Yu.N. Klikushin - Dr.Sc. (Eng.), Professor, Department «Technology of Electronic Equipment», Omsk State Technical University. Е-mail: iit@omgtu.ru V.Yu. Kobenko - Ph.D. (Eng.), Associate Professor, Department «Technology of Electronic Equipment», Omsk State Technical University. Е-mail: kobra_vad@rambler.ru

Intellectual power of systems of data processing in many respects depends on their ability to carry out automatic classification of analyzed signals. Classification is understood as division of group of objects into some parts - subgroups in which objects have the general properties. The option of classification can be realized, using the methods accepted in applied statistics and the theory of recognition of images. However, because of mathematical and algorithmic complexity, these methods practically aren\'t applied in systems of real time. The offered method uses new methodology of classification and recognition of signals. This methodology is based on ideas and models of identification measurements of signals. Identification measurements are understood as the measurements which are quantitatively characterizing a form of signals or their characteristics. The offered method consists in the following: 1) distribution of a signal is represented in the form of the histogram with three intervals; 2) each fashion is considered as a code of monochrome color in system of RGB-vectors; 3) on a RGB code of distribution classification is carried out. The offered classification system allows to divide or to group analyzed signals, both in a form of distribution of instant values, and in properties of symmetry-asymmetry, camber-concavity. Measurability of a color tree on axes of coordinates (brightness-asymmetry) opens opportunities for automation of the solution of classification tasks by embedding of an analysis algorithm and a database at program level in microprocessor means of signals processing. Thus, the new prospect of creation of intellectual control facilities, measurements, control and diagnostics opens. From the informative point of view, the classification «color» model of distributions can form a basis of creation of algebra, as analytical device for the analysis and forecasting of results of interaction of signals.

 

1. Prikladnaja statistika: klassifikacija i snizhenie razmernosti: sprav. izd. / pod red. S.A. Ajjvazjana. M.: Finansy i statistika. 1989. 607 s.
2. Vasilev V.I. Raspoznajushhie sistemy. Kiev: Naukova Dumka. 1969. 292 s.
3. Zagorujjko N.G. Prikladnye metody analiza dannykh i znanijj. Novosibirsk: Izd-vo in-ta matematiki im. S.L. Soboleva SO RAN. 1999. 270 s.
4. Klikushin JU.N. Identifikacionnye instrumenty analiza i sinteza formy signalov: monografija. Omsk: Izd-vo OmGTU. 2010. 216 s.
5. Gumenjuk A.S., Klikushin JU.N., Kobenko V.JU. i dr. Algoritmy analiza struktury signalov i dannykh: monografija. Omsk: Izd-vo OmGTU. 2010. 272 s.
6. SHtejjnberg SH.E. Identifikacija v sistemakh upravlenija. M.: EHnergoatomizdat. 1987. 80 s.
7. Klikushin JU.N., Koshekov K.T., Koshekova B.V. Transljator kak universalnyjj instrument identifikacionnykh izmerenijj // Informacionno-izmeritelnye i upravljajushhie sistemy. 2008. T. 6. № 9. S. 24-30.
8. Klikushin JU.N. RGB - shkala dlja izmerenija raspredelenijj [EHlektronnyjj resurs] // ZHurnal Radioehlektroniki. 2008. № 3.Rezhim dostupa: http://jre.cplire.ru (data obrashhenija: 18.02.2013).
9. Klikushin JU.N., Kobenko V.JU., KolmogorovA.S. «Cvetovaja» model raspredelenijj [EHlektronnyjj resurs] // ZHurnal Radioehlektroniki. 2011. № 10.Rezhim dostupa: http://jre.cplire.ru (data obrashhenija: 18.02.2013).