Accession Number : ADA204888

Title :   Robust Detectors for Very Heavy-Tailed Noise,

Corporate Author : PRINCETON UNIV NJ INFORMATION SCIENCES AND SYSTEMS LAB

Personal Author(s) : Warren, Douglas ; Thomas, John B

PDF Url : ADA204888

Report Date : Oct 1986

Pagination or Media Count : 11

Abstract : This paper treats the robust detection problem for the case where the noise mixture model is not strongly unimodal; that is, has heavier than exponential tails. Under a constraint on the support of the contaminating density, an optimal detector is derived. Its performance is investigated using both computer-generated noises and noise records from underwater noise fields. Keywords: Robust detectors, Signal processing, Non Gaussian noise. (JHD)

Descriptors :   *COMPUTER APPLICATIONS, *DETECTION, *NOISE, *UNDERWATER, CONTAMINATION, DENSITY, DETECTORS, GAUSSIAN NOISE, MIXTURES, MODELS, OPTIMIZATION, RECORDS, SIGNAL PROCESSING

Subject Categories : Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE