Improved DOA estimation via pseudorandom resampling of spatial spectrum
Alex B. Gershman, J.F. Böhme · IEEE Signal Processing Letters · 1997
A new pseudorandom spatial spectrum resampling scheme for improving the threshold performance of eigenstructure direction of arrival (DOA) estimators is proposed. Our approach is based on combining additional information arising when several underlying DOA estimators are available simultaneously, e.g., are calculated in parallel manner using given estimate of covariance matrix. Pseudorandomly generated weighted MUSIC estimators are exploited as appropriate underlying spatial spectral functions. Simulations show that our approach significantly improves the DOA estimation signal-to-noise ratio (SNR) threshold relative to the conventional and beamspace MUSIC estimators, and performs in the threshold region almost as well as the stochastic maximum likelihood (ML) estimator.