Исследование параметров окна частотно-временного преобразования при проведении декомпозиции сигнального радиопрофиля

Abstract

Known methods of time-frequency transformation, such as the Gabor-Ville transform, wavelet transform, are complex and cumbersome in calculations, require the correct selection of basic expansion functions, may have interference points and negative values of the spectral density function. The presence of space-time uncertainty inherent in any frequency-time decomposition, as well as the lack of the possibility of correcting the found expansion coefficients, make modern methods of frequency-time transformation ineffective in the analysis of signal radio profiles (SRP). Decomposition of SRP with windowed correlation feedback eliminates the above disadvantages by multiple enumeration of parameters, however, this occurs at the expense of a significant increase in the transformation time. The transformation window here is understood as a rectangular function that moves along the time axis from the origin. In this work, the window width is understood as the number of sampling points falling into the rectangular window function, which is equivalent to the length of the function along the time axis. The window step is the number of sampling points to which the rectangular window function (transformation window) moves in the time-frequency decomposition. The purpose of this work is to increase the efficiency of the frequency-time decomposition of the SRP by analyzing the parameters of the transformation window in decomposition with windowed correlation feedback. In the study, experimental research methods are used to register the electrical component of the electromagnetic field radiated by the product - SRP. Correlation analysis methods for determining the best match between the reference and reconstructed SRP. Methods of digital signal processing with frequency-time decomposition. The paper presents the registered SRPs and their parameters, a correlation analysis of the SRP data is carried out. The dependences of the correlation functions on the width and step of the window are given for a different number of sampling points of the registered SRP. The results obtained for the first time showed the possibility of using certain parameters of the transformation window to improve the accuracy and significantly reduce the time of SRP decomposition, which makes the work novel. It has been established that with a window width equal to half the length of the SRP itself and a window step of not more than a tenth of the length of the SRP, it is possible to achieve an increase in the cross-correlation between the original and restored SRP up to 0.95. Recommendations are given on the speed of recording equipment when receiving PDS (sampling rate below 10 GS/s). The practical significance of the work lies in the reduction of iterations of the multi-parameter approximation when using correlation feedback in the course of SRP decomposition, which significantly reduces the conversion time and improves its accuracy.