Signal Processing in the Linear Statistical Model

Abstract

This report summarizes our work on four general problems during the three-year period of the contract: (1) estimation of frequency-wavenumber, (2) matched subspace filters, (3) maximum likelihood estimation of modes from space-time data, and (4) statistical inference within the wavelet representation. We have generalized the theory of multiwindow estimators of the power spectrum to multiwindow estimators of the frequency wavenumber spectrum and of the related correlation sequence. We are now applying these ideas to the derivation of adaptive filters. We have developed a theory of matched subspace detectors for detecting signals which lie in low-dimensional model subspaces. The theory bridges the gap between generalized likelihood ratio theories and invariance theories. We have generalized the theory of maximum likelihood for identifying time domain modes and space domain directions of arrival. We have characterized subband decompositions for perfect reconstruction, developed filter design algorithms for constructing near perfect reconstruction filterbanks from nonorthogonal analysis filters, and derived algorithms for predicting and filtering in periodically correlated time series.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 04, 1994
Accession Number
ADA286255

Entities

People

  • Clifford T. Mullis
  • Louis L. Scharf

Organizations

  • University of Colorado Boulder

Tags

Communities of Interest

  • Energy and Power Technologies
  • Sensors

DTIC Thesaurus Topics

  • Adaptive Filters
  • Algorithms
  • Decomposition
  • Detectors
  • Digital Signal Processing
  • Engineering
  • Estimators
  • Filters
  • Filtration
  • Frequency
  • Maximum Likelihood Estimation
  • Modal Analysis
  • Power Spectra
  • Signal Processing
  • Spectra
  • Spectrum Analysis
  • Two Dimensional

Fields of Study

  • Engineering

Readers

  • Mathematical Modeling and Probability Theory.
  • Statistical inference.
  • Wave Propagation and Nonlinear Chaotic Dynamics.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • Space
  • Space - Space Objects