A useful approximation for nonGaussian distributions
Abstract
It is shown that the firstorder probability density function (pdf) of a nonGaussian function may be approximated by a weighted sum of zeromean Gaussians or by a weighted sum of displaced Gaussians. The method is used to approximate a firstorder Rayleigh distribution and the distribution of a sum of a Gaussian noise and a filtered shot noise. The output distribution of a filter forced by an impulsive noise process at its input may be approximated by a sum of conditional Gaussian distributions for some specific combinations of the amplitude distribution of the impulses and filter characteristics, facilitating the evaluation of system components.
 Publication:

International Journal of Electronics
 Pub Date:
 January 1976
 Bibcode:
 1976IJE....40...81C
 Keywords:

 Approximation;
 Communication Theory;
 Noise Spectra;
 Normal Density Functions;
 Rayleigh Distribution;
 Statistical Analysis;
 Control Theory;
 Random Noise;
 Shot Noise;
 Signal Processing;
 Communications and Radar