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.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1993
- Accession Number
- ADA273498
Entities
People
- James H. Michels
Organizations
- Rome Laboratory