Generation of non-Gaussian stationary stochastic processes

Abstract

A procedure is developed to generate a non-Gaussian stationary stochastic process with the knowledge of its first-order probability density and the spectral density. The procedure is applicable to an arbitrary probability density if the spectral density is of a low-pass type, and to a large class of probability densities if the spectral density is of a narrow band, with its peak located at a nonzero frequency. © 1996 The American Physical Society.