ROBUST HIGH-RESOLUTION DIRECTION-OF-ARRIVAL ESTIMATION VIA SIGNAL EIGENVECTOR DOMAIN

We present a new robust high-resolution direction-of-arrival (DOA) est imation approach for coherent/noncoherent sources. The proposed approa ch is based on the fact that the signal eigenvectors of the covariance matrix are a linear combination of the direction vectors which contai n the DOA information. By applying a high-resolution frequency estimat ion algorithm to an element sequence of a combination of the signal ei genvectors, the proposed approach has better performance at low SNR th an the conventional methods. We show that the improvement in performan ce increases with the number of snapshots. For example, the resolution improvement of the proposed signal eigenvector domain approach over s patial-smoothed Minimum-Norm is about 2.5 dB and 7 dB for 20 and 100 s napshots, respectively.