Correlation Function Estimator Performance in Non-Gaussian Spherically Invariant Random Processes

Abstract

In this report, analytic expressions are developed for the variance, error variance and bias of the time-averaged correlation function estimator for stationary, discrete, non-Gaussian complex processes. The expressions derived here pertain to the general class of non-Gaussian processes known as Spherically Invariant Random Processes (SIRP's) Specific results are shown for K-distributed processes which form a special case of the SIRP's. Furthermore, these equations are derived for the general case of processes with unconstrained quadrature components; i.e., for processes exhibiting elliptical symmetry For the special case of complex processes with constrained correlation between the quadrature components (i.e., circular symmetry), the resulting analytic expressions attain a simplified form. Validity of the analytic expressions is presented using Monte-Carlo simulations. Correlation function estimator, Estimation, Spherically invariant random processes, Ergodicity, Non-Gaussian random processes.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 01, 1993
Accession Number
ADA273498

Entities

People

  • James H. Michels

Organizations

  • Rome Laboratory

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Air Force
  • Command And Control
  • Computational Science
  • Covariance
  • Data Science
  • Estimators
  • Gaussian Processes
  • Information Processing
  • Information Science
  • Monte Carlo Method
  • Random Variables
  • Shape
  • Simulations
  • Stationary
  • Statistical Algorithms
  • Statistics
  • Symmetry

Readers

  • Approximation Theory.
  • Statistical inference.