Parameterized Channel Estimation Considering Phase Deviations of the Multi-Antenna Array

A multi-antenna communication system usually experiences distinct phase deviations across the antennas, which can severely affect the performance of parametric channel estimation. But this problem has long been ignored in the literature. In this paper, we take the phase deviations into consideration and parameterize a time-varying multi-antenna channel by the multipath gains, the angle-of-arrivals (AoAs), the time delays, the Doppler frequency offsets, and the phase deviations of the multi-antennas as well. To tackle the high-dimensional problem of parameter estimation, we propose to use Zadoff-Chu sequences as the pilot for reducing the dimensionality associated with the time-delays and the Doppler frequency shifts. Using an iterative procedure that alternately updates the phase deviations and the other channel parameters, we can accurately estimate the time-varying channel even in presence of severe phase deviations. Numerical examples validate the effectiveness of our method.