PARAFAC-Based Channel Estimation for Relay Assisted mmWave Massive MIMO Systems

Millimeter wave (mmWave) communication has been regarded as the core technology with great potential in the fifth generation (5G) and the next generation mobile cellular networks. For relay assisted mmWave massive multiple input multiple output (MIMO) systems, we propose a tensor-based channel estimation algorithm which can reduce the pilot overhead. In the proposed method, the received signal is formulated into a parallel factor (PARAFAC) tensor model at the destination node, and a channel estimation algorithm based on PARAFAC decomposition is proposed. In addition, we also study the influence of the form of relay matrix on the performance of the proposed algorithm. The proposed algorithm needs a small number of iterations till convergence and has high estimation accuracy with less pilot overhead. Numerical examples verify the effectiveness of the proposed PARAFAC decomposition-based channel estimation method.