The thesis launched new and advanced already known theoretical and methodological principles in the development of new efficient technologies for signal processing in computer systems, including primary components data convert, digital data exchange communications components implementation and identification and diagnosis of information and measurement systems. In the study, methods of implementation gas flow meters proposed to use additional measurement information sources, random fluctuations of variable amplitude information entropy, which are generated by the measuring environment, including, at a flow generator and broadband measurement signals. This processing method insensitivity to zero drift signal amplitudes of the primary device. Investigated the frequency spectrum measurement signals and set the frequency spectrum band information flow noise of the medium, which are characterized by maximum entropy dependent changes in estimates of the value of the current signal flow that allowed to realize the optimization procedure processing measurement data limited the relevant portions of the spectrum. In order to create information-measuring system flow of gaseous media for standard calibration facilities developed appropriate structural solutions, algorithms and software. The analysis of existing information transfer methods possible to propose alternative approaches to the formation and digital processing of wideband signals in the data paths. In particular character representation of the information entropy values posts random signals when creating a statistical estimation and entropy values during processing, creates a number of advantages: efficient use of the frequency band of the data channel, simplifying the hardware and software, provide adequate immunity at low signal / noise ratios. First obtained characteristics of the developed method, which found an increase in speed signal processing, increased noise immunity at the same time complexity of computer systems and communication, as well as uniformly bandwidth usage compared to conventional correlation methods. Developed and structural concepts of digital imaging devices and processing of broadband signals with manipulated information entropy based on universal chip microcomputer. The investigation and analysis of digital signal processing techniques in computer systems, traditionally, allocating the informative part of the signal is realized on the basis of statistical methods, spectral and correlation analysis. Consider effective correlation methods, but they do not work for signals that do not have acceptable correlation properties. Digital processing system based on the analysis of amplitude, frequency, phase, etc. signals characteristics require significant computational costs and allow us to estimate the characteristics of the overall signal. Proposed the use of new additional signal parameters, in particular, estimates of the information entropy. Developed projection methods for identifying objects on which to build projections realized by statistical estimation of the information entropy values pieces of binary matrices their representations, the processing method of diagnostic signals, based on the use of probabilistic representations of sequential fragments of their amplitudes. This approach yields a number of significant advantages, including: low dependence on the signal strength, insensitivity to zero drift amplitude of the primary device, simplifying the hardware and software. Modeling in computational experiments, experimental research and practical application of the developed digital tools and algorithmic software solutions confirmed the adequacy of the proposed approaches and the effectiveness of the developed methods. Keywords: information entropy, wideband signal, computer system, information-measuring system, sensor, forming, processing.
