Generation of Correlated Log-Normal Sequences for the Simulation of Clutter Echoes

摘要

The present investigation arose in connection with an attempt to simulate radar clutter echoes. Such echoes are random and usually correlated over several radar pulses. Among the detector envelope probability amplitude distributions which have been used in radar clutter modeling and analysis are the log-normal and Weibull distributions. The relationships between the means, variances and autocorrelation functions of normal and log-normal random variables are given. The generation of correlated normal sequences using the power spectral density is discussed. The relationship between the autocorrelation functions is then invoked to produce appropriately correlated log-normal sequences. Illustrations of the method are presented in the final section. The discussion is limited to sequences which represent wide-sense stationary processes. These are processes whose means are independent of time and whose correlation functions depend only on the time interval between samples.