A Monte Carlo Performance Analysis of Accelerated SVD (Singular Value Decomposition)-Based High Discrimination Algorithms

Abstract

High discrimination algorithms are increasingly being considered for the task of processing data from arrays of sensors and in the form of time series. Many such algorithms rely on singular value decomposition of a data matrix or eight analysis of the corresponding covariance estimate, thereby imposing a heavy computational requirement. However, if the data is oversampled, or if the solution vector may be constrained, in terms of its angular extent or frequency range for example, it is often possible to pre-process the data matrix in such a way as to reduce its size. This may be carried out by means of a fixed matrix pre-multiplication, and can lead to a substantial acceleration of the subsequent analysis. The method is described, and its use exemplified in combination with a number of well-known high discrimination algorithms. A number of results from a Monte Carlo analysis are given which show that the new technique can lead to significantly improved parameter estimates being obtained from the high discrimination algorithms. Great Britain.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1988
Accession Number
ADA200043

Entities

People

  • J. L. Mather

Organizations

  • Royal Signals and Radar Establishment

Tags

Communities of Interest

  • Energy and Power Technologies
  • Sensors

DTIC Thesaurus Topics

  • Algorithms
  • Cross Correlation
  • Data Analysis
  • Data Science
  • Detection
  • Detectors
  • Eigenvalues
  • False Alarms
  • Far Field
  • Frequency
  • Information Science
  • Linear Arrays
  • Matched Filters
  • Monte Carlo Method
  • Noise
  • Statistics
  • Time Domain

Readers

  • Linear Algebra
  • Statistical inference.
  • Systems Analysis and Design