Tensor-Based Blind Structured Channel Estimation for Multichannel Systems

In this work, we propose a blind channel estimation method in multiple-input multiple-output systems using a tensor decomposition approach. Consequently, fundamental link between convolutive channels and block-term decomposition (BTD) is established. The proposed approach leverages the second-order statistics of the received signals to construct a third-order tensor by stacking covariance matrices at different time lags. Then, the channel estimation process consists of two stages. In the first stage, the tensor is decoupled using the type-2 BTD technique to extract the loading factors. In the second stage, a Toeplitz constraint is imposed on the loading factors to obtain the channel matrix. The loading factors are constraint to be identical and have a Toeplitz structure. The numerical simulations show the effectiveness of the proposed method.