N.S. Konnova - Post-graduate Student, Bauman Moscow State Technical University
M.A. Basarab - Dr.Sc.(Phys.-Math.), Professor, Bauman Moscow State Technical University. E-mail: bmic@mail.ru
D.A. Basarab - Ph.D.(Med.), Cardiovascular department, St. Ioasaf's Belgorod Regional Hospital
D.D. Matsievsky - Ph.D.(Eng.), Scientific Research Institute on General Pathology and Pathophysiology RAMS, Moscow

The paper is devoted to the research of the data obtained by using the Doppler blood flow velocity sensor. It is substantiated the importance of this research direction, problems solved through the work are leaded. The relevance of the work theme and its originality, the development level of these studies in the world and in Russia in particular are discussed. We describe the data studied, the method of their preparation and registration, and provide a scheme of the experiment - from meter reading to digitization and signal processing on a computer, with a description of the equipment used and its operating principles. Various methods of digital signals under study processing are considered. In particular, it is the signal preprocessing methods, such as detection of noise to decide the type of noise, the smoothing, the elimination of the trend, and post-processing methods, such as the study of the general spectral pattern and the internal organization of the time series representing a signal, the construction of time-frequency characteristics of the signal, etc. We present arguments in favor of applying the methods of nonlinear dynamics (in particular, R/S-analysis, fractal analysis, calculation of the Lyapunov exponents), wavelet analysis, modeling, etc. The review description of the algorithms used is given, and the results of their application are briefly expounded. The article is also deals with the phase analysis of blood flow velocity curves, which is central to this article. Because of consideration of curves decomposition for phases, a detailed description of the cardiac cycle structure is leaded, and currently used mechanisms of blood flow data study and recording are described. Also a lot of attention is paid to the peculiarities of coronary blood flow, which distinguishes it from peripheral blood flow and, accordingly, distinguishes applied to it processing methods and calculation parameters. The following is a description of the phase blood flow curve analysis in the form, in which it is usually produced by an expert (manually), mainly used in the diagnosis of diseases and CVS states blood flow indicators and formulas for calculating them are pointed. There are discussed the reasons for difficulty with obtaining the synchronized data of linear blood flow velocity, blood pressure and ECG for coronary blood flow, with the purpose to use these data for splitting curves in phases of cardiac cycle (using specific "markers" on a given curve).For these reasons, a description of the proposed algorithm for the phase curve analysis based on cluster analysis is leaded. In this case, for specified bounds on the curve it builds automatically the clusters for the potential phase boundaries, then the algorithm selects the best points from each clusters and clarifies the position of points, if there are no signs of missing phase border point or, conversely, excessive certain point. It is possible to perform phase analysis of the curve, as in automatic mode with the optional adjustment of results by specialist, as well as in entirely manual mode. Examples of the algorithm working are provided. In conclusion the findings of the usefulness of these methods application in blood flow signal processing are given, in order to confirm and assist in decision making in the CVS states diagnosis, including intraoperative cases.