-based method and jointly estimated the tensor rank and channel factor matrices. It differs from most existing works by assuming that the number of channel paths is unknown, yet it can accurately estimate channel parameters without prior knowledge of the number of multipath components. The tensor rank is estimated by a novel sparsity-promoting prior that is incorporated into a standard alternating least squares (ALS) function. We introduced a weighting parameter to control the impact of the previous estimate and the tensor rank estimation bias compensation in the regularized ALS. The channel information is then extracted from the estimated factor matrices. Secondly, we proposed a CE framework based on the tensor decomposition method to address the challenge of slow-moving CE in a reconfigurable intelligent surface (RIS)-aided MIMO communications system. We formulated the uplink training signals as a third-order tensor, which admits a CP model with three factor matrices containing the channel state information (CSI) and phase information of the RIS. We further utilized the third factor matrix featured in the Vandermonde structure and decomposed the received tensor signals with a higher rank into three factor matrices. Additionally, we developed algorithms to estimate parameters of the low-velocity user terminal (UT)-RIS channel from the resultant factor matrices. The results of this dissertation contribute to the body of knowledge in wireless communications. Advisors: Lamar Yaoqing Yang and Andrew Harms