University proceedings. volga region. technical sciences

Background. The following tasks are set in the work: analysis of the possibility of using emotional agents in SOS models of two groups; development of proposals for the use of emotional agents. The first group of models are deterministic and stochastic models (complete formalization of processes), the second are models with less formalization (Foresight method, situational analysis, etc.). Simulation of emotional states is relevant for systems with incomplete and inaccurate information, when the chosen decisions are often the result of emotions and can be irrational. Materials and methods. Methods that fully formalize emotional processes using modern mathematical apparatus and the possibility of simulating emotions with only partial formalization are considered. Results. It is proposed to use the imitation of the emergence of human emotional states by including certain types of agents in the model. Conclusions. The proposed use of agent types allows us to take into account emotions resulting from events that occurred long before the considered modeling interval in systems located outside the contour of the modeled system.

Background. Currently, multilayer convolutional networks of artificial deep learning neurons are actively used to recognize people's faces. Their testing is carried out according to the ISO/IEC 19795-1-2007 standard by testing laboratories in unfriendly countries, which may distort the test results. Materials and methods. The basic international standard stipulates the volume of the test base of real people’s faces. It is possible to significantly reduce the size of the test base through morphing synthesis of new biometric images by crossing the images of parents according to the domestic standard GOST R 2633.2- 2010. At the same time, incorrect crossing of parent images can lead to a distortion of the test results. The situation is complicated by the fact that the neural network face recognition tool will work with real data of people's faces of different quality. Results. It is proposed to eliminate the threat of possible distortion of test results by providing the testing laboratory by the customer with a number of statistical points describing the real working databases of people’s faces. It is shown that in addition to mathematical expectation and standard deviation, it is necessary to use the third and fourth statistical moments. When calculating statistical points, it is proposed to train the tested neural network to recognize specific biometric images of the faces of people who have given their consent to the use of their personal data.

Background. On-board computing systems (OCS) play a key role in the functioning of robotic complexes (RC), ensuring the transition from remote control to semi- andfully autonomous systems. This transition requires increased complexity and adaptability of hardware and software components. Due to the wide variety of solutions from the hardware, algorithmic, architectural and other aspects, as well as the limited computing resources of the UAV RTK, the choice of a rational composition of functional software during reconfiguration during operation is a non-trivial task. Materials and methods. A method for selecting a rational composition of the functional software of the UAV RTK in extreme conditions is presented, which allows selecting a rational configuration of the system depending on the requirements for solving the assigned tasks based on the current state of the external environment, the internal state of the UAV RTK and the existing limitations. Results and conclusions. The method works on multi-version libraries of functional programs of the UAV RTK. The method uses the solution of the generalized problem of a multiplicative multidimensional knapsack with multichoice and additional constraints. This approach allows taking into account complex dependencies between elements of functional software.

Background. Numerous studies have shown that one of the critical elements of low-bit integrating analog-to-digital converters (ADC) is a capacitor included in the integrator, the dielectric absorption of which limits the maximum achievable linearity of the conversion function. The relevance of studying its influence today is determined in connection with the development of multi-bit integrating analog-to-digital converters of the Σ- architecture, in which the summation of the conversion results of adjacent partial cycles is carried out for a full cycle. The execution of the operation of summation of partial and full conversion cycles leads to the emergence of a methodical component of the error from edge effects due to their asynchronism, which can lead to inequality of the integral values of the converted and reference voltages for a full conversion cycle. Materials and methods. These issues are investigated based on the developed ΣT-ADC using simulation modeling in the NI Multisim environment. An experimental study was conducted and the values of the root mean square deviation (RMS) of the linearity of the ADC conversion function were obtained using an ideal integrating capacitor, as well as with the addition of an integrating capacitor equivalent circuit with a number of absorption coefficient values to the model. The algorithm performance was assessed and the effect of the absorption coefficient on the final conversion result was analyzed for different numbers of partial cycles in a full cycle. Results. The study of the linearity error of the conversion function was conducted taking into account the effect of dielectric absorption achieved by including elements of the equivalent circuit of the integrating capacitor in the converter model. The values of the RMS of the conversion linearity were calculated for different numbers of partial cycles in a full conversion cycle. Conclusions. The obtained RMS values indicate the ability of the proposed structural and algorithmic solution to completely eliminate the influence of the instrumental component of the error caused by dielectric absorption phenomena in the integrating capacitor. For all values of the number of partial conversion cycles in a full cycle, absorption phenomena do not affect the final conversion result. For all values of the absorption coefficient, the mean square error of the linearity of the ΣT-ADC conversion function remains unchanged.

Background. One of the dominant factors affecting the macro- and microgeometric accuracy parameters of parts is the dynamic quality of automated metal-cutting machines. It is evaluated based on the results of measuring vibroacoustic vibrations of the dynamic system (DS) of machines, which are then processed by various methods. Dynamic quality can vary for machines of the same model depending on their technical condition, as well as for each machine depending on the assigned cutting mode, cutter wear and other factors. Studies performed at the Yuri Gagarin State Technical University showed that it is advisable to assess the dynamic quality of the machines by the DS stability margin determined as a result of special processing of vibration measurement information and identification of the DS oscillation autocorrelation function (ACF). The purpose of the work is to algorithmize the process of processing vibration measuring information when generating estimates of the dynamic quality of the machine by the stability margin of the dynamic system in order to identify the initial phase of catastrophic wear of the cutter. To achieve the goal, firstly, a theoretical justification of the type of ACF vibrations during cutting is carried out, which is then used to identify ACF from experimental data, secondly, a method and algorithm for processing vibration measuring information is proposed, and thirdly, a practical test of the method for recognizing the beginning of catastrophic wear of a cutter on a CNC lathe is carried out. Materials and methods. Two main areas of analysis and evaluation of the dynamic quality of machines can be distinguished, the first being based on theoretical models, in which the results of experiments are used to a certain extent, and the second - on experimental studies of vibrations of various machines nodes with further mathematical processing of measurement information to identify the DS model and form estimates of dynamic quality. The results of many years of research of turning and grinding machines have shown that in production conditions it is the second direction that allows you to quickly assess the dynamic quality of the machines, subject to computer processing of measuring information. Results and conclusions. The results of theoretical analysis ofstochastic oscillations in the dynamic system of the machine and experimental studies of the accuracy of machining bearing rings on a CNC lathe with simultaneous automated measurement of vibroacoustic oscillations and calculation of oscillation indicators of the dynamic system confirmed their correlation and the possibility of determining the criterion for detecting the initial phase of catastrophic wear of the cutter by a significant change (by more than 50%) of the first differences in oscillation indicators.

Background. The purpose of the research is to develop a computer model for optimizing the loading of machines of the same classification group, but of different types, of the existing fleet of technological equipment of the enterprise. Materials and methods. The research provided for the development of a calculation methodology and algorithm implemented in the MatchCAD mathematical package in the form of a computer model for comparative evaluation and optimization of calculating the number of parts processed by machines of different types based on the limitations available in production. Results. The developed technique made it possible to create a calculation algorithm based on it. The implemented computer model in the MatchCAD mathematical package allows optimizing the values of the number of parts processed in each group, but with machines of different types, by alternately optimizing to minimize operating costs, to minimize energy consumption and to minimize labor costs. Conclusions. The found difference in values between the option of the minimum production cycle for the manufacture of a batch of parts and the display is stable.