Single-Vector-Sensor Blind-Identification for Noncircular Signals in the Presence of Colored-Noise

The presence of colored-noise (the noise at different elements are correlated and/or have distinct variances) is often the case in vector-sensor array signal processing. In this paper, we develop two types of blind identification algorithms to extract the desired noncircular signals and/or their DOA´s from data contaminated by uncorrelated interferences and correlated interferences, respectively, in the presence of colored-noise. The former is based on the second-order-statistics (SOS) only, whereas the latter involves both the SOS and the higher-order-statistics (HOS). The desired steering-vectors or array response are blindly estimated via regularized ESPRIT in both blind identification algorithms. Simulation results are given to evaluate the proposed blind identification algorithms and compare them with the existing pure-HOS-based counterpart