PARAFAC-Based Channel Estimation and Positioning in Millimeter-Wave Massive MIMO-OFDM Systems

In millimeter wave (mmWave) massive MIMO-OFDM systems, our focus is on downlink channel estimation. These systems employ extensive planar antenna arrays for precise beamforming. In contrast to the narrowband model of channel studies, our attention centers on the estimation of wideband mmWave models featuring frequency selectivity, ideally suited for the selected systems. Capitalizing on the sparse scattering exhibited by mmWave models, we present a PARAFAC-based approach to estimate pivotal channel parameters, including angles of arrival and departure, time delays, and fading coefficients that could be used for positioning and mapping. Our proposed method leverages a third-order tensor representation of the received signal at the transmitter. This tensor representation employs a low-rank CANDECOMP/PARAFAC decomposition (CPD) structure. This allows estimation of channel parameters from factor matrices, while ensuring CPD uniqueness even for smaller tensors. Simulation results validate our method's efficacy and performance enhancement for mmWave massive MIMO-OFDM channel estimation.