Estimation of powers of uncorrelated sources in linear arrays

Abstract

The performance of the MUSIC estimator for the powers of uncorrelated sources arriving at a linear array is studied and is compared with the Cramer-Rao bound on the direction estimates of two closely spaced sources investigated in (Swingler, 1993). In general, it is found that if the source directions can be accurately estimated as when the sources are well separated, the estimator will function well and has a percentage variance which depends solely on the number of data samples. However, when the source separation is small and the direction estimates are poor, the normalized standard deviations of the power estimates will increase drastically and may become larger than that of the direction estimates. © 1995.