Methods of Computer-Aided Analysis of Non-Gaussian Noise and Application to Robust Adaptive Detection. Part 2.

We present a methodology for the modeling of certain non-stationary and non-gaussian random time series data with application to weak signal detection. Some components of the noise, which give it its nongaussian characteristics, can be individually modeled, synthesized and subtracted to provide a ga...

Full description

Bibliographic Details
Main Authors: Kirsteins,I, Tufts,D W
Other Authors: RHODE ISLAND UNIV KINGSTON DEPT OF ELECTRICAL ENGINEERING
Format: Text
Language:English
Published: 1984
Subjects:
Statistics and Probability
Cybernetics
*TIME SERIES ANALYSIS
*SIGNALS
*COMPUTER AIDED DIAGNOSIS
ALGORITHMS
METHODOLOGY
DETECTION
MULTIVARIATE ANALYSIS
MATRICES(MATHEMATICS)
ESTIMATES
ADAPTIVE SYSTEMS
QUANTIZATION
COVARIANCE
LOW STRENGTH
ACOUSTIC DATA
MATRIX DISPLAYS
Nongaussian noise
Robust procedures
Arctic
Online Access:http://www.dtic.mil/docs/citations/ADA148879
http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA148879
Description
Summary:We present a methodology for the modeling of certain non-stationary and non-gaussian random time series data with application to weak signal detection. Some components of the noise, which give it its nongaussian characteristics, can be individually modeled, synthesized and subtracted to provide a gaussian residual. Further, it is shown that this process can also be carried out when signals are present. The proposed methodology is applied to some Arctic Acoustic data using a combination of adaptive differential quantization and adaptive signal estimation algorithms based on singular-value-decomposition of a data matrix which we have developed. The combination of adaptive differential quantization with low-rank approximations to data matrices or estimated covariance matrices is believed to be a new and effective method for multivariable, robust, adaptive detection.

Similar Items