Joint Phase-Noise and Channel Estimation in mmWave Massive MIMO Systems With Hybrid Structures Using Nested Tensor Decomposition

This paper focuses on the phase-noise (PN) and channel estimation for mmWave massive MIMO systems with hybrid analog-digital structures. By considering the PN caused by the impairment of radio frequency (RF) chains, we construct the training signal as a nested CANDECOMP/PARAFAC (CP) model. After separating PN and compressed channel (inner tensor) through CP decomposition (CPD) of outer tensor, we further explore the Vandermonde structure of the factor matrix in inner tensor and propose Tensor-Train-Vandermonde-Structure-CPD (TTVSCPD) to estimate mmWave channel parameters. The simulation results verify that the accuracy and robustness of proposed algorithm are superior to traditional methods.