An Expectation-Maximization Approach to Spectral Structure Identification and Estimation

Abstract

The method of Expectation-Maximization is applied to find an approximate maximum-likelihood estimate of the amplitudes, phases and frequency shift of a known collection of spectral lines in a signal with Gaussian noise. The technique works well for many frequency components provided that these frequencies are well separated. It is applied to the identification of the frequencies present in a radar return from a propeller driven aircraft, arising from modulation by blades of the propellers.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 01, 2001
Accession Number
ADA401234

Entities

People

  • Alan Bolton
  • Marian Viola
  • William Moran

Organizations

  • Defence Science and Technology Group

Tags

Communities of Interest

  • Advanced Electronics
  • Air Platforms
  • Materials and Manufacturing Processes
  • Sensors

DTIC Thesaurus Topics

  • Algorithms
  • Australia
  • Electrical Engineering
  • Electronics
  • Engineering
  • Frequency
  • Frequency Shift
  • Information Exchange
  • Information Processing
  • Mechanical Engineering
  • Propellers
  • Radar
  • Random Variables
  • Signal Processing
  • Standards
  • Surveillance
  • Universities

Fields of Study

  • Physics

Readers

  • Aerodynamics.
  • Radar Systems Engineering.
  • Radio communications and signal processing.