DOI: 10.7256/2454-0714.2017.4.21909

Abstract: To date, the problem of developing systems, algorithms and methods for spatial orientation and navigation of robots remains one of the most urgent problems. The subject of the study are multi-angle and bipolar systems of technical vision, as well as an algorithm for the formation of 3D models of the surface of objects. The author examines in detail such aspects in the development of mobile robots, as ways of obtaining information about the three-dimensional scene, estimation of the depth of the scene, integration, the method of constructing epipolar lines, some elements of virtual reality are touched upon. Applied approaches in the work are based on the methods of digital image processing, image recognition, discrete transformations and system analysis, a method for constructing epipolar lines is given. The work also considers the formation of 3D models of the surface of the object in the systems of technical vision, applied to mobile robots, an algorithm for determining the distance to the points on the surface of the object in the systems of technical vision on the mobile robot. The results of the research can be applied in systems for modernizing the management, monitoring and processing of information, providing more accurate technical vision in the development of mobile and autonomous robots, which in turn can affect the import substitution and the overall economic component of the use and development of robotic systems and systems.

DOI: 10.7256/2454-0714.2017.4.22887

Abstract: The research is devoted to computer technologies for processing personal data. In the article particular attention is paid to the preparatory stage of data processing, which precedes the meaningful analysis of data. This stage of work largely determines the quality of the results of the whole work. The preparatory stage is very time-consuming and requires a lot of time. Automation of the work of participants in the process of collecting information makes it possible to increase the efficiency of the entire system for the analysis of personal data. The purpose of this study is to develop a software tool that allows the data collected in various ways to be combined into a single database. To analyze various schemes for organizing information collection and transferring information to the computer medium, the methodology of system analysis was used. In addition, the practical experience of using various Internet services for organizing online surveys was used. Computer technology has the property of universality, allowing the use of a wide range of schemes and methods of organizing questionnaires. The advantage of the technology is its ease of use and availability to a wide range of users who use the EXCEL environment in their work. The effectiveness of the developed software was confirmed in the course of practical work on the study of the socio-economic problems of the region.

DOI: 10.7256/2454-0714.2017.4.24437

Abstract: The subject of the study is the concept of hierarchical bitmap-indexes proposed by the author. It is that in order to improve the processing performance of queries on the time filter, the indices are supported not only for the values of the basic unit of time, but also for arbitrary larger multiple units. The object of the study is the construction of an analytic probability model of such indices for the particular case of the exponential distribution of a random stream of recording records in a database. The author focuses on such an aspect as the calculation of the discrete distribution of the number of indices involved in the processing of the query. The methodology of the study is probability theory, combinatorial methods, measure theory, computational experiment. In addition, it is shown that the latest concepts of the theory of cellular automata, such as the Zaitsev's neighborhood, can be used to study the features of the proposed model. The main results of the work can be formulated as follows: introduced an original, intuitive concept of building indexes; new, meaningful optimization problems for selecting a hierarchical index system are formulated; a mathematical model is constructed and verified, allowing to estimate the efficiency of using the chosen hierarchy of indices. It is shown that in the limiting case the model naturally tends to a set of fractal nature, in particular, one of the varieties of Cantor dust, for which the formula for calculating its Hausdorff-Besicovitch dimension is derived through the application parameters of the initial problem.

DOI: 10.7256/2454-0714.2017.4.24785

Abstract: The most common type of sounding signals of sonar systems are broadband linear-frequency-modulated signals (LFM). Due to the use of such signals it is possible to increase the range and resolving power of hydroacoustic devices, but this does not solve the problem of mutual interference of pulses. One type of broadband signals are chaotic oscillations, which theoretically are able to become an alternative to chirp signals due to increased resistance to crosstalk. The paper compares chaotic and chirp signals by simulating their passage in an aqueous medium. Chaotic signals generated by systems with different numbers of basins of attraction are investigated. The study shows that chaotic signals are more resistant to mutual interference, and also experience less attenuation when propagating in an aqueous medium compared to the types of signals currently used. The stability of chaotic broadband signals to interference and attenuation correlates with the number of basins of attraction of the attractor of the generating chaotic system. Based on the results of the work, it can be concluded that the chaotic signals are used as probing pulses of sonar systems.

DOI: 10.7256/2454-0714.2017.4.24534

Abstract: The problem of the normalization of frequency-time correlation functions is considered and solved. The aim of the work is to create a methodology for calculating the coefficients for normalizing the time-frequency correlation functions and integrating them into the known computational algorithm. At the same time, the tasks were to ensure the possibility of normalizing each frequency component of the time-frequency correlation function independently and maintaining high performance of the original algorithm. The latter imposed restrictions on the application of the filtering operation in the time domain and the use of additional discrete Fourier transforms in the algorithm. To minimize the computational costs in calculating and rationing the time-frequency correlation functions, a technique was developed for calculating the normalizing coefficients from the samples of the complex signal spectrum. The main result of the work is a new algorithm for calculating the normalized time-frequency correlation function, which differs by an insignificant increase in computational complexity in comparison with the original algorithm. At the same time, the coefficients obtained can be used both for simultaneous normalization of all the frequency components of the time-frequency correlation function, which is necessary to ensure the independence of the result from the scale of the input signals, and for independent normalization of each of its frequency components. The latter is useful in solving problems of detecting weak correlated components in signal mixtures.

DOI: 10.7256/2454-0714.2017.4.24655

Abstract: The article explores the applicability of long short-term memory (LSTM) recurrent networks for the binary classification of text messages of the social network Twitter. A three-stage classification process has been designed, allowing a separate analysis of pictograms and verification of the text for neutrality. The accuracy of the classification of the emotional polarity of text messages using the LSTM network and vector representations of words was verified. The percentage of coincidences of vector representations of words with a training set of data is determined, which makes it possible to obtain an acceptable classification accuracy. The estimation of the learning speed of the LSTM network and the use of memory was carried out. To solve the task of classifying text messages, methods of processing natural language and machine learning using precedents are applied. The algorithmic base for processing text data from social networks, obtained as a result of the application of LSTM neural networks, has been optimized. The novelty of the proposed solution method is due to the implementation of pre-processing of messages, which allows to improve the accuracy of classification, and the use of the neural network configuration taking into account the specifics of text data of social networks.

DOI: 10.7256/2454-0714.2017.4.24786

Abstract: An approach to the selection of a finite-difference scheme of a chaotic pseudo-random sequence generator based on the use of step diagrams (h-diagrams) is proposed. As a test problem, a generator is considered based on the random Rössler system discretized by explicit, implicit and semiquant numerical methods of the first and second order of algebraic accuracy. The sequences generated by different variants of the generator are randomly checked by a battery of NIST statistical tests. Advantages of the proposed approach in the design of chaotic signal generators are shown, consisting in an essential (by an order of magnitude) acceleration of the device design time due to a new method of selecting the discretization step and the discrete operator. The effectiveness of using semi-implicit finite difference schemes in the generation of pseudo-random sequences by the method of numerical solution of chaotic differential equations is confirmed. The obtained results can be used in cryptography applications, in the design of secure communication systems, in solving problems of numerical simulation of dynamical systems and mathematical statistics.

DOI: 10.7256/2454-0714.2017.4.24251

Abstract: The research focuses on enterprises and organizations of various industries, leading project-oriented business. The subject of the study are the decision-making processes that are present in the implementation of various projects. Increasing the effectiveness of decisions can be achieved through the use of expert systems. At the same time, the expert system should be based on modern methods of processing information in conditions of uncertainty. The author suggests using expert systems based on methods and models of fuzzy logic. Particular attention in the article the author pays to the functional requirements, which must correspond to fuzzy expert system. The author examines in detail the existing list of software for the implementation of fuzzy expert system, and to identify the optimal software which meets the requirements, the method of analysis of complex systems by the criterion of functional completeness of Professor G.N. Khubaeva.As a result of the analysis it was established that the existing software solutions do not meet the functional requirements in many respects, therefore the development of a new and effective tool is an actual task.The analysis also made it possible to identify software tools with a similar set of functions, to estimate the degree of similarity and the degree of correspondence of the systems of the "reference" model of the information system that takes into account the requirements of the user.