A Method for the Generation of Correlated Random Processes

Abstract

In this paper the authors propose a method which will allow the generation of random discrete variables with a given probability distribution and autocorrelation sequence. The method is applicable in cases where, once experimental measurements have established the statistical characteristics of a stationary random process, an algorithm is to be implemented to generate random discrete variables with the same statistical properties in terms of amplitude distribution and correlation among the values. The method proposed allows a correlated random process to be generated by combining two ergodic independent statistical uncorrelated random processes. A case study is given to apply the method proposed to modelling of variable bit rate video sources by simulation.