Martyniuk T. Theoretical basis and organization of high-effective computers with parallel difference cuts data processing

The object of research is the process of parallel processing of numeric one-and two-dimensional data sets by difference cuts (DC) for neurocalculation and image classification; the purpose of the research is to improve the efficiency of computers of intelligent systems through the development of theoretical propositions of difference cut processing of data sets and creation on this basis the new mathematical, algorithmic and structural models oriented hardware implementation of parallel computing structures with enhanced functionality; methods of research are based on the general provisions of the theory of discrete data transformations, theory of algorithms, methods of the theory of digital processing and analysis of signals and images, pattern recognition theory, methods of theory of analysis and synthesis for computing systems, method of mapping iterative algorithms on regular structures, methods of mathematical and simulation modeling; theoretical results - theory of eye-computing data processing by the introduction and justification of the method of DC guidelines is generated, including first: theoretical foundations of linear vector data transformation based on the formation of difference cuts (SM- conversion) are developed, the method of element processing of discriminant function (DF) as a matrix for the image classification by DC is developed, a formal description of mathematical models of numbers parallel processing by DC with the using SAA V.M. Glushkov is proposed; practical results - the functionality of computational tool in the previous processings and analysis and in pattern recognition is enhanced, which allows to extend the application of the proposed computational tools, the hardware implementation on FPGA the pipelined and matrix processors for compact devices in intelligent systems is developed, the programs to simulation modeling for parallel processing of one-and two-dimensional data sets by DC with definition of temporal parameters are developed that allows to define the maximum time spent in the implementation of the developed methods by DC on a specific hardware. The degree of implementation - the results of the thesis introduced in enterprises: Federal State Unitary Enterprise "NPO Astrophysics " (Russia, Moscow ), JSC "Space Informatic Analytic Systems" (Russia, Moscow), Center for Optoelectronic Technology Research Institute of Microdevices of NAS of Ukraine (Kyiv), "Infrakon" (Vinnitsa), Vinnitsa National Medical University named M.I. Pirogov. Sector (industry) use - hardware intelligent recognition systems, robotic systems, simulation tools in biomedical engineering.