The method of segmentation of stochastic cyclic signals for the problems of their processing and modeling

Abstract

The segmentation method (signal partitioning into specific sections, segments) of a cyclic stochastic signal is considered. Taking into account cyclicity attributes, segmental cyclic and segmental zone structure, there is proposed a method for solving the segmentation problem of a cyclic stochastic signal with a stable or replaceable rhythm. The information obtained on the segment structure allows us to analyze the rhythm: to estimate the value of the period in the case of a stable rhythm, and in the case of a variable rhythm - the rhythmic structure (discrete rhythm function). There are given the examples of segmentation of modeled and real cyclic signals, and the accuracy of the developed segmentation method is estimated. The developed method can be used in digital process automation systems (diagnostics and forecasting) of cyclic data: cardiac signals of various physical nature, cyclic economic processes, gas consumption and energy consumption processes, surface processes of relief formations of modern materials.