Showing 1–1 of 1 results for author: Berman, M M
-
Simultaneous Detection and Estimation, False Alarm Prediction for a Continuous Family of Signals in Gaussian Noise
Authors:
D. Michael Milder,
Robert G. Lindgren,
Morris M. Berman
Abstract:
New problems arise when the standard theory of joint detection and estimation is applied to a set of signals drawn from a continuous family; decision thresholds must be determined as a function of the continuous parameter x characterizing the signals, and false alarms occur, not with a discrete probability, but with a density in x. A Bayes decision structure over the domain of signal parameters yi…
▽ More
New problems arise when the standard theory of joint detection and estimation is applied to a set of signals drawn from a continuous family; decision thresholds must be determined as a function of the continuous parameter x characterizing the signals, and false alarms occur, not with a discrete probability, but with a density in x. A Bayes decision structure over the domain of signal parameters yields a state estimate of the signal parameter x as an integral part of a signal declaration. The decision criterion is converted to a form in which detection and false alarm densities appear and from which is derived a relation between them for all x. The limiting case of additive Gaussian noise and a high detection threshold allows a simplified decision criterion and a state estimate of signal location in x that approaches the Cramer-Rao bound. Also in this limit, an analytic form for the false alarm probability density over x, a quantity not readily obtained in general, is evaluated here through its relation to the detection probability. The false alarm density expression and state accuracy prediction are tested through Monte Carlo simulations, and the comparison demonstrates excellent agreement.
△ Less
Submitted 15 July, 2014;
originally announced July 2014.
