Multisource delay estimation: nonstationary signals (underwater acoustics)

The authors study delay estimation in a very general framework: multiple sources, correlated noises, nonstationary random signals, and varying delays. The optimal maximum likelihood (ML) delay estimator is derived; its performance via the Cramer-Rao inequality is analyzed; and it is tested by experimentation with synthetic data. The present time-varying delay estimator extends to the nonstationary/multisource environment the estimators of L.C. Ng and Y. Bar-Shalom (1986), L.R. Kirlin and L.A. Dewey (1985), and C.H. Knapp and G.C. Carter (1976). However, the present receiver significantly departs from the correlator structures of those authors.<
>