Implementation of Noise in the Discrete Simulation of Continuous Systems

摘要

This report discusses the implementation of system and process noise, including sensor errors and atmospheric turbulence, in the simulation of continuous models by digital computers. The concept of discrete white noise is first introduced in a manner in which it reduces to continuous white noise as the integration time interval reduces to zero. The derivation of first and second order Markov processes from white noise is then discussed. The discussion includes the consideration of a suitable difference algorithm to approximate differentiation. The treatment throughout has been aimed at providing the reader with the tools to implement a required power spectral density (or autocorrelation function) as noise in any discrete digital simulation model. (Author)