A Parafac-Based Blind Channel Estimation and Symbol Detection Scheme for Massive MIMO Systems

In this paper, a multi-user massive multiple-input and multiple-output (MIMO) uplink system is considered, in which multiple single antenna users communicate with a target BS equipped with a large antenna array. We assume both the BS and K users have no knowledge of channel statement information. For such a system, by utilizing the unique factorization of three-way tensors, we proposed a parafac-based blind channel estimation and symbol detection scheme for the massive MIMO system, the proposed system can ensure the unique identification of the channel matrix and symbol matrix in a noise-free case. In a noisy case, a novel fitting algorithm called constrained bilinear alternating least squares is proposed to efficiently estimate the channel matrix and symbols. Numerical simulation results illustrate that the proposed scheme has a superior bit error ratio and normalized mean square error performance than traditional least square method. In addition, it has a faster convergence speed than typical alternation least square fitting algorithm.