Software & Systems

Articles of journal № 4 at 2013 year.

Order result by: Public date | Title | Authors |

41. Intelligent control of a robot manipulator based on soft computing [№4 за 2013 год]
Authors: Nikolaeva A.V., Ulyanov, S.V.
Visitors: 12820
The problem of designing control systems using soft computing is described with 3 degrees of freedom manipulator as an example. The article investigates a general methodology of robust knowledge base design using special intelligent toolkit (Knowledgebase Optimizator on soft computing technology). Variations of organization coordinated control are proposed: single knowledge base containing information about manipulator’s three links, and a divided control method (independent knowledge bases for particular links control). The efficiency of constructed intelligent control systems using soft computing technology is described in comparison with control system using constant settings of regulator defined by genetic algorithm. Set of performance criteria taking into consideration for automatic control theory methods of transition estimations and tailored to concerned control object (3 degrees of freedom robot manipulator) is used to estimate control systems work. Estimate of control system work is performed by results of MatLab/Simulink simulation and experiment series on practical control object.

42. Kalman fuzzy filter in the structure of the algorithm solving inverse problems for economic objects [№4 за 2013 год]
Author: Павлов Д.А.
Visitors: 12678
A method solving inverse problems using Kalman fuzzy filter considering specific economic facilities, such as the inability to build a physical model of the process or phenomenon has been proposed. The algorithm is based on the transition from the system description in the form of a Fredholm integral equation of the type I to the form of the discrete Kalman filter, as well as the using fuzzy logic to set the elements of filter equations. The method for solving inverse problems is based on object output process data, which is usually the result of any parameters measurement. The measurement noise results in a further decrease of the solutions accuracy. Therefore it is appropriate to use filtration which reduces this impact. As for economic object, the measurement noise can be interpreted as statistical data precision, inaccurate analysis and expert models used in the interpretation of its operating results and others. To reduce experts subjectivity contributed to the knowledge base, it is proposed to apply fuzzy cluster analysismethods and adaptive neuro-fuzzy systems methods allowing automatical synthesizing the knowledge base structure. MatLAB has been used as a software development environment, because it has a library to create adaptive neuro-fuzzy systems (ANFIS-editor).

43. Intelligent decision support system to manage complex objects using dynamic fuzzy cognitive maps [№4 за 2013 год]
Authors: Palyukh B.V., Dli M.I., Какатунова Т.В., Багузова О.В.
Visitors: 11188
The article discusses the necessity of using cognitive modeling algorithms to determine the consistency of strategies for social, economic and innovative development of the region. It also shows the way of implementing cognitive modeling techniques based on the use of fuzzy dynamic multi-level cognitive model of regional innovation processes management as a computer program. The authors identify the main functional blocks in the process of strategic innovation management in the region. It allows choosing the key features of the proposed intelligent decision support system for complex objects management. Based on the results of the structural and functional modeling of business processes for strategic management of innovative activity in the regions, an adaptive intelligent information-analytical system has been developed. This system provides decision- making support for innovative development in the region. The authors propose an architecture of adaptive intelligent information and analytical decision support system for managing innovations in the region. Proposed system consists of four blocks: the formation of a regional strategy for innovation management, modeling, performance evaluation of investment decisions to support of innovation; innovation controlling.

44. Software implementation of slabs automatic identification system [№4 за 2013 год]
Authors: Провоторов А.В., Орлов А.А.
Visitors: 10364
The structure of hardware and software complex (SALI) for slabs moving tracking between warehouse and main manufacture equipment is described in the article. It considers automatic metalrolled manufacture complex production monitoring. Distinctive feature of the system is algorithm of automatic slabs identification (bulky metal units) which allows identifying the moving object with high accuracy. It can make the workers of crane free from their routine of slab number or code identifying. One of the logic units of algorithm is automatic step-by-step product label identifying. The result of label detection steps is calibration of camera moving mechanism position and zoom for getting primary hi-res image of load with label, its analysis, label scanning and recognition. There are the results confirming efficiency and reliability of the method.

45. Robust intelligent control based on quantum computing technology [№4 за 2013 год]
Authors: A.G. Reshetnikov, Керимов Т.А., Ulyanov, S.V.
Visitors: 12402
The article describes the technology of robust intelligent control system design based on quantum fuzzy inference in two versions: using remote object connection and using verified mathematical model. Developed design methodology is used based on knowledge self-organization. It improves the robustness of intelligent control systems in unforeseen situations. The information sources for fuzzy controllers design are devided into two types: the first includes inaccurate mathematical model (with sub-definite parameters) of control object, and the second is based on the process of measuring the teaching signal from the real control object. A comparison of simulation and experimental results for physical test of PID controller, fuzzy controllers and quantum fuzzy controllers with different types of correlations is considered. The efficiency of using the quantum fuzzy inference to improve the robustness of intelligent control in emergencies management situations is confirmed.

46. Soft computing technologies in intelligent control [№4 за 2013 год]
Authors: A.G. Reshetnikov, Керимов Т.А., Ulyanov, S.V., Barkhatova I.A.
Visitors: 14010
The article consideres the methods and capabilities of knowledge base design based on models of intelligent control systems in the autonomous mode and remote-controlled mode (USB cable and Wi-Fi). As an example authors take a physical dynamically unstable “cart – pole” type object (typical Benchmark). The paper focuses on developing intelligent control methodology with /without ill-structured mathematical model of control object. The teaching signal designed off-line using a mathematical model, and on-line directly from a physical object using “knowledge base optimizer” tool. The comparison of control systems based on PID controller and fuzzy controllers in three modes is introduced. The possibility of effective remote configuration of a fuzzy controller without a mathematical model of the control object is discussed.

47. Performance investigation of self-configuring genetic algorithm for spacecrafts' control systems effective variant choice [№4 за 2013 год]
Authors: Semenkin E.S., Семёнкина М.Е., Попов Е.А.
Visitors: 11151
The work of the spacecraft control system is modeled with Markov chains. The models for the technological and command-programming control contours are developed. The way of the calculation of the control contour effectiveness indicators (readiness coefficients and others) is described. The problem of choosing an effective variant for a spacecraft's control system is formulated as a multi-scale optimization problem with algorithmically given functions. Special self-configuring genetic algorithm is suggested to use for choosing effective variants of spacecraft control systems as it is very reliable, has no requirements to settings determination and parameter tuning and requires no expert knowledge in evolutionary optimization from end users (aerospace engineers). The high performance of the suggested algorithm is demonstrated by solving optimization problems. It was shown that the algorithms require the examination of small part of the search space to find solution that greatly accelerates the process of search for a solution of the problem of choosing effective variants of spacecraft control systems.

48. Software system for ionospheric measurements data processing [№4 за 2013 год]
Author: Skripachev V.O.
Visitors: 20551
The authors propose a software system to determine the ionosphere parameters by radio probing means. This system helps to get information about environment parameters in order to further diagnose the disturbances, both natural and artificial character, and to specify the known ionospheric models. The developed software system is designed to restore the profiles of the electron density in the ionosphere using signals from navigation satellites GLONASS/GPS. Object-oriented programming was used to develop the system. This reduces the time develop new classes, inheriting them from already existing in used software libraries. The object-oriented approach has been implemented when developing the code responsible for the total electron content calculation according to signals characteristics of navigation spacecraft. Therefore interaction between system blocks is also performed based on object data structures. Prospects of future work suppose building the capacity of the represented software system using the software products for ionospheric sounding data processing, and that data is obtained by vertical sounding of the ionosphere and incoherent scattering techniques.

49. Interactive Data Language application for radiometric systems data processing [№4 за 2013 год]
Authors: Skripachev V.O., Polushkovsky Yu.A., Стрельцов А.М.
Visitors: 20983
When researching the "atmosphere – the underlying surface" system, a large amount of data is used that must be processed quickly. It is advisable to use applications written in the Interactive Data Language for detailed on-line radiometric data processing. A syntax is quite simple in the Interactive Data Language. It is not necessary in cycles preparation and run-time functions optimization for working with arrays. The paper discusses the features of the language, descriptions of its classes and objects. There is a brief overview of the scientific data visualization by Interactive Data Language tools including mapping capabilities. The possibilities of Interactive Data Language to build a graphical user interface are considered. These features have been implemented for creating data processing applications of satellite’s microwave radiometers to determine the meteorological parameters. The HDF data format and developed application architecture are given.

50. “Parallel Matlab” cloud service [№4 за 2013 год]
Authors: Созыкин А.В., Goldshtein M.L, Gribova, V.V., F.M. Moskalenko
Visitors: 15615
The approach to provide cloud service for parallel computation in Matlab is proposed. The distinctive feature of the approach is providing not only a parallel Matlab cluster, but also a virtual personal workspace with Matlab GUI as a cloud service. The users of the cloud service have the ability to execute jobs on the Matlab cluster from the Matlab GUI. In addition, Matlab GUI can be used to analyze results of the computation. An architecture of the “Parallel Matlab” cloud service is described. The suggested architecture is implemented at the supercomputer center of the Institute of Mathematics and Mechanics UrB RAS using the Computational cloud platform of UrB RAS and supercomputer “URAN”. User workspaces are provided based on the virtual machines from Computational cloud platform of UrB RAS, which have Matlab GUI and Parallel Computing Toolbox installed. Matlab Distributed Computing Server is running on the supercomputer “URAN”. Integration of the virtual user workspaces with the cluster is done with the help of the SLURM resource manager. “Parallel Matlab” cloud service provides the integrated solution for parallel computation in the popular Matlab environment. Users do not have to install and configure parallel Matlab by themselves, the cloud service is ready to use and requires only a Web-browser. The installation and support of parallel Matlab is done by qualified cloud-platform administrators.

◄ ← Preview | 1 | 2 | 3 | 4 | 5 | 6 | Next → ►