PARAFAC Modeling/Estimation of Time-Varying Space-Time Multipath Channels

In this paper, we address the problem of multipath parameter estimation of time-varying space-time wireless channels based on PARAFAC modeling. We exploit the different varying rates in the structure of the multipath channel over multiple data-blocks to build a 3D tensor from the received signal. The proposed estimation methods make use of a training sequence that is periodically extended over multiple data-blocks, across which the fading amplitudes are assumed to vary. A PARAFAC-based estimator using the ALS algorithm is proposed for joint recovering of the directions of arrival, time delays and complex amplitudes of the multipaths. Simulation results are provided to demonstrate the accuracy of the proposed approach under different channel conditions.