DoA and Polarization Estimation for Arbitrary Array Configurations

In this paper, algorithms for joint high-resolution direction-of-arrival (DoA) and polarization estimation using real-world arrays with imperfections are proposed. Both azimuth and elevation angles are considered. Partially correlated and coherent signals may be handled as well. Unlike most of the work available in the open literature, we consider the case when polarization sensitive antenna arrays, which may be disposed on a conformal surface, may have unknown (but fixed) geometries, be composed of elements with individual beampatterns and be subject to cross-polarization as well as mounting platform effects. Herein, recent results on steering vector modeling from noise-corrupted array calibration measurements are employed. This allows for incorporating all nonidealities of an antenna array into the estimation algorithms in a general and convenient manner. The proposed estimators can be implemented using the fast Fourier transform or polynomial rooting techniques regardless of the array configuration. The stochastic Cramér-Rao lower bound for the estimation problem at hand is established using the results from array steering vector modeling as well. The proposed expression takes into account array nonidealities, making such a bound tight even for real-world arrays. Extensive simulation results are provided using several real-world antenna arrays. The proposed algorithms outperform conventional algorithms available in the literature and have a performance close to the stochastic Cramér-Rao lower bound.