Спектральний аналіз сигналів із застосуванням його представлення складовим векторним випадковим процесом

Abstract

The work investigates the SVRP model of the process vector, which consists of subvectors of shorter length, and its properties for the analysis and modeling of non-stationary data with trend and seasonal components. The change in the frequency of a single-mode signal with an increase in the length of the subvector was studied and formulas for its estimation were proposed. The use of the SVRP model and decimation as methods of long-term analysis are compared. It is shown that the use of a long-term model of the AR signal in the SVRP presentation allows to obtain an increase in resolution.