Multiscale Surveillance Information Acquisition and Fusion.

Abstract

The goals of this research effort were to develop robust and adaptive array processing techniques and study adaptive radar waveform selection strategies for optimal range-Doppler imaging in a given time horizon. Specific achievements include the development of two novel wavelet based least mean squares algorithm (LMS) filtering and beamforming procedures. These procedures exploit the sparse structure of the wavelet transform of a wide class of autocorrelation matrices to derive a data dependent transformation of the input process. Their numerical complexities are comparable to those of the fast Fourier transform and discrete cosine transform based LMS techniques. However, their convergence rates are appreciably faster than those of time domain and other transform domain LMS procedures. Another achievement of this research effort is a new waveform selection strategy for optimal radar range-Doppler imaging. In particular, we have designed an optimal waveform selection strategy for the case where the radar targets to be imaged are known to belong to one of two classes. We have also identified an extension of this approach to the case where the target can belong to more than two classes. A complete study of the numerical and performance issue involved in that extension is currently under way. (MM)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 1995
Accession Number
ADA295600

Entities

People

  • A. H. Tewfik

Organizations

  • University of Minnesota

Tags

DTIC Thesaurus Topics

  • Acquisition
  • Algorithms
  • Autocorrelation
  • Convergence
  • Fast Fourier Transforms
  • Filtration
  • Mathematics
  • Radar Targets
  • Surveillance
  • Targets
  • Time Domain
  • Waveforms
  • Wavelet Transforms

Fields of Study

  • Engineering

Readers

  • Approximation Theory.
  • Computer Vision.
  • Phased Array Antenna Design.