Analytic conditions for asymptotic optimality of the maximum-likelihood estimator with application to problems in remote sensing
Nicholas C. Makris · The Journal of the Acoustical Society of America · 2003
Analytic conditions necessary for a general multivariate maximum-likelihood estimate (MLE) to become asymptotically optimal are derived from the first principles of estimation theory following Makris and Naftali [J. Acoust. Soc. Am. 110, 1917–1930 (2001)]. The conditions are given as sample size or signal-to-noise ratio requirements for the MLE to become asymptotically unbiased and attain the minimum variance possible for any unbiased estimate known as the Cramer–Rao lower bound (CRLB). Applications to problems common in remote sensing such as time-delay and Doppler shift estimation, beamforming, source localization in a waveguide, pattern recognition in 2-D intensity images, surface orientation estimation, and Kalman filtering will be presented.