Array processing for estimating multiple emitter parameters
Wei-wen Gu · 2005
In this paper the Multiple Signal Classification Method (MUSIC) is used to generate estimates of angles of arrival of multiple signals. The method involves determining the covariance matrix of the signals seen on the elements, then utilizing the eigenstructures in an algorithm to find estimates of the numbers of signal arrivals, angles of arrival, and power and cross-power of the arrivals. Various array layouts with uniform and non-uniform element spacing are examined Estimates are made using source pairs widely separated and closely spaced; equal and unequal amplitudes; uncorrelated and coorelated signals; and long and short signal sample records. Although correlated sources are viewed by the algorithm as a single signal, a small uncorrelated component in one of the arrivals makes it possible to resolve the two arrivals if enough samples are observed.