Tensor Train-Based Channel Estimation for RIS-Assisted mmWave MIMO Systems

In this letter, we consider a Reconfigurable intelligent surface (RIS)-assisted point to point cascade channel in millimeter wave multiple-input multiple-output orthogonal frequency division multiplexing (mmWave MIMO-OFDM) systems. Unlike most existing studies that adopt low-dimensional channel models, by exploiting the structure of the RIS-assisted MIMO channel across multiple dimensions, we model it in form of a low-rank higher-order CANDECOMP/PARAFAC (CP) tensor. Motivated by the inherent Vandermonde structure and higher-order low-rank characteristics of the tensor, we exploit Tensor-train (TT)-based processing to efficiently estimate the RIS-assisted channel. Numerical results confirm that the proposed method outperforms existing approaches in terms of both estimation accuracy and computational complexity.