Prony's method modification of multi frequency signals spectral analysis.

The issues of digital spectral analysis of signals were considered. The possibility of accurate representation of a complex discrete sequence in the form of a linear combination of multiple harmonics without attenuation has been proved and implemented. A finite algorithm modifying the Prony's method was proposed, with a guaranteed arrangement of the poles of the corresponding autoregressive model on the circle of a unit radius. A scheme of economization of the order of the model in conditions of data redundancy in relation to the estimated limited number of harmonic components of the signal was developed.

Window Processing of EEG Records by the Prony’s Method

The paper considers the application of the Prony analysis for processing digital data in a sliding window. The stages of the algorithm are stated; the basic equations are given. On a model example, the features of constructing the Prony spectra are shown. Particular attention is paid to biometric data (EEG recordings). For them, the method of windowing with the extraction and visualization of complex roots is discussed. The possibility of using window processing to identify mental correlates in EEG records is shown and substantiated. Isolation of features in individual sections of the EEG can be of significant importance in the problems of biometric identification