Vol 20 No 2 (2021): Informatics and Automation

In modern conditions, in the field of the creation and use of existing and advanced space vehicles (SV), the issues of autonomy and survivability acquire particular relevance in the development and operation of small-mass spacecraft (SMS) for Earth remote sensing (ERS). The specificity of the small spacecraft lies in the fact that it is difficult to directly apply to the process of their creation the standard practice of using the system for ensuring the reliability of the rocket and space industry due to the lack of the ability to provide full structural redundancy of its onboard systems (OBS) associated with mass-dimensional and other restrictions. In this case, the tasks of developing model-algorithmic methods and approaches to ensuring the required level of indicators of structural reliability, survivability and, in general, the effectiveness of the functioning of the MCA OBS become of particular relevance. The problem of increasing the level of indicators of autonomy, survivability, efficiency of functioning of complex technical objects (CTO), which, in particular, SMS belong, is considered in the scientific literature in conjunction with solving problems of control, assessment and technical diagnostics of the state of the CTO reconfiguration (structural, functional, structural-functional reconfiguration) of CTO structures, management of its reserves, alternative and multi-mode control, analysis of fault tolerance and disaster recovery of CTO. However, all of these studies are fragmented, both at the methodological and methodological and technological levels. The article provides a generalized description of the combined methods and algorithms developed by the authors for solving the problems of synthesis of technologies and programs for controlling the OS reconfiguration to increase the survivability of the SMS. At the same time, these tasks are solved not in isolation, but in a comprehensive manner within the framework of the general problem of proactive management of the structural dynamics of SMS with or without the use of GCC tools, which ensures the efficiency, validity, completeness, isolation and consistency of synthesized management decisions. The novelty of the approach proposed in the article is that its authors, based on the concepts of integrated (system) modeling, proactive control of the structural dynamics of the OS SMS, as well as the intellectualization of the processes of proactive control of the OS SMS, developed methods and algorithms for the synthesis of technologies and programs. Control of the reconfiguration of the MCS BS, providing, firstly, the situational choice of the optimal sequence of operations and the allocation of SMS resources with and without the use of GCC facilities, and, secondly, effective parrying not only of the calculated ones, but also off-design emergency flight situations (EFS), as well as the operational restoration of the operability of its OS. The constructiveness of the proposed approach is illustrated by the example of solving the problem of flexible redistribution of information processing tasks between the OS SMS and the SMS GCC.

The purpose of the study is to select the optimal conditions for collecting non-coordinate information about a spacecraft by a space optical-electronic means at the time objects pass the vicinity of the points of the minimum distance between their orbits. The quantitative indicator is proposed that characterize the measure of the possibility of obtaining non-coordinate information about space objects with the required level of quality. The arguments of the function characterizing the indicator are the distance between spacecraft; their relative speed; phase angle of illumination of a spacecraft by the Sun in relation to the optical-electronic means; the length of the time interval during which both objects are in the vicinity of the point of a minimum distance between their orbits. The value of the indicator is computed by solving three particular research problems. The first task is to search for neighborhoods that include the minimum distances between the orbits of the controlled spacecraft and optical-electronic means. To solve it, a fast algorithm for calculating the minimum distance between orbits used. Additionally, the drift of the found neighborhoods is taken into account on the time interval up to 60 hours. The second task is to estimate the characteristics of motion and the conditions of optical visibility of a controlled spacecraft in the vicinity of the minimum points of the distance between the orbits of spacecraft. The solution to this problem is carried out by using the SGP4 library of space objects motion forecast. The third task is justification and calculation of an index characterizing the measure of the possibility of obtaining an optical image of a spacecraft for given conditions of optical visibility. To solve the problem, the developed system of fuzzy inference rules and the Mamdani algorithm is used. The presented method is implemented as a program. In the course of a computational experiment, an assessment was made of the possibility of obtaining non-coordinate information on low-orbit and geostationary space objects. The proposed indicator provides an increase in the efficiency of the procedure for collecting non-coordinate information about space objects by choosing the most informative alternatives for monitoring space objects from the available set of possible observations at a given planning interval for collecting information about space objects.

The paper proposes a method for fusioning multi-angle images implementing the algorithm for quasi-optimal clustering of pixels to the original images of the land surface. The original multi-angle images formed by the onboard equipment of multi-positional location systems are docked into a single composite image and, using a high-speed algorithm for quasi-optimal pixel clustering, are reduced to several colors while maintaining characteristic boundaries. A feature of the algorithm of quasi-optimal pixel clustering is the generation of a series of partitions with gradually increasing detail due to a variable number of clusters. This feature allows you to choose an appropriate partition of a pair of docked images from the generated series. The search for reference points of the isolated contours is performed on a pair of images from the selected partition of the docked image. A functional transformation is determined for these points. And after it has been applied to the original images, the degree of correlation of the fused image is estimated. Both the position of the reference points of the contour and the desired functional transformation itself are refined until the evaluation of the fusion quality is acceptable. The type of functional transformation is selected according to the images reduced in color, which later is applied to the original images. This process is repeated for clustered images with greater detail in the event that the assessment of the fusion quality is not acceptable. The purpose of present study is to develop a method that allows synthesizing fused image of the land surface from heteromorphic and heterogeneous images. The paper presents the following features of the fusing method. The first feature is the processing of a single composite image from a pair of docked source images by the pixel clustering algorithm, what makes it possible to isolate the same areas in its different parts in a similar way. The second feature consists in determining the functional transformation by the isolated reference points of the contour on the processed pair of clustered images, which is later applied to the original images to combine them. The paper presents the results on the synthesis of a fused image both from homogeneous (optical) images and from heterogeneous (radar and optical) images. A distinctive feature of the developed method is to improve the quality of synthesis, increase the accuracy and information content of the final fused image of the land surface.

Spectral analysis of signals is used as one of the main methods for studying systems and objects of various physical natures. Under conditions of a priori statistical uncertainty, the signals are subject to random changes and noise. Spectral analysis of such signals involves the estimation of the power spectral density (PSD). One of the classical methods for estimating PSD is the periodogram method. The algorithms that implement this method in digital form are based on the discrete Fourier transform. Digital multiplication operations are mass operations in these algorithms. The use of window functions leads to an increase in the number of these operations. Multiplication operations are among the most time consuming operations. They are the dominant factor in determining the computational capabilities of an algorithm and determine its multiplicative complexity. The paper deals with the problem of reducing the multiplicative complexity of calculating the periodogram estimate of the PSD using window functions. The problem is solved based on the use of binary-sign stochastic quantization for converting a signal into digital form. This two-level signal quantization is carried out without systematic error. Based on the theory of discrete-event modeling, the result of a binary-sign stochastic quantization in time is considered as a chronological sequence of significant events determined by the change in its values. The use of a discrete-event model for the result of binary-sign stochastic quantization provided an analytical calculation of integration operations during the transition from the analog form of the periodogram estimation of the SPM to the mathematical procedures for calculating it in discrete form. These procedures became the basis for the development of a digital algorithm. The main computational operations of the algorithm are addition and subtraction arithmetic operations. Reducing the number of multiplication operations decreases the overall computational complexity of the PSD estimation. Numerical experiments were carried out to study the algorithm operation. They were carried out on the basis of simulation modeling of the discrete-event procedure of binary-sign stochastic quantization. The results of calculating the PSD estimates are presented using a number of the most famous window functions as an example. The results obtained indicate that the use of the developed algorithm allows calculating periodogram estimates of PSD with high accuracy and frequency resolution in the presence of additive white noise at a low signal-to-noise ratio. The practical implementation of the algorithm is carried out in the form of a functionally independent software module. This module can be used as a part of complex metrologically significant software for operational analysis of the frequency composition of complex signals.

The problem of stability analysis and its components of reliability and survivability is quite popular both in the field of telecommunications and in other industries involved in the development and operation of complex networks. The most suitable network model for this type of problem is a model that uses the postulates of graph theory. At the same time, the assumption of the random nature of failures of individual links of the telecommunications network allows it to be considered in the form of a generalized Erdos–Renyi model. It is well known that the probability of failure of elements can be interpreted in the form of a readiness coefficient and an operational readiness coefficient, as well as in the form of other indicators that characterize the performance of elements of a telecommunications network. Most approaches consider only the case of bipolar connectivity, when it is necessary to ensure the interaction of two end destinations. In modern telecommunications networks, services such as virtual private networks come to the fore, for which multipoint connections are organized that do not fit into the concept of bipolar connectivity. In this regard, we propose to extend this approach to the analysis of multi-pole and all-pole connections. The approach for two-pole connectivity is based on a method that uses the connectivity matrix as a basis, and, in fact, assumes a sequential search of all combinations of vertex sections, starting from the source and drain. This method leads to the inclusion of non-minimal cross-sections in the general composition, which required the introduction of an additional procedure for checking the added cross-section for non-excess. The approach for all-pole connectivity is based on a method that uses the connectivity matrix as a basis, and, in fact, assumes a sequential search of all combinations of vertex sections, not including one of the vertices considered terminal. A simpler solution was to control the added section for uniqueness. The approach for multipolar connectivity is similar to that used in the formation of the set of minimal all-pole sections and differs only in the procedure for selecting the combinations used to form the cross-section matrix, of which only those containing pole vertices are preserved. As a test communication network, the Rostelecom backbone network is used, deployed to form flows in the direction of "Europe-Asia". It is shown that multipolar sections are the most general concept with respect to two-pole and all-pole sections. despite the possibility of such a generalization, in practical applications it is advisable to consider particular cases due to their lower computational complexity.

In this article the task of determining the current position of pneumatic actuators is considered. The solution to the given task is achieved by using a technical vision system that allows to apply the fuzzy clustering method to determine in real time the center coordinates and the displacement position of a color label located on the mechatronic complex actuators. The objective of this work is to improve the accuracy of the moving actuator’s of mechatronic complex by improving the accuracy of the color label recognition. The intellectualization of process of the color shade recognition is based on fuzzy clustering. First, a fuzzy model is built, that allows depending on the input parameters of the color intensity for each of the RGB channels and the color tone component, to select a certain color in the image. After that, the color image is binarized and noise is suppressed. The authors used two defuzzification models during simulation a fuzzy system: one is based on the center of gravity method (CoG) and the other is based on the method of area ratio (MAR). The model is implemented based on the method of area ratio and allows to remove the dead zones that are present in the center of gravity model. The method of area ratio determines the location of the color label in the image frame. Subsequently, when the actuator is moved longitudinally, the vision system determines the location of the color label in the new frame. The color label position offset between the source and target images allows to determine the moved distance of the color label. In order to study how noise affects recognition accuracy, the following digital filters were used: median, Gaussian, matrix and binomial. Analysis of the accuracy of these filters showed that the best result was obtained when using a Gaussian filter. The estimation was based on the signal-to-noise coefficient. The mathematical models of fuzzy clustering of color label recognition were simulated in the Matlab/Simulink environment. Experimental studies of technical vision system performance with the proposed fuzzy clustering model were carried out on a pneumatic mechatronic complex that performs processing, moving and storing of details. During the experiments, a color label was placed on the cylinder, after which the cylinder moved along the guides in the longitudinal direction. During the movement, video recording and image recognition were performed. To determine the accuracy of color label recognition, the PSNR and RMSE coefficients were calculated which were equal 38.21 and 3.14, respectively. The accuracy of determining the displacement based on the developed model for recognizing color labels was equal 99.7%. The defuzzifier speed has increased to 590 ns.

A new fast method for pupil detection and eyetracking real time is being developed based on the study of a boundary-step model of a grayscale image by the Laplacian-Gaussian operator and finding a new proposed descriptor of accumulated differences (point identifier), which displays a measure of the equidistance of each point from the boundaries of some relative monotonous area (for example, the pupil of the eye). The operation of this descriptor is based on the assumption that the pupil in the frame is the most rounded monotonic region with a high brightness difference at the border, the pixels of the region should have an intensity less than a predetermined threshold (but the pupil may not be the darkest region in the image). Taking into account all of the above characteristics of the pupil, the descriptor allows achieving high detection accuracy of its center and size, in contrast to methods based on threshold image segmentation, based on the assumption of the pupil as the darkest area, morphological methods (recursive morphological erosion), correlation or methods that investigate only the boundary image model (Hough transform and its variations with two-dimensional and three-dimensional parameter spaces, the Starburst algorithm, Swirski, RANSAC, ElSe). The possibility of representing the pupil tracking problem as a multidimensional unconstrained optimization problem and its solution by the Hook-Jeeves non-gradient method, where the function expressing the descriptor is used as the objective function, is investigated. In this case, there is no need to calculate the descriptor for each point of the image (compiling a special accumulator function), which significantly speeds up the work of the method. The proposed descriptor and method were analyzed, and a software package was developed in Python 3 (visualization) and C ++ (tracking kernel) in the laboratory of the Physics and Mathematics Faculty of Kamchatka State University of Vitus Bering, which allows illustrating the work of the method and tracking the pupil in real time.

The paper proposes a hybrid method for training convolutional neural networks. The method consists of combining second and first-order methods for different elements of the architecture of a convolutional neural network. The hybrid convolution neural network training method allows to achieve significantly better convergence compared to Adam; however, it requires fewer computational operations to implement. Using the proposed method, it is possible to train networks on which learning paralysis occurs when using first-order methods. Moreover, the proposed method could adjust its computational complexity to the hardware on which the computation is performed; at the same time, the hybrid method allows using the mini-packet learning approach. The analysis of the ratio of computations between convolutional neural networks and fully connected artificial neural networks is presented. The mathematical apparatus of error optimization of artificial neural networks is considered, including the method of backpropagation of the error, the Levenberg-Marquardt algorithm. The main limitations of these methods that arise when training a convolutional neural network are analyzed. The analysis of the stability of the proposed method when the initialization parameters are changed. The results of the applicability of the method in various problems are presented.