Software & Systems

Articles of journal № 4 at 2021 year.

Order result by: Public date | Title | Authors |

11. Diagnosing the condition of a technical object using machine learning classification [№4 за 2021 год]
Authors: Lomovtseva N.A., Yu.E. Kuvayskova , Klyachkin, V.N.
Visitors: 3389
Diagnosing the functioning of complex technical systems is necessary to ensure their safety and reliability. Sometimes the diagnosis is reduced to the division of objects into healthy and faulty: there is a binary classification of machine learning methods according to precedents (with the teacher). However, when there is a need to describe an object’s state with several possible options (not just two: a healthy object or a faulty object), a more detailed study is often needed. In this case, a multi-class classification of the object's states is carried out. Machine learning techniques can be used effectively as for binary classification. The sample obtained from the preliminary tests is divided into two parts: training and test. The training part is for building models that help to divided objects into a given number of classes. It is assumed that there is some connection between the object’s performance indicators and states. Based on the training sample, it is necessary to build an algorithm that provides a sufficiently accurate object’s state assessment for a given set of performance indicators. The paper presents a developed multi-class classification program allowing building an algorithm model for reliable diagnosis of the object’s condition. At the same time, cross-validation is used to eliminate retraining. The three quality measures of the built models are used to take into account the specifics of the training sample applying different types of classifiers. As a numerical example, the authors consider the robot's navigation: according to the results of 24 distance sensors, one of the four directions of its movement is determined.

12. Using discrete optimization tools to classify cognitive deficits: special aspects of using the minimax and additive criterion [№4 за 2021 год]
Authors: Razumnikova O.M., Mezentsev Yu.A., Pavlov P.S., Tarasova I.V., Trubnikova O.A.
Visitors: 3373
The paper devoted to the development of discrete optimization methods for solving the applied problem of clustering the cognitive resources of patients with coronary artery disease (CAD). The methods reflect the prospects of their surgical treatment. Many indicators of different cognitive functions and brain activity are used to determine the cognitive deficits associated with aging and concomitant cerebrovascular atherosclerosis. Coronary artery bypass grafting, which is widely used to treat CAD patients, increases the risk of postoperative cognitive deficits. In this regard, it is im-portant to identify the most informative markers of the cognitive status in patients in the preopera-tive state. To classify this state, the authors use the hemispheric activity characteristics, i.e. lateral-ized power of the theta, alpha, and beta rhythms together with the indicator of minimal cerebral dys-function (MMSE) and the integral cognitive indicator based on a set of parameters obtained during a recording sensory-motor responses and testing attention and memory in 114 male patients admitted to the clinic for coronary artery bypass grafting. The average patient’s age is 55.9 ± 5.3 years; 90 of them had secondary education and 32 had higher education. The results of computational experiments with clustering indicators of psychometric and neuro-physiological testing of CAD patients have shown the effectiveness of the developed toolkit for clustering by the discrete optimization means and the best discriminatory capabilities due to the ad-ditive criterion.

13. Semiotic control system for a mobile robotic platform [№4 за 2021 год]
Authors: M.A. Rovbo, Sorokoumov P.S.
Visitors: 3694
The paper considers the problem of combining a control system based on a semiotic model of a ro-bot’s world and human-machine interfaces, in particular, a voice interface, into a single system that processes both user commands and the robot’s autonomous behavior. The developed system allows controlling a robot using a voice interface, executing extended be-haviors, eliminating command ambiguities by resolving spatial relationships and taking into account the operator's gaze direction. This is achieved by integrating all the information necessary for mak-ing a decision and choosing the robot's action into the world model, applying logical inference to supplement it, and using heuristics to eliminate ambiguities in the speech command when there are not enough known facts for this. The decision-making system based on this world model also pro-vides the robot’s autonomous response to special situations. The paper describes the developed system architecture and demonstrates the applicability of this approach using a simulated model of a mobile robot in Gazebo. The simulation showed the possibil-ity of controlling a mobile platform using the developed system and the indirect control proved to be more ergonomic. Although, it should be noted that the computational load increases substantially when it is necessary to process a large number of objects in the world and it is impossible to process conflicting information in the current implementation. The computational load is partially compen-sated by algorithms that make it possible to infer only the information necessary to process the current command.

14. The environment status analysis and the issues of emission neutralization in China on the example of the Beijing area [№4 за 2021 год]
Authors: R.I. Solnitsev , Lei Wang, Kuzmin S.A., Kupriyanov G.A.
Visitors: 2959
This paper is devoted to the problem of compensation for emissions of pollutants into the atmos-phere. It shows the application of a comprehensive assessment of the emission concentration in space. An appropriate model of emission propagation is constructed depending on the type of a pol-lution source. The analysis is based on the data on emissions of pollutants in a specific area of the city of Beijing (China). Based on the analysis of the main stationary and mobile sources (vehicles) of pollution in this area, the authors have determined the gradients of emissions concentration in the atmosphere and the level of air quality during the day. Based on the analysis results, the paper proposes some approaches and design solutions to com-pensate for emissions based on the concept of constructing a closed system called Nature-Technology. Such closed-loop control systems include means to compensate for emissions from both stationary and non-stationary sources of pollutants. The solution of the control synthesis prob-lems according to the criterion of emission minimization in extreme sections of emission flows for both stationary and non-stationary (motor transport) sources of pollutants is proposed. The construction of closed-loop emission compensation control systems is associated with form-ing of an appropriate knowledge base. The knowledge base is based on the ontology components in-cluding the types of air pollutants and the ways to compensate for these substances. The ontology contains such concepts as the environment in which a person is currently located, a pollutant that is currently in the environment, and the current degree of compensation for sources of pollutants. In addition to the ontology, the knowledge base contains the rules that make it possible to conclude on the degree of compensation efficiency by the proposed means based on the environmental state in-cluding urban infrastructure or an enclosed space, and the type of pollutants (dust, toxic gases, or neutral gases). The ontology was developed in the Protégé editor. The web application for working with the knowledge base is based on JSP technology and the Java programming language. Based on the knowledge base, geographic information systems for data visualization and blockchain technol-ogy for registration and reliable storage of data are being built. The paper shows the prospects for the development of this approach for solving the complex problem of compensating emissions from both stationary sources of pollution and from vehicles.

15. A simulation model for estimating the service life of the Internet of Things under the conditions of attacking effects emitting the node energy [№4 за 2021 год]
Authors: Tatarnikova, T.M., P.Yu. Bogdanov
Visitors: 3737
The low power of the sensor nodes of the Internet of Things determines the search for solving sev-eral urgent problems: increasing the service life of sensor nodes and the security of the Internet of Things. Sensor nodes use batteries with limited resources as a power source, therefore if a sensor network is installed and deployed in a remote geographic space to observe physical phenomena, then recharging or replacing sensor nodes may become impossible or expensive due to the long dis-tance. Power consumption is one of the important quality indicators of the Internet of Things defined as the amount of energy used and spent by sensor nodes. Energy consumption determines the network lifespan – the time when the sensor network is fully functional. On the other hand, the implementa-tion of IoT security mechanisms requires additional energy costs associated with their implementa-tion. However, the lack of security mechanisms causes the proliferation of attacks that emit the node energy, as well as reduced service life of the Internet of Things. The paper presents the results of a simulation experiment proving that timely detection of at-tacks contributes to an increase in the service life of the network compared to a network with no se-curity mechanisms. To understand the operation principles of the simulation model, there is a a de-scription of its main modules, which simulate real objects of the Internet of Things network: sensor nodes, routers, protocols, communication channels, attacks, data packets. The estimates of energy consumption and service life are given in the form of graphs of dependences on various parameters of the Internet of Things network.

16. Thermodynamic constraints and information conditions of intelligent cognitive control stability, controllability, and robustness [№4 за 2021 год]
Authors: Ulyanov, S.V., Shevchenko A.A., Shevchenko A.V., Tyatyushkina O.Yu.
Visitors: 3089
The paper considers information and physical (entropy and energy) patterns, as well as the features of the model of a quantum strong artificial computational intelligence as a self-organizing intelli-gent control system. The model is based on the principles of minimal information entropy (in the “intelligent” space state of control signals) and the minimal generalized thermodynamic measure of the entropy production in the unified system “control object + intelligent cognitive controller”. The main result of applying the self-organization process is the guaranteed possibility of achieving the necessary reliability and flexibility level of the reproducible structure of the cognitive intelligent control system. The paper briefly describes the main physical principles of management processes allowing es-tablishing the relationship between the qualitative characteristics of the dynamic behavior of the control object and the executive device of the automatic control system: control stability, controlla-bility, and robustness. To achieve this purpose, it uses the information and thermodynamic ap-proaches that combine dynamic stability (Lyapunov function), controllability and robustness crite-ria by a homogeneous condition. The authors give the relations between the amount of pure work, information and the extracted free energy, which confirm the possibility of increasing the intellectual control system robustness due to the production of entropy of a cognitive controller that reduces the loss of the useful re-source of the control object. In turn, the negative entropy of cognitive control reduces the require-ments for the minimum initial information to achieve robustness. Based on the retrieved infor-mation from the cognitive controller knowledge base, it is possible to obtain an additional resource for useful work, which is equivalent to a targeted action on the management object, ensuring the management goal achievement.

◄ ← Preview | 1 | 2