Selective Properties of Structural Transformations in Multisinusoidal Signals Analysis

摘要

Determining periodic components in a multi-sinusoidal signal is a complex process, especially if the signal is noise-contaminated or contains measurement interference. This paper asserts selective properties of structural transformations used for structural and parametric identification of multisinusoidal signals. The concept of structural transformation is based on the structural function of the signal. Practical studies have shown that consistent application of the third structural transformation provides interference filtering and step-by-step selection of harmonics of a multisinusoidal signal, starting with maximum amplitude harmonic. Selective properties of structural transformations also allow us to determine the number of signal harmonic components and all their parameters without additional a priori assumptions concerning properties of the signal. Structural transformations efficiency is exemplified by computer simulation of noise-contaminated multisinusoidal signal with two harmonics, digitally processed with a certain selected discretization step. The use of structural transformations with gradual allocation of periodic components in the studied signal permits to determine the number and parameters of harmonics under а priori uncertainty.