Detection and Classification of Fluctuating Targets

Abstract

The structure of echoes observed in echo-location systems such as radar and sonar can exhibit significant variability caused by environmental effects and fluctuations in the target aspect. A technique for countering these effects and thereby improving the detection and classification of the target is investigated. The technique relies on being able to characterizes the target signal space using a low dimensional signal basis. For this purpose the Karhunen-Loeve expansion is used to construct the target signal space and in conjunction with the likelihood ratio test form the appropriate test statistic, which is the basis of the detector/classifier structure for distinguishing target returns from nontarget ones. This structure is sometimes referred to as the separable kernel receiver. With this structure in place the performance of the detector/classifier can be evaluated as a function of clutter type and target mismatch. Performance comparisons are also made with a peak detector and a matched filter processor detecting Rayleigh and Rician targets in Rayleigh clutter. The results are presented in the usual form of receiver operating characteristic (ROC) curves.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 19, 2001
Accession Number
ADA389350

Entities

People

  • David M. Drumheller
  • Henry Lew

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Energy and Power Technologies
  • Sensors

DTIC Thesaurus Topics

  • Aspect Angle
  • Classification
  • Data Science
  • Detection
  • Detectors
  • Eigenvalues
  • Eigenvectors
  • False Alarms
  • Filters
  • Frequency
  • Gaussian Processes
  • Machine Learning
  • Matched Filters
  • Random Variables
  • Scattering
  • Standards
  • Warning Systems

Readers

  • Radar Systems Engineering.
  • Statistical inference.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • Space
  • Space - Space Objects