Rapidly convergent #x201C;superresolving#x201D; direction finders of noise radiation sources in adaptive arrays

摘要

For a number of known "superresolving" algorithms intended for direction-finding of point sources of noise radiation, we compare the requirements to the size of training sample in the estimates of input signals' covariance matrix. We discuss the reasons of the distinctions between the analyzed algorithms, and propose new kinds of methods with improved robustness to the finite sample size.