APPROXIMATION AND GENERATION OF GAUSSIAN AND NON-GAUSSIAN STATIONARY-PROCESSES

Combining a linear dynamical filter system and a nonlinear transform makes it possible to adapt the power density spectrum of an artificial stationary process as well as its probability distribution function to given quantities. A linear and a nonlinear approximate method for Gaussian processes are presented. An extension of the linear system by a static polynomial transform yields a non-Gaussian process. Varying the polynomial coefficients its distribution can be adapted. Using the analytical input-output relation of the power spectra, a suitable target spectrum for the linear filter can be evaluated. The entire adaptation concept is derived and principally discussed. © 1990.