SYSTEM IDENTIFICATION IN THE PRESENCE OF NOISE

摘要

A method of identification of a dynamic process, which is governed by a certain form of exponential function of time and contaminated with noise, is presented. The approach that is used consists of numerical filtering of experimental data, numerical derivative of filtered numerical data, numerical filtering of the derivative and natural logarithm of derivative of filtered experimental data, and linear curve fit to the logarithm of derivative of assumed exponential function. All the steps are carried out in the EXCEL spreadsheet (EXCEL is a trade mark of Microsoft Corporation).