Search

The article proposes a methodology for optimizing the process of irrigation of crops using a phytoindication system based on computer vision methods. We have proposed an algorithm and developed a system for obtaining a map of irrigation for maize in low latency mode. The system can be installed on a center pivot irrigation and consists of 8 IP cameras connected to a DVR connected to a laptop. The algorithm consists of three stages. Image preprocessing stage - applying an integrated excess green and excess red difference (ExGR) index. The classification stage is the application of the method that we choose depending on the system’s operating conditions. At the final stage, a neural network trained using the Resilient Propagation method is used, which determines the rate of watering of plants in the current sector of the location of the sprinkler. The selected methods of pretreatment and classification made it possible to achieve an accuracy of plant identification up to 93%, growth stages - up to 92% (with unconsolidated maize sowing and good lighting). System performance up to 100 plants in one second, which exceeds the performance of similar systems. The neural network showed an accuracy of 92% on the training set and 87% on the test set. Dynamic analysis of spatial and temporal variability leads to an increase in productivity and efficiency of water use. In addition, given the ubiquitous distribution of agribusiness management systems, this approach is quite simple to implement in the farm’s conditions.

Increasing the profitability of sales agents has been and continues to be given great attention. The article discusses various approaches to solving this problem, one of which is reinforcement learning, which is actively used to solve algorithmic trading problems. Increasing the efficiency (profitability) of algorithmic trading is possible in two ways, on the one hand, improving the algorithms, on the other hand, enriching the data that is transmitted to the input to the algorithms. The study confirmed the feasibility of using derived financial indicators for the tasks of applying trading algorithms based on reinforcement learning algorithms. The main idea of ​​the research implementation is aimed at obtaining the results of the sales agent's work based on Q-learning on technical indicators and on derived technical indicators (the agent is implemented in Python). The substantiation of the choice of the Q-learning method for solving the problem is carried out, the basics of decision-making, policy, strategy, and reinforcement learning are considered. The paper considers the issues of increasing the efficiency (profitability) of a sales agent based on the Q-learning algorithm by transferring derivative technical indicators to him, determined and substantiated derived technical indicators, verified the results of a sales agent's work on technical indicators and derived technical indicators. In the study presented in the article, an empirical test of the possibility of creating synthetic financial features to improve the efficiency of learning algorithms was carried out, in addition, the verification of obtaining the necessary results when using reinforcement learning algorithms was carried out. Empirical confirmation has been carried out that the use of derived financial indicators to increase the efficiency (profitability) of sales agents based on the Q-learning method with the use of reinforcement learning algorithms is expedient.

At the moment, there is a large volume of literature on exchange trading. Obviously, every year the mathematical base of work is becoming more complicated along with an increase in computing power, machines can process more metrics from year to year and produce more accurate solutions per unit of time. The use of deep learning has already proven itself well, as the application of this approach has given a quantum leap in algorithmic trading. The article presents an algorithm for trading long contracts with one asset in the financial market in the Python programming language using the LSTM neural network using the Keras library, which is used as a demo example in the Reinforcement Learning discipline. The formalized LSTM model solves the vanishing gradient problem, which can hold the gradient of the objective function relative to the state signal. As applied to our problem, such an improvement in the model allows us to collect data on certain patterns of price changes, that is, when predicting the price of the next step, we rely not only on the data of the previous step, but also on earlier data, when there was a similar state of the environment. Sharpe Ratio is used to determine the optimal strategy and make decisions at each time of application. The optimal minimum time period for the model operation has been determined; the signal transmission delay from the moment the market situation changes until the signal is received by the model, which will be infinitely small, and the computing power will be considered infinitely large. These assumptions give the right to say: when the market situation changes, the model is instantly ready to react and make a decision to sell, buy or hold an asset.

Incorporating brain-computer interfaces (BCIs) into daily life requires reducing the reliance of decoding algorithms on the calibration or enabling calibration with the minimal burden on the user. A potential solution could be a pre-trained decoder demonstrating a reasonable accuracy on the naive operators. Addressing this issue, we considered ambiguous stimuli classification tasks and trained an artificial neural network to classify brain responses to the stimuli of low and high ambiguity. We built a pre-trained classifier utilizing time-frequency features corresponding to the fundamental neurophysiological processes shared between subjects. To extract these features, we statistically contrasted electroencephalographic (EEG) spectral power between the classes in the representative group of subjects. As a result, the pre-trained classifier achieved 74% accuracy on the data of newly recruited subjects. Analysis of the literature suggested that a pre-trained classifier could help naive users to start using BCI bypassing training and further increased accuracy during the feedback session. Thus, our results contribute to using BCI during paralysis or limb amputation when there is no explicit user-generated kinematic output to properly train a decoder. In machine learning, our approach may facilitate the development of transfer learning (TL) methods for addressing the cross-subject problem. It allows extracting the interpretable feature subspace from the source data (the representative group of subjects) related to the target data (a naive user), preventing the negative transfer in the cross-subject tasks.

We trained an artificial neural network (ANN) to distinguish between correct and erroneous responses in the perceptual decision-making task using 32 EEG channels. The ANN input took the form of a 2D matrix where the vertical dimension reflected the number of EEG channels and the horizontal one—to the number of time samples. We focused on distinguishing the responses before their behavioural manifestation; therefore, we utilized EEG segments preceding the behavioural response. To deal with the 2D input data, ANN included a convolutional procedure transforming a 2D matrix into the 1D feature vector. We introduced three types of convolution, including 1D convolutions along the x- and y-axes and a 2D convolution along both axes. As a result, the F1-score for erroneous responses was above 88%, which confirmed the model’s ability to predict perceptual decision-making errors using EEG. Finally, we discussed the limitations of our approach and its potential use in the brain-computer interfaces to predict and prevent human errors in critical situations.

The article deals with the concept of building an automated system for the harvesting of apple crops. This system is a robotic complex mounted on a tractor cart, including an industrial robot and a packaging system with a container for fruit collection. The robot is equipped with a vacuum gripper and a vision system. A generator for power supply, a vacuum pump for the gripper and an equipment control system are also installed on the cart. The developed automated system will have a high degree of reliability that meets the requirements of operation in the field.

The development of mathematical models and efficient technologies for the processing of protein-containing dairy and vegetable raw materials and the production of food and feed concentrates with controlled functional properties is one of the most promising areas within the agricultural industry. In this work, the suitability of the electroflotation coagulation method for the combined extraction of vegetable and milk proteins was established by changing the electrolysis parameters and directed regulation of the isoelectric state of proteins. The research methodology is based on modern achievements of leading domestic and foreign researchers in the field of electrolysis of solutions and the creation of reagentless technologies for extracting proteins, as well as on the use of guest methods of physicochemical analysis, pH-metry, potentiometric and organoleptic analysis, methods of cyclic chronovoltammetry and currentless chronopotentiometry. The paper presents technological schemes for the extraction of vegetable and milk proteins, based on the combination of electroflotation and electrocoagulation processes. We carried out technological tests, which made it possible to determine the optimal conditions that ensure the highest yield of the product and its quality indicators. Ready-made isolates and concentrates of chickpea proteins and curd whey were obtained.

Currently, a significant group of industrial facilities can be classified as chemically hazardous facilities (CHFs). To predict the spread of harmful impurities in the programs being developed, Gaussian and Lagrangian models are actively used, on the basis of which the complexes used both in the EMERCOM of Russia and in research organizations are being implemented. These complexes require the introduction of a large amount of information, including the characteristics of the wind field in the distribution of an emergency chemically hazardous substance, which limits their use. In systems, the formation of which is influenced by a large number of different random factors, spatial scaling (similarity) is often found, and one or another parameter can be described using the methods of fractal geometry, which in the past few decades has been actively and successfully applied to the description of various physical objects. The purpose of this study is to analyze the possibility of using the random-addition method for early prediction of the distribution of harmful impurities in the surface air layer during the short-term release of a substance on the surface as a result of an emergency.

Mathematical computer models of the permittivity of silicon-based nanostructures upon interaction with electromagnetic radiation in a wide frequency range have been developed. To implement computer models for studying the electrophysical properties of the structures under study, algorithms and a set of programs have been developed. The results of the study of materials will not only provide fundamental information about the physical effects occurring in composite nanostructures but will also be useful for solving problems related to calculations for given electrophysical problems. For a nanocomposite based on ceramics and semiconductor oxides of zinc grains, resonant bursts of permittivity are observed within a wavelength of 300–400 nm; it has been found that this is due to the presence of electronic polarization of the nanocomposite core. The paper presents the results of modeling the current-voltage characteristics of a nanocomposite based on ceramics and semiconductor grains of zinc oxide. The obtained results show that the geometrical parameters, such as the number of layers and sample width, affect the CVC of the nanocomposite, and the operating point of the CVC shifts. This may be of interest in the development of materials with desired electrical characteristics for the creation of varistors.

To predict the spread of the new coronavirus infection COVID-19, the critical values of spread indicators have been determined for deciding on the introduction of restrictive measures using the city of Moscow as an example. A model was developed using classical methods of mathematical modeling based on exponential regression, the accuracy of the forecast was estimated, and the shortcomings of mathematical methods for predicting the spread of infection for more than two weeks. As a solution to the problem of the accuracy of long-term forecasts for more than two weeks, two models based on machine learning methods are proposed: a recurrent neural network with two layers of long short-term memory (LSTM) blocks and a 1-D convolutional neural network with a description of the choice of an optimization algorithm. The forecast accuracy of ML models was evaluated in comparison with the exponential regression model and one another using the example of data on the number of COVID-19 cases in the city of Moscow.

The authors carried out modeling of the electrophysical properties of composite media. The frequency dependences of the dielectric constant on the type of inclusions in the composite are investigated. On the basis of the nanocomposite considered in the work, based on Si, B, and SiO2, the authors model a reflecting screen, the lattice elements of which have a layered hierarchically constructed structure similar to a fractal formation. The influence of the level of fractality on the optical properties of the object was also investigated, and it was found that the proposed structure makes it possible to increase the operating frequency range of the reflecting screen and the efficiency, in comparison with reflecting screens that have a lattice of traditional structure. The results obtained can be of practical interest for broadband and nonlinear radar devices, localization devices and mobile objects, microelectronics, as well as intelligent applications in the field of information security.

In recent years, various methods and directions for solving a system of Boolean algebraic equations have been invented, and now they are being very actively investigated. One of these directions is the method of transforming a system of Boolean algebraic equations, given over a ring of Boolean polynomials, into systems of equations over a field of real numbers, and various optimization methods can be applied to these systems. In this paper, we propose a new transformation method for Solving Systems of Boolean Algebraic Equations (SBAE). The essence of the proposed method is that firstly, SBAE written with logical operations are transformed (approximated) in a system of harmonic-polynomial equations in the unit n-dimensional cube Kn with the usual operations of addition and multiplication of numbers. Secondly, a transformed (approximated) system in Kn is solved by using the optimization method. We substantiated the correctness and the right to exist of the proposed method with reliable evidence. Based on this work, plans for further research to improve the proposed method are outlined.

A universal solution to an applied problem related to the study of deviations occurring in the joints of manipulation robots, for example, due to elastic deformations or gaps in them, is proposed. A mathematical (dynamic) model obtained by the Lagrange–Euler method is presented, making it possible to investigate such deviations. Six generalized coordinates, three linear and three angulars, were used to describe the variations of each joint in the dynamic model. This made it possible to introduce into consideration joints with six degrees of freedom (6-DOF joints). In addition, mathematical models for external forces acting on the links of manipulation robots are presented. When composing matrices of coefficients of equations of motion, elements identically equal to zero were excluded, which significantly increased the computational efficiency of these equations. The dynamic model based on the obtained equations can be used in the computer simulation of manipulation robots.

An algorithm for parallel calculations in a dynamic model of manipulation robots obtained by the Lagrange–Euler method is developed. Independent components were identified in the structure of the dynamic model by its decomposition. Using the technology of object-oriented programming, classes corresponding to the structures of the selected components of the dynamic model were described. The algorithmization of parallel computing is based on the independence of the calculation of objects of individual classes and the sequence of matrix operations. The estimation of the execution time of parallel algorithms, the resulting acceleration, and the efficiency of using processors is given.

Composite materials consisting of a dielectric matrix with conductive inclusions are promising in the field of micro- and optoelectronics. The properties of a nanocomposite material are strongly influenced by the characteristics of the substances included in its composition, as well as the shape and size of inclusions and the orientation of particles in the matrix. The use of nanocomposite material has significantly expanded and covers various systems. The anisotropic form of inclusions is the main reason for the appearance of optical anisotropy. In this article, models and methods describing the electrical conductivity of a layered nanocomposite of a self-similar structure are proposed. The method of modeling the electrical conductivity of individual blocks, layers, and composite as a whole is carried out similarly to the method of determining the dielectric constant. The advantage of the method proposed in this paper is the removal of restrictions imposed on the theory of generalized conductivity associated with the need to set the dielectric constant.

Recently, we have seen the rapidly growing popularity of unmanned aerial vehicles. This is due to some advantages, namely portability, the ability to fly over hard-to-reach areas without human intervention. They are also widely used for commercial purposes, agriculture, delivery, automation in warehouses. The potential of unmanned aerial vehicles is vast and demonstrates promising opportunities. However, when using these devices, the issue of safety is acute. This article presents a developed software application that is used to improve the efficiency of flight research of groups of unmanned aerial vehicles, based on a new method for assessing flight safety by comparing the complexity of specified air routes. A practical approach to modeling and evaluating the search for a safe way is proposed. A suitable method of research is computer and simulation modeling. It is suggested to use the spectrum of dynamic characteristics of the sequence as a formal attribute for analyzing routes. The method is illustrated by an example of comparing air trajectories according to the flight safety criterion. The software application is intended for use in the educational process when training specialists in transport security, robotics, and system analysis.

The thermomechanical stability of the anion–cation exchange matrix “Polikon AC” on viscose nonwoven materials is investigated. In this work, a molecular model of a solvation environment for experimentally obtained “Polikon AC” mosaic membranes is refined. Mosaic membranes on a viscose fiber base were fabricated by the method of polycondensation filling. The temperature dependence of deformation was investigated for dry and wet anion and cation exchange membrane components at a constant tensile load of 1.5 N and a heating rate of 8 °C/min. The effect of moisture content on the deformation of anionite and cationite fragments under a constant external tensile load of 1.5 and 3 N in a temperature range up to 100 °C was studied.

A model of a layered hierarchically constructed composite is presented, the structure of which demonstrates the properties of similarity at different scales. For the proposed model of the composite, fractal analysis was carried out, including an assessment of the permissible range of scales, calculation of fractal capacity, Hausdorff and Minkovsky dimensions, calculation of the Hurst exponent. The maximum and minimum sizes at which fractal properties are observed are investigated, and a quantitative assessment of the complexity of the proposed model is carried out. A software package is developed that allows calculating the fractal characteristics of hierarchically constructed composite media. A qualitative analysis of the calculated fractal characteristics is carried out.

Analysis of the rise in prices for consumer goods is a state’s priority task. The state assumes the obligation to regulate pricing in all spheres of consumption. First of all, the prices for essential commodities to which agricultural products belong are analyzed. The article shows the changes in prices for consumer goods of agricultural products (sugar) during a pandemic. The analysis of forecasting prices for sugar and its impact on the development of its production is carried out. The construction of the forecast model was based on extrapolation. The structure of a forecast model for price changes was based on the analysis of the time series of the Autoregressive Integrated Moving Average (ARIMA) class. This model consists of an autoregressive model and a moving average model. A forecast of the volume of domestic sugar transportation by rail has been completed. The algorithms implemented this model for searching for initial approximations and optimal parameters for the predictive model. The Hirotsugu Akaike Information Criterion (AIC) was used to select the best model. The algorithms were implemented in the Python programming language. The quality check of the description was performed with a predictive model of actual data. An economic interpretation of the rise in sugar prices and proof of the forecast’s truth obtained from a financial point of view were carried out.