Completed projects – Research Institute for Intelligent Computer Systems

“Virtual Master Cooperation Data Science” (ViMaCs)

Grant holder – Dortmund University of Applied Sciences (FH Dortmund)

Managers: prof. Dr. Carsten Wolf, Prof. Stefan Recker (FH Dortmund)

Partner universities:

Kyiv National University of Construction and Architecture (KNUBA) – prof. Sergey Bushuyev.
Ternopil National University of Economics (TNEU) – prof. Anatoly Sachenko
National University “Zaporizka Polytechnic” (NUZP) – prof. Galina Tabunshchik

Duration: 2019 – 2021.

The goal: to build a virtual training and laboratory infrastructure for online teaching and a portfolio of modules in the field of “Data Science”.

Task:

Building a common IT environment, including a common e-learning platform
Development of 4 online educational modules for 6 ECTS:

– KNUBA “Business analysis and decision-making”

– TNEU “Data Collection and Processing”

– NUSP “Artificial Intelligence and Data Analysis”

– FH Dortmund “Cloud data processing systems”

Pilot training in specialized schools 4 times a year
Pilot training of teachers for interuniversity distribution of modules
Creating a community of practice for continuous content evolution

TNEU team:

Anatoliy Sachenko;
Pavlo Bykovyy;
Oleksandr Osolinskyi;
Mykhailo Dombrovskiy;
Iryna Turchenko.

Link to project web-site: https://go-study-europe.de/vimacs/

DAAD program “International Study and Training Partnerships” (ISAP)

Grant holder: Berlin University of Applied Sciences (HTW Berlin), prof. Juergen Sieck.

The head of the program from TNEU is Doctor of Technical Sciences, Prof. Anatoliy Sachenko

Duration: 2019 – 2021.

Goal: strengthening partnership relations and cooperation between German universities and universities in the countries of the Middle East / Southeast and Eastern Europe, as well as in the Caucasus and Central Asia; promoting cooperation for the reconciliation of academic degrees (Bologna Process)

Main tasks:

strengthening existing and starting new stable partnerships;
exchange of scientists, graduates and students;
structural improvement of research and training conditions in partner countries;
internationalization of German and foreign universities.

TNEU team:

Anatoliy Sachenko;
Pavlo Bykoviy;
Iryna Turchenko.

Methods and means of structural-statistical identification of hierarchical objects by characteristic points of their contours

Principal investigator of project – PhD Diana Zahorodnia;

Co-investigator – PhD Hrystyna Lipyanina-Goncharenko.

Duration: 2019 –2020

Objectives: development of methods and tools aimed at improving the efficiency of automated video surveillance systems by reducing the amount of data processed.

Main results of the project:

An analysis of known methods of identification and classification of objects for video surveillance systems.
A method of combined adaptive identification of objects based on a hierarchical principle has been developed.
A method of data classification based on cluster analysis methods has been developed.
Algorithmic solutions of the proposed method of combined adaptive identification of objects based on the hierarchical principle are developed.
Algorithmic solutions of the proposed method of data classification based on cluster analysis methods are developed.
Experimental researches of the offered methods and algorithms are carried out.

Team:

Diana Zahorodnia;
Vitaliy Dorosh;
Hrystyna Lipyanina-Goncharenko;
Ivan Kit;
Andriy Kaniovskyi;
Denys Zolotukhin;
Andriy Sydor;
Dmytro Lendiuk.

Methods for intelligent processing and analysis of Big Data based on deep neural networks

Principal investigator of project – Prof. Anatoliy Sachenko;

Co-investigator – Dr Myroslav Komar.

Duration: 2018 –2019

Objectives: to increase the efficiency and performance of Big Data intelligent processing and analysis by developing effective methods of data compression and classification, and pattern recognition using deep neural networks.

Main results of the project:

Known methods of data protection against computer attacks were analyzed.
Data compression method based on deep neural networks was developed, using network traffic parameters in an intrusion detection system.
A method for data classification based on deep neural networks was developed in order to prevent attacks against information telecommunication networks.
An image recognition method was developed based on knowledghe of Big Data using deep neural networks.
A method for parallel deep neural network training was developed to solve the problems of Big Data compression and classification.
Algorithms of the proposed methods of intelligent processing and analysis of Big Data based on deep neural networks were proposed.
Deep neural networks architucture was proposed to solve the problems of Big Data compression and classification.
Experimental studies of the proposed methods and algorithms have been carried out

Team:

Anatoliy Sachenko;
Myroslav Komar;
Volodymyr Kochan;
Vasyl Koval;
Vladimir Golovko;
Vasyl Yatskiv;
Nadiia Vasylkiv;
Taras Lendyuk;
Pavlo Bykovyy;
Diana Zahorodnia;
Vitaliy Dorosh;
Oleksandr Osolinskyy;
Grygoriy Gladiy;
Oleksiy Roshchupkin;
Volodymyr Turchenko

Erasmus+ALIOT

Grantholder – Prof. Chris Phillips, Newcastle University, Newcastle, UK

National coordinator – Prof. Vyacheslav Kharchenko, National Aerospace University KhAI, Kharkiv

Leader of ICS TNEU team – Prof. Anatoliy Sachenko, ICS, Ternopil National Economic University

Duration: 2016 – 2020

Objectives: to develop and update curricula for masters, graduate students and industrial company specialists in the field of development, research and application of Internet of Things (IoT) in accordance with the needs of modern society.

Interim project results:

Three working meetings of all project participants were held in Chernivtsi, February, 2018; Kyiv, May, 2018, and Newcastle and Leeds, UK, July, 2018 to announce the interim results of the team and the tasks for a given period.
Curricula were developed.
The content of the developed courses and modules was discussed.
The structure of books and manuals was developed and discussed according to the proposed courses and modules.
Regular working meetings of the ICT-TNEU team were held (see information on the websites http://www.tneu.edu.ua/, iosu.tneu.edu.ua та www.ics.tneu.edu)

Team:

Anatoliy Sachenko;
Myroslav Komar;
Volodymyr Kochan;
Vasyl Yatskiv;
Vasyl Koval;
Grygoriy Gladiy;
Iryna Strubytska;
Zbyshek Dombrovskiy;
Mykhailo Dombrovskiy
Oksana Dunets;
Pavlo Bykovyy;
Diana Zahorodnia;
Oleksandr Osolonskyy;
Vitaliy Dorosh

DAAD programme “Eastern Partnerships”

Project Co-investigator: Prof. Anatoliy Sachenko

Co-investigator – Dr Iryna Turchenko

Duration: 2017 – 2019

Objectives:

Strengthening partnerships and cooperation between German HEI and HEI in the Middle East/ South Eastern and Eastern Europe as well as Caucasus and Central Asia
Fostering cooperation for alignment of academic degrees (Bologna
process)

Main project results:

Strengthening of existing and initiating new sustainable partnerships
Research, graduate and student exchanges
Sustainable structural improvement of conditions for conducting research and studying in partner-countries
Contribution to internationalisation of German and foreign HEI

Team:

Anatoliy Sachenko;
Pavlo Bykovyy;
Iryna Turchenko.

Theoretical Foundations and Hardware for Improving the Productivity of Wireless Sensor Networks

Principal investigator of project – Dr. Vasyl Yatskiv

Duration: 2017 – 2018

Objectives: The project is aimed at solving the scientific and applied problem of improving the productivity of Wireless Sensor Networks (VSN) by developing effective methods of noise-immune encoding and adaptive data transmission schemes, providing error-immune and asymmetric computing complexity methods of data compression. At the same time, important criteria for evaluating the developed methods are the following ones: hardware complexity, computational complexity and energy costs for the implementation of algorithms.

Main project results:

development of methods for correction of multiple errors based on modular correction codes with low computational complexity of the decoding algorithms;
study of computational complexity of the correction codes of the Residue Number System with a special system of modules;
development of the method of data transmission in WSN on the basis of adaptive error control scheme and modular correction codes;
investigation of the influence of noise on algorithms of data compression in WSN;
development of new data compression methods resistant to noise and error propagation during decoding with asymmetric computational complexity of coding algorithms (the complexity of coding algorithms is less than the complexity of decoding algorithms);
conducting experimental research of the transmission of compressed data under the influence of various types of noise;
development and implementation on the FPGA of the reconfigurable special processor of noise-immune data encoding on the basis of modular correction codes;
writing data compression algorithms in Verilog language and implementation of data processing devices in WSN on FPGA.

Team:

Vasyl Yatskiv;
Anatoliy Sachenko;
Volodymyr Kochan;
Mykhailo Kasyanchuk;
Natalia Yatskiv;
Ihor Yakymenko;
Stepan Ivasiev;
Orest Volynskyy;
Taras Tsavolyk.

Methods of Protection against Computer Attacks based on Neural Networks and Artificial Immune Systems

Principal investigator of project – Prof. Anatoliy Sachenko;

Co-investigator – Dr Myroslav Komar.

Duration: 2016 – 2017

Objectives: The development of a new intelligent information technology based on the theory of artificial neural networks, fuzzy logic and artificial immune systems to increase the reliability of computer attacks detection and classification.

Main project results:

An analysis of known methods of protection against computer attacks was carried out.
A modified method for constructing a detector of computer attacks based on neural networks and artificial immune systems was developed.
A method for reducing the amount of information based on neural networks of high trust with the use of multichannel neural network detectors for constructing a hierarchical classifier of computer attacks was developed.
A generalized architecture of intelligent computer-based system to prevent computer attacks was developed.
Experimental studies of developed methods and algorithms were conducted, which confirmed the reliability of detection and classification of computer attacks and improvement of the safety level.
An approach was proposed to improve the security of the system designed to prevent computer attacks by implementing neural network detectors on FPGA and introducing a subsystem of decision-making based on the rules of the Mamdani fuzzy inference.

Team:

Anatoliy Sachenko;
Myrolav Komar;
Volodymyr Kochan;
Vladimir Golovko;
Vasyl Yatskiv;
Lesia Dubchak;
Pavlo Bykovyy;
Diana Zahorodnia;
Vitaliy Dorosh;
Taras Tsavolyk;
Stepan Ivasiev;
Grygoriy Sapozhnyk;
Andriy Karachka.

Distributed Sensor Networks with Computing Nodes Reconfiguration

Principal investigator: Prof. Anatoliy Sachenko

Co-investigator: Dr. Igor Maykiv

Foreign partner: Technical University of Moldova, Moldova

Duration: 2014 – 2015

Objectives: Development of methods for structural synthesis of universal modules with the reconfiguration possibility.

Main project results:

Method for structural synthesis of universal modules comprising functional analysis, structural synthesis and the search for a set of optimal solutions was developed on the basis of morphological analysis and synthesis. The proposed method combines lexicographical criterion advantages (L-criterion) for the selection of electronic components during functional analysis and absolute criterion of preference (optimality Pareto, π-criterion) during the search for a set of optimal solutions that are considered in scientific literature as alternative methods for finding optimal solutions. The combination of L- and π-criteria allows us to reduce the number of alternatives synthesized during structural synthesis. A formalized discrete optimization solution is versatile for a wide range of problems of optimal structural synthesis of computing systems.
A new universal module structure with improved functional properties was designed due to separate data processing and sharing as well as reconfiguration of hardware and software using Field Programmable Gate Arrays (FPGAs).
A 4-level model that graphically shows information relationships between different processes of receiving and transmitting messages in the controller serial interfaces, which is an effective tool of their implementation both during functional analysis and structural synthesis, was developed.
An experimental model of network application processor with the capability of reconfiguring was created and the methodology of its testing was developed.

Team:

Anatoliy Sachenko;
Igor Maykiv;
Volodymyr Kochan;
Nadia Vasylkiv;
Oleksiy Roshchupkin;
Diana Zahorodnia;
Yuriy Ivanyshak;
Olexandr Osolinsky;
Taras Lendyuk;
Oksana Dunets.

Wireless Multimedia Sensor Networks on the Base of Modular Arithmetics and Galois Codes for Videomonitoring Systems

Principal investigator: Prof. Anatoliy Sachenko

Co-investigator: Dr. Vasyl Yatskiv

Foreign partner: Pedagogical University Huazhong, China.

Duration: 2013 – 2014

Objectives: developing of improved methods for training artificial neural networks on heterogeneous parallel computing systems referring to Grid, whih provide high efficiency of parallelization and development of grid-based library functions for paralel training of artificial neural networks.

Main project results:

New methods of data coding and transmitting based on modular arithmetic were developed, which enable increased efficiency of wireless multimedia sensor networks (WMSN). Methods were designed for devices with limited hardware resources and autonomous power supply.
Method of network coding is based on data of Residue Number System. The overall bandwidth of wireless sensor networks was investigated as well as the scope of data transmission schemes for different residues.
Method of coding and redundancy reducing of multimedia data without the loss in Residue Number System, which allows us to reduce image processing in 2-3 times by splitting the image into the modules of Residue Number System and parallel encoding of the obtained residues, was developed. Application of Huffman codes for residues compressing provides lossless compression ratio depending on the class of images: 1,6 – 4 – for photo-realistic images; 4 – 8 – for images with large areas of the same color.
Method of improving data reliability based on modified correcting code of Residue Number System, which is characterized by a lower computational complexity and allows us to increase the efficiency of encoding about 5 times comparing with R – source code RNS and Reed – Solomon RS (127, 87), was developed.

Team:

Anatoliy Sachenko
Yaroslav Nykolaychuk
Natalia Yatskiv
Vasyl Yatskiv
Orest Volynskyy
Petro Humenyy

Neural network method for improving the accuracy of information-measurement systems of ultraviolet radiation

Principal investigator: Prof. Anatoliy Sachenko

Project was completed within inter-university network Erasmus Mundus together with partners from Alaxender Ioan Kuza University, Iassi, Romania.

Duration: 2013 – 2014

Objectives: development of new neural network method for improving the accuracy of information measuring systems for measurement of ultraviolet radiation.

Research methods: structural and functional analysis (error analysis in measuring systems for measuring UV radiation level and UV sensors); methods of neural networks theory, the method of gradient ascent in the space of weight coefficients and neurons thresholds (for NN training); simulation methods (for experimental research of developed methods); technique for primary transformer investigation.

Project results:

The methods of signal processing of multiparameter sensors were proposed. Simulations were conducted in MathLab.
The software for modeling of the real multiparameter sensors behavior was developed. The software allows us to enter the model random and systematic errors and identify the limits of the proposed methods.
Ukrainian Patent application for invention and useful model was received.

Team:

Anatoliy Sachenko
Oleksiy Roshchupkin
Volodymyr Kochan

Methods and Tools of Building Wireless Multimedia Sensor Networks Based on Modular Arithmetic

Principal investigator – Prof. Yaroslav Nykolaychuk

Duration: 01.01.2013 – 31.12.2014

Objectives: development of methods and tools for data encoding and transmitting in wireless multimedia sensor networks aimed at improving the reliability of their operation and functionality.

Abstract: New methods and algorithms for data encoding and transmitting using mathematical tools of modular arithmetic were developed, aimed at improving the performance of wireless multimedia sensor networks (WMSM). A Verilog – encoder model for noise-immune data encryption using modified correcting codes was designed.

Main results:

The method of adaptive coding and transmission of multimedia data based on modular arithmetic and multipath routing using adaptive distribution packages and their transfer from multipath routing, is developed, which provides the efficiency of the total bandwidth of wireless sensor networks.
The method of network data coding based on the Residue Number System (RNS), which provides reduction of data amount by 50%, including the retransmission of packages that are necessary for message recovery, was developed. The proposed method allows us to select relatively simple modules of various bit-widths, though the bit-width of residues transmitted through the common route is approximately equal to the bit-width of residues on specific routes. The developed method of network coding improves overall network bandwidth by about 60%.
A modified correcting code of Residue Number System was developed, which is characterized by the simplified procedure of check symbols formation, providing increased efficiency of encoding approximately in 5 times as compared with other correcting codes. Using modified correcting codes of RNS in wireless sensor networks allows us to improve the reliability and overall network bandwidth by reducing the number of retransmissions.

Team:

Yaroslav Nykolaychuk
Anatoliy Sachenko
Vasyl Yatskiv
Natalia Yatskiv
Natalia Vozna
Petro Humenny
Orest Volynsky

Efficient Parallel Batch and Single Pattern Neural Network Training Algorithms Using Open MPI and GPU-computing

Principal investigator: Dr. Volodymyr Turchenko

Partners: Prof. Jack Dongarra, Innovative Computing Lab, University of Tennessee, Knoxville, TN, USA.

Grant: Fulbright Scholar Program 2012/13

Duration: 09/2012 – 06/2013

Objectives: test enhanced batch pattern parallel algorithm for NN training by changing the parameters of the internal algorithms of MPI collective functions on different parallel architectures;

develop GPU-based versions of the parallel batch and single pattern algorithms for NN training; test experimentally the efficiency of the improved GPU-based version of the algorithms in comparison with their Open MPI implementations.

Main results:

The parallelization efficiency of the neural network training algorithm on the example of the recirculation neural network model has been researched. The Open MPI, OpenMP and CUDA-based versions of the parallel batch pattern training algorithm for recirculation of neural network were implemented using C language. The parallelization efficiency of the developed algorithms has been researched on many-core parallel machine with 48 AMD Opteron 6180 SE processors, on computational cluster with 48 Intel Xeon E5520 processors, on 60-core Intel GPU Xeon Phi Coprocessor 5110P card and Nvidia Tesla C2050 GPU card using its 64 cores only (total is 1024). The experimental research of the developed algorithm using Open MPI technology showed the parallelization efficiency of 75% on 48 processors of the many-core system, 60% on 48 processors of the cluster, 70% on 60 processors of the Intel GPU Xeon Phi card. The experimental research of the developed algorithm using OpenMP technology showed lower figures, 40% of parallelization efficiency on 48 processors of the many-core system. The experimental research of the developed algorithm using CUDA technology showed 14-times speedup on one Nvidia Tesla GPU card. The developed algorithms are included to the developing library PaGaLiNNeT capable to speed-up scientific computations based on neural networks on general-purpose and hybrid (CPU+GPU) high performance computing systems.
The research project entitled “An Adaptive End-to-End Approach for Terabit Data Movement Optimization” was investigated. The goal of this project is to develop a novel architecture and related approaches to the end-to-end optimization of terabyte size data movement on next-generation networking and storage system technologies. The moving scientific data sets at terabits per second transfer rates over wide-area networks between geographically dispersed data centers were modeled. The set of events which describe a drop of the bandwidth in the communication network was obtained. A predictive model based on artificial neural networks to predict the duration of the event and the value of the maximum bandwidth drop was developed. I have used the developed library for parallel neural network training PaGaLiNNeT (developed by me within my previous project) and the model of a multi-layer perceptron. The experimental researches showed that the modeled events have stochastic nature and therefore it is necessary to tune the neural network model to provide desirable prediction results. This scientific collaboration with the host institution will be continued in the future.

Published results:

Turchenko V., Bosilca G., Bouteiller A. and Dongarra J. “Efficient Parallelization of Batch Pattern Training Algorithm on Many-core and Cluster Architectures”, Proceedings of the 7^th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems, Sep 12-14, 2013, Berlin, Germany, pp. 692-698.

Neural Network Methods for Evaluation of Microprocessor Power Consumption While Performing Instructions

Principal investigator: Dr. Zbyshek Dombrovsky

Duration: 2010 – 2012

Objectives: development of hardware-software complex, which allows to built mathematical models of processor cores power consumption.

Main tasks:

development of appropriate specialized hardware, which allows to evaluate power consumption of instruction execution in normal microprocessor operation mode;
development of testing methods (calibration) of created hardware;
using artificial neural networks to predict power consumption of the instruction execution modes (addressing, conditions, etc.) which were not completely tested experimentally;
using the experiment planning methods for additional decreasing of experiments volume.

Team:

Anatoliy Sachenko
Volodymyr Kochan
Andrii Borovyi
Oleh Havryshok
Ihor Maykiv
Orest Volynskyy

Published results:

Borovyi, V. Kochan, Th. Laopoulos, Sachenko A. Improved Sorting Methodology of Data-processing Instructions, International Journal of Computing, vol. 10, issue 1, 2011, pp. 50-55.
Borovyi, I. Maykiv, R. Kochan, Z. Dombrovskyy, V. Kochan. The Unit of Measurement of Consumers Pulse Energy, Patent of Ukraine 90922 UA, MPK (2009) G05F 5/00 G01K 17/00, no. А2008 06325 ; applied 13. 05. 2008; published 10. 06. 2010, Bulletin no. 11.
Time-domain analysis of ARM7TDMI core instructions [Text] / A. Borovyi, V. Kochan, Th. Laopoulos, A. Sachenko // Proceedings of the 6th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems (IDAACS’2011). – Vol. 2. – [S. L. : s. N.], 2011. – September 15-17. – P. 785 –790.

Human Biometric Identification in Video Surveillance Systems

Foreign partner: Technical University of Sofia, Bulgaria

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Bulgaria: Dr. Ognian Bumbarov

Duration: 2009 – 2010

Objectives: design of intelligent biometrical sub-system for detection and recognition of humen faces in the video surveillance systems for monitoring of public places, database support of staff or factory’s visitors etc.

Main tasks:

development of methods and algorithms for movement detection on the captured videoframes;
development of methods and algorithms of videoframes preliminary processing by skin color;
improvement of methods and algorithms for detection and tracing of human face;
development of methods and algorithms for face recognition.

Team:

Anatoliy Sachenko
Ihor Paliy
Yuriy Kurylyak
Taras Leshko

Published results:

Ihor Paliy, Anatoliy Sachenko, Yuriy Kurylyak, Ognian Boumbarov, Strahil Sokolov. Combined Approach to Face Detection for Biometric Identification Systems // Proceedings of 5^th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications, 21-23 September 2009, Rende (Cosenza), Italy, pp. 425-429.
Ognian Boumbarov, Strahil Sokolov, Plamen Petrov, Anatoliy Sachenko, Yuriy Kurylyak. Kernel-based Face Detection and Tracking with Adaptive Control by Kalman Filtering // Proceedings of 5^th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications, 21-23 September 2009, Rende (Cosenza), Italy, pp.434-439.
Kurylyak, I. Paliy, A. Sachenko, A. Chohra, K. Madani. Face Detection on Grayscale and Color Images using Combined Cascade of Classifiers // International Journal of Computing. –Ternopil (Ukraine). – 2009. – Vol. 8, Issue 1. – pp. 61-71.
Kurylyak A Real-Time Motion Detection for Video Surveillance System // Proceedings of 5^th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS2009). – Rende (Cosenza), Italy, 2009. – pp.386-389.
Paliy I.O. Methods of Face Detection in Systems of Computer Recognizing on the Base of Combined Cascade of Neural Network Classifiers. – PhD Thesis, Ternopil National Economic University. – Ternopil. – 2009.

Parallel Grid-aware Library for Neural Networks Training – PaGaLiNNeT

Principal investigator: Dr. Volodymyr Turchenko

Scientific advisor: Prof. Anatoliy Sachenko

Partners: Prof. Lucio Grandinetti, Center of Excellence on High Performance Computing, Department of Electronics, Computer Science and Systems, University of Calabria, Italy.

Grant No FP7 MC IIF 221524 – 908524 according to the 7^th EU Frame Programme, Grant of Marie Curie for researches from the third countries (International Incoming Fellowships – IIF), return phase

Duration: 2011 – 2012

Objectives: development of the enhanced training methods for artificial neural networks in heterogeneous parallel computing systems within the Grid; providing the high efficiency of parallelization and development of the Grid-based library for parallel neural networks training.

Main results:

As a part of the project design three levels of grid-based library are created: (i) at the level of a single supercomputer / cluster homogeneous computing nodes, (ii) at the heterogeneous computing nodes within a cluster, (iii) at the grid of computing system with heterogeneous hosts and heterogeneous communication channels between them. A parallel version of the library for the level (i) was installed on parallel machines with ccNuma architecture. A strategy for resource brokering based on Pareto optimization [1] is implemented in C programming language and included in the library. The developed library for the level (i) which includes the routines for parallel training of multilayer perceptron [2] and recurrent neural network was used for the prediction of the stock price for financial markets. The results are published in [6]. A parallel version of the library for the level (ii) was developed and installed on the computing cluster of heterogeneous architectures. The resource brokering sub-routine based on Pareto optimization [1] is called from the code of resource broker separately before executing the main task. The performance analysis of computing nodes of the cluster is based on a modified BSP-based model with improved computational complexity of parallel training algorithm for multilayer perceptron [2]. The results are published in [5];
Within the application of parallel algorithms for neural network training to speed up the execution of practical tasks, an application task of convolution neural network for the detection of the number of micronucleus in the human lymphocytes is considered. The accurate detection of the number of micronucleus in the human lymphocytes can be used as biological dosimeter in order to relive the presence and the action of carcinogenic factors and could enhance the correctness of the final medical response. It was proven the application of convolution NN for the development of this task because this NN model provides good detection properties and showed good detection results of the more complicated task of human face detection. The human lymphocyte images were acquired by the image flow cytometer which causes the different types of noise that influence on the acquired image. We have tested the CNN for the images altered by a zoom factor. The CNN provides no false alarms for each zoom factor. The number of false negative detections is much lower in comparison with the pattern matching method, implemented as a LABVIEW routine (IMAQ Match Pattern method) inside the flow cytometer. The detection rate of 87.5% provided by the CNN is much higher than 25% of detection rate by the IMAQ Match Pattern method on the considered example images. The results are published in [3, 4].

Published results:

Turchenko V.O. Brokering methodology of Grid-resources using Pareto-optimality // Measuring and Computing Technologies Equipment in Technological Processes. – 2011. # 1. – pp. 312-318.
Turchenko V.O. Efficiency Comparison of Multilayer Perceptron Group Training on Parallel Computer and Computation Cluster // Transactions KPI. Informatics, management and computing technology: Proceedings – Kyyiv: Vek+. – 2011. – No. 54. – pp. 130-138.
Paliy I., Lamonaca F., Turchenko V., Grimaldi D., Sachenko A. Detection of Micro Nucleus in Human Lymphocytes Altered by Gaussian Noise Using Convolution Neural Network, Proceedings of 2011 IEEE International Instrumentation and Measurement Technology Conference (I2MTC 2011), 2011, Binjiang, Hangzhou, China, pp. 1097-1102.
Lamonaca F., Turchenko V., Grimaldi D. Aspetti innovativi della progettazione hardware e software di citofluorimetro ad immagini, Atti del XXVIII Congresso Nazionale Gruppo Misure Elettriche ed Elettroniche, 2011, Genova, Italy, pp. 289-290.
Turchenko V., Puhol T., Sachenko A., Grandinetti L. Cluster-Based Implementation of Resource Brokering Strategy for Parallel Training of Neural Networks, Proceedings of the 6^th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems IDAACS2011, Sep 15-17, 2011, Prague, Czech Republic, pp. 212-217.
Turchenko V., Beraldi P., De Simone F., Grandinetti L. Short-term Stock Price Prediction Using MLP in Moving Simulation Mode, Proceedings of the 6^th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems IDAACS2011, Sep 15-17, 2011, Prague, Czech Republic, pp. 666-671.
Turchenko V. Efficiency Comparison of Batch Pattern Training Algorithm of Multilayer Perceptron on Parallel Computer and Computational Cluster, Scientific Journal of National Technical University of Ukraine “Kyiv Polytechnic Institute”, Kyiv, 2011, No 54, pp. 130-138 (in Ukrainian).
Sachenko A., Kulakov Yu., Kochan V., Turchenko V., Bykovvy P., Borovyy A. Computer Networks: A Tutorial, Ternopil, Ekonomichna dumka, 2012, 476 p. // Chapter 15. Grid-computations based on network technologies, pp. 416-439 (in Ukrainian).
Turchenko V., Grandinetti L., Sachenko A. Parallel Batch Pattern Training of Neural Networks on Computational Clusters, Proceedings of the 2012 International Conference on High Performance Computing & Simulation (HPCS 2012), July 2 – 6, 2012, Madrid, Spain, pp. 202-208.
Turchenko V., Golovko V., Sachenko A. Parallel Batch Pattern Training of Recirculation Neural Network, Proceedings of the 9^th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2012), July 28 – 31, 2012, Rome, Italy, pp. 644-650.
Turchenko V., Golovko V., Sachenko A. Parallel Training Algorithm for Radial Basis Function Neural Network, 7^th International Conference on Neural Networks and Artificial Intelligence (ICNNAI’2012), October 10-12, 2012, Minsk, Belarus, pp. 47-51.

Development of Intelligent Video Surveillance Systems

Principal investigator: Dr. Volodymyr Kochan

Project executed together with the Glushkov Institute for Cybernetics, Prof. Vitaliy Boyun.

Duration: 2009 – 2010

Objectives: development of highspeed and relevant video surveillance system on the basis of intelligent videocamera, which allows us to decrease information streams between camera and worksatation central processor, as well as to read and process large images with high frame rate.

Main tasks:

increasing of efficiency of communication channels between intelligent videocamera and personal computer;
development of methods and algorithms for videoframes preliminary processing by skin color and movement;
development of methods and algorithms for human face recognition on the basis of the combined cascades classifiers, classifier training paralleling, and improvement of neural network training method in the frame of combined cascade;
development of algorithms for faces tracing;
development of software and highlevel programe interface for interaction with intelligent camera; coding of developed algorithms in processor computer code for digital processing of intelligent videocamera images.

Team:

Anatoliy Sachenko
Ihor Paliy
Yuriy Kurylyak

Published results:

Kurylyak Y.O., Sachenko A.O. Method of background image renewal for movement segmentation // Proceedings of 10-th International Conference “Modern Information and Electronic Technologies” (SIET’2009). – Odessa (Ukraine), 2009. – Vol. 1. – pp. 44.
Paliy I.O. Training of neural network classifiers with combined cascade for face detection // Proceedings of 10-th International Conference “Modern Information and Electronic Technologies” (SIET’2009). – Odessa (Ukraine), 2009. – Vol. 1. – pp. 42.
Paliy I. Face detection on grayscale and color images using combined cascade of classifiers // International Journal of Computing. – 2009. – Vol. 8. – Issue 1. – pp.61-71.

Development of 3D Localization Methods for Navigation of Mobile robot

Foreign partner: Kaunas Technical University, Lithuania

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Lithuania: Prof. Rimvydas Simutis

Duration: 2009 – 2010

Objectifes: developing the unified structure for autonomous mobile robot control and providing 3D localization and navigation in non-structured environment with dynamical objects by using new methods and means which allow us to improve the navigation characteristics of mobile robots and use already known methods for new applications.

Main tasks:

Analysis of already known methods for designing the structure of control system for mobile robots (MR) and development of unified structure for autonomous MR control.
Development of Dataflow Diagram (DFD) for robot control system and analysis of time characteristics of DFD main modules. Setting of requirements for main MR modules.
Development of improved methods and means of MR control system:
1. Development of new method of acquisition and processing of sensor data;
2. Development of MR 3D localization methods.
Development of hardware and software for autonomous MR.
MR composing according to the requirements set in point 2, taking into account the applied problems and MR hardware/software means developed in points 3-4.
Verification and testing of MR prototype functioning.

Team:

Anatoliy Sachenko
Vasyl Koval
Oleh Adamiv
Viktor Kapura

Published results:

Roth H., Sachenko A., Koval V., Chanim J., Adamiv O., Kapura V. The 3D Mapping Preparation using 2D/3D Cameras for Mobile Robot Control // Artificial Intelligence journal, Donetsk, Ukraine. – 2008. – Vol. 4. – pp. 512-521.
Adamiv O., Sachenko A., Kapura V. Gradient Method for Autonomous Robot Navigation // Proceedings of the Ninth International Conference “Modern Problems of Radio Engineering, Telecommunications and Computer Science” (TCSET’2008). – Lviv-Slavsko (Ukraine), 2008. – pp. 640-642.
Adamiv, V. Koval, V. Dorosh, G. Sapozhnyk, V. Kapura Mobile Robot Navigation Method for Environment with Dynamical Obstacles // Proceedings of the 5-th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS’2009). – Rende (Cosenza), Italy, 2009. – pp.515-518.
Adamiv, A. Lipnickas, A. Knyš. A stereovision system for autonomous robot navigation in 3-D // Proceedings of 10-th International Conference “Modern Information and Electronic Technologies” (SIET’2009). – Odessa (Ukraine), 2009. – Vol. 1. – pp. 28.

Development of Stereovision Methods and Devices for Autonomous Navigation of Mobile Robots

Foreign partner: University of Sigen, Germany

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Germany: Prof. Hubert Roth

Duration: 2008 – 2009

Objectives: Development of stereovision methods for autonomous navigation of mobile robots.

Main tasks:

Development of stereo camera preliminary data processing methods for future integration with a mobile robot:
Methods of generation of stereo images;
Image filtering and analysis methods.
Development of stereo image fusion and mobile robot 3D environment map generation techniques:
Image description methods;
Stereo image corresponding points search and 3D environment map generation methods.
Development and implementation of sensor data fusion algorithms.
Verification and testing of the developed methods using a mobile robot.

Team:

Anatoliy Sachenko
Vasyl Koval
Oleh Adamiv
Viktor Kapura

Published results:

Roth H., Sachenko A., Koval V., Chanim J., Adamiv O., Kapura V. The 3D Mapping Preparation using 2D/3D Cameras for Mobile Robot Control // Artificial Intelligence journal, Donetsk, Ukraine. – 2008. – Vol. 4. – pp. 512-521.
Adamiv O., Sachenko A., Kapura V. Gradient Method for Autonomous Robot Navigation // Proceedings of the Ninth International Conference “Modern Problems of Radio Engineering, Telecommunications and Computer Science” (TCSET’2008). – Lviv-Slavsko (Ukraine), 2008. – pp. 640-642.
Roth, A. Sachenko, V. Koval, O. Adamiv, V. Kapura Evaluation of Camera Calibration Methods for Computer Vision System of Autonomous Mobile Robot // Proceedings of 10-th International Conference “Modern Information and Electronic Technologies” (SIET’2009). – Odessa (Ukraine), 2009. – Vol. 1. – pp. 29.

Development of Design and Optimization Methods for Breach Detection Systems

Foreign partner: Institute of Technology, Gebze, Turkey

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Turkey: Dr Serkan Aksoy

Duration: 2008 – 2009

Objectives: development of a Computer Aided Design (CAD) system for development of perimeter security systems optimized for quality-price, reliability-price criteria and further testing of the CAD system on real security systems.

Main tasks:

Analysis of existing solutions and formation of a set of criteria and limitations for functional and cost analysis of security systems. Development of improved components and database for security systems.
Development of methods and algorithms for structural synthesis and multi-criteria optimization of security systems. Development of a CAD system for security systems design based on the developed methods and algorithms.
Development of a pilot security system with the use of the developed CAD. Testing of the pilot system.
Carrying out a comparative analysis of the developed pilot system against existing systems. Introduction of necessary changes to the CAD system based on the conducted analysis.
Carrying out the pilot security system testing to measure risks of undetected intrusions and risks of false alarms. Introduction of necessary changes to the pilot security system based on the conducted tests.
Testing of the CAD system.

Team:

Anatoliy Sachenko
Volodymyr Kochan
Volodymyr Turchenko
Pavlo Bykovyy

Published results:

Bykovyy P. Design optimization of distributed technical security systems using a genetic algorithm // Visnyk of Vinnitsa Polytechnic Institute. – 2008, Issue #6, pp 28-34.
Bykovyy P., Pigovsky Yu., Kochan V., Sachenko A., Markowsky G., Aksoy S. Genetic Algorithm Implementation for Distributed Security Systems Optimization // Proceedings of the IEEE International Conference on Computational Intelligence for Measurement Systems and Applications (CIMSA 2008), 14-16 July 2008. – Istanbul, Turkey. – pp. 120-124.
Bykovyy P.Ye., Kochan V.V. Cryptographically secure protocol for networks of security sensors // Proceedings of 10-th International Conference “Modern Information and Electronic Technologies” (SIET’2009). – Odessa (Ukraine), 2009. – Vol. 1. – pp. 189.
Bykovyy P.Ye. Distributed sensor network for security systems // International journal of Computing. – Ternopil (Ukraine). – 2009. Vol. 8, Issue 2. – pp. 157-164.
Bykovyy, V. Kochan, Y. Kinakh, A. Sachenko, O. Roshchupkin, S. Aksoy, G. Markowsky. Data Communication Crypto Protocol for Security Systems Sensor Networks // Proceedings of 5^th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems(IDAACS’2009). – Rende (Cosenza), Italy, 2009. – pp. 375-379.
Bykovyy, Y. Pigovsky, A. Sachenko, A. Banasik. Fuzzy Inference System for Vulnerability Risk Estimation of Perimeter Security // Proceedings of 5^th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems (IDAACS’2009). – Rende (Cosenza), Italy, 2009. – pp. 380-384.

Computer Telecommunication System Based on Noise Signals

Principal investigator: Prof. Yaroslav Nikolaychuk

Project is executed together with JSC Ternopil Radio Plant ‘Orion’, chief designer Volodymyr Kordyak.

Duration: 2007 – 2009

Objectives: to increase noise-immunity and active range of radio stations, produced by the Orion plant; introduce a mode of code based on division of transmission channels; develop a computerized system of data acquisition based on autonomous sensors.

Project tasks:

Design of a noise-signal based radio station with a low range of operation for construction companies;
Analysis of possible application areas for 2D noise signals;
Analysis of possible application areas and prospective customers of computer systems based on autonomous sensors.
Preparation of project solutions related to radio system serving and construction areas.

Team:

Yaroslav Nykolaychuk
Oleh Zastavnyy
Nazar Krutskevych

Published results:

Nykolaychuk Y., Krutskevych N., Zastavniy O. Multibases Processors of Two-dimensional Correlation for Noise Immunity of Transfer Information // Proc. Of the IEEE International Workshop on Intelligent Data Acquisition and Advancing Computing Systems (IDAACS’2007). – 2007. – Dortmund (Germany). – pp. 315-317.

Dynamically Reprogrammable Network Capable Application Processor with Internet Capability

Foreign partner: Esensors Inc., USA

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from USA: Dr Darold Wobschall, PhD

Grant #UE2-2534-TE-07.

Duration: 2007 – 2009

Objectives: to enter the US smart sensors market with the Network Capable Application Processor (NCAP) developed within the project CRDF #UE2-2534-TE-03 – device oriented on software data processing in smart distributed measurement and control systems which uses adaptive software reconfiguration for intelligent functions execution (self-adapting and self-training). The developed NCAP will have the following features:

ability to work in distributed measurement control systems utilizing the Internet;
online remote reprogramming of user application software;
support of a wide set of network interfaces;

Main tasks:

the minimal set of the design documentation sufficient for production of a prototype NCAP was developed;
two prototype NCAP devices have been developed and undergo testing;
testing of certain NCAP modules was performed, the NCAP software was developed as well.

Team:

Anatoliy Sachenko
Volodymyr Kochan
Roman Kochan
Andrew Stepanenko
Ihor Maykiv
Iryna Turchenko
Natalia Vozna

Published results:

Maykiv I., Stepanenko A., Wobschall D., Kochan R., Kochan V., Sachenko A., Vasylkiv N. Remote Reprogrammable NCAPs: Issues and Approaches // Proc. Of the IEEE International Workshop on Intelligent Data Acquisition and Advancing Computing Systems (IDAACS’2007). – 2007. – Dortmund (Germany). – pp. 109-113.
Maykiv I.M., Kochan V.V., Bilousov I.A. Project analysis of methods of serial interfaces controllers realization // Transactions of Ternopil State technical University. – 2009. – No. 1. – pp. 110-115.
Maykiv I.M. Investigation of I2C interface controllers realizations method on the programmed logical matrix // Proceedings of 5-th International Youth Conference “Modern Problems of Radiotechnics and Telecommunication”. – Sevastopol (Ukraine), 2009. – pp. 284.
Maykiv I.M., Kochan V.V. Software-hardware controller of consecutive interfaces in network nodes of data acquisition // Proceedings of 10-th International Conference “Modern Information and Electronic Technologies” (SIET’2009). – Odessa (Ukraine), 2009. – Vol. 1. – pp. 138.
Maykiv I.M. Methodology of structural synthesis of netwok capable application processors // Proceedings of National Conference in Ternopil Ivan Pul’uj State Technical University. – Ternopil (Ukraine), 2009. – pp. 176.
Maykiv I.M. Software-hardware method of secuential interfaces controllers realization // Proceedings of 11-th International Conference “System Analysis and Information Technologies” (SAIT-2009). – Kyyiv (Ukraine), 2009. – pp. 437.
Maykiv I.M. Netwok capable application processor for distributer measuring-control systems // Transaction “Problems of Informatization and Control”, Kyyiv (Ukraine). – 2009. – No. 2 (28). – pp. 187-191.
Maykiv I.M. Universal controlle of serial interfaces // Transactions of Chernivstsi University. Series: Physics. Electronics, Chernnivtsi (Ukraine). – 2009. – No. 3 (186). – pp. 130-135.
Maykiv I.M., Stepanenko A.V., Wobschall D. A method for structural synthesis of network capable application processors. // International Journal of Computing – Ternopil (Ukraine). – 2009. – Vol. 8. – Issue 2. – pp.126-138.
Maykiv, D. Wobschall, A. Stepanenko, R. Kochan, A. Sachenko, V. Kochan. Multi-port Serial NCAP using IEEE1451 Smart Transducer Standard // Proceedings of IEEE Sensor Application Symposium (SAS-2009). – New Orleans, LA, (USA), 2009. – pp. 293-297.
Maykiv, A. Stepanenko, D. Wobschall, R. Kochan, V. Kochan, A. Sachenko. Universal Controller of Serial Interfaces // Proceedings of the 5-th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS’2009). – Rende (Cosenza), Italy, 2009. – рp. 121-125.
Iryna Turchenko. Methods for Improving Efficiency of Data Processing Obtained from Multi-parameter Sensors in Distributed Computer Systems. Ph. D. Thesis on speciality 05.13.05 – Computer Systems and Components.- Ternopil National Economic University.- Ternopil.- 2008.- 200 p. (in Ukrainian)
Natalia Vozna. Forming and Organizing of Structured Data Movement in Multilevel Distributed Computer Systems. Ph. D. Thesis on speciality 05.13.05 – Computer Systems and Components.- Ternopil National Economic University. – Ternopil. – 2009. (in Ukrainian)

Ternopil Education Communication Center

Foreign partner: University of Maine, USA

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from USA: Prof. George Markowsky

Project is granted by NATO Program of Security through Science Network Infrastructure Grant, and performed together with the University of Maine, USA.

Duration: 2006 – 2009

Objectives: create common communication center for universities of Ternopil, integrate educational networks of Ternopil Universities, introduce high-speed network for training and research.

Main tasks:

Connect educative institutions of Ternopil to Internet through Ternopil Education Communication Center;
Make a basis for cooperation of all universities of Ternopil;
Make a basis for educative and research cooperation between universities of Ternopil and University of Maine and other researchers;
Provide high-speed access to UARNET and GEANT networks;
Provide abilities for holding video-conferences between Ternopil and other cities;
Develop a prototype of a system, that can be implemented in other areas of Ukraine;
Implement 16 processor clusters for GRID-processing that will be used in universities – project members;
Introduce on-line library;
Provide Wi-Fi service for universities of Ternopil.

Team:

Anatoliy Sachenko
Serhiy Voznyak
Ihor Romanets’
Roman Romanyak

Published results:

Sachenko A. Ternopil Education Communication Center // Innovation and Communication Security (ICS) Panel Meeting. – 2006. – Kyiv (Ukraine).
Markowsky, A. Sachenko, S. Voznyak, V. Spilchuk, R. Romanyak, V. Turchenko, I. Romanets. The Ternopil Educational Communication Center – A NATO Project to Integrate Regional Information Technology Resources. Computing, 2008, Vol. 7, Issue 1.
Palagin O., Alishov N., Markowsky G., Sachenko A., Turchenko V. Security Tools for GRID-systems // Proceedings of the 2007 International Conference on Security ans Management. -2007. Las Vegas, NV (USA).

Instruction Parameters Analysis for Power Modeling of Embedded Microprocessors

Foreign partner: Aristotle University of Thessaloniki, Thessaloniki, Greece

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Greece: Prof. Theodore Laopoulos

Project is granted by Ministry of Education and Science of Ukraine and Greek Government (agreement #М/85-2006), and performed together with the Aristotle University of Thessaloniki, Greece.

Duration: 2006 – 03.2008

Objectives: to determine power consumption of each parameter while executing the following instructions by the processor: determining number and value of registers, immediate values, values and addresses of operands, address of command call, pipeline panel and substitution, examination and analysis of correlation of instruction parameters in power consumption of instructions; examination and analysis of each parameter in power consumption of instructions; developing accurate power models for execution level of ARM7TDMI processor instructions.

Main tasks:

Additional investigating of instruction parameters power consumption and developing of measurement methodology using existing measurement setup; developing new approach in measurement methodology that can determine processor configuration. Due to this approach it is possible to measure and analyze correlation of instruction power consumptions according to instruction parameters; determine power consumption; analyze and process power consumption values; develop power models for instructions; experimentally prove achieved theoretical results.

Team:

Anatoliy Sachenko
Volodymyr Kochan
Volodymyr Turchenko
Andrii Borovyi

Published results:

Borovyi A., Kostandakos V., Kochan V., Sachenko A., Yaskilka V. Analysis of CPU’s Instructions Energy Consumption Device Circuits // Proceedings of Fourth IEEE International Workshop on Intelligent Data Acquisition and Advancing Computing Systems (IDAACS’2007). – 2007. – Dortmund (Germany). – pp. 42-46.
Borovyi A., Kochan V. Analysis of Microcontroller Instructions Power Consumption Measurement Circuits. Visnyk of Khmelnytskyy National University. – 2007. – Vol. 1. – #2. – pp. 105-109.
Borovyi A.M., Kochan V.V., Turchenko V.O. Stand for investigation of current moment value consumed by microprocessor // Transaction of Ternopil State Technical University. – 2009. – No. 1. – pp. 131-137.
Borovyi A.M. Analysis of power consumption by ARM7TDMI processor kernel // Proceedings of National Conference in Ternopil Ivan Pul’uj State Technical University. – Ternopil (Ukraine), 2009. – pp. 101.
Borovyi, V. Kochan, Z. Dombrovskyy, V. Turchenko, A. Sachenko Device for Measuring Instant Current Values of CPU’s Energy Consumption // Proceedings of the 5-th IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS’2009). – Rende (Cosenza), Italy, 2009. – рр.126-130.

Financial Analytics Method with Applications of Knowledge Bases

Principal investigator from RIICS: Prof. Anatoliy Sachenko

This is a joint project between National University of the State Taxation Department of Ukraine, Irpin, Ukraine and Research Institute for Intelligent Computer Systems, Ternopil, Ukraine.

Duration: 09.2008 – 11.2008

Objectives: evaluation of the present state and selection of priority directions for implementation of intelligent information technologies of financial analytics and knowledge bases in governmental resource management processes.

Main tasks:

evaluation of the present state and investigation of theoretical research in information technologies for financial analytics with application of knowledge bases in management of governmental institutions;
investigation of possible intelligent computer technologies application in the domain of financial analytics ontologies in governmental management;
evaluation of the state and perspectives of ontology intelligent tools using methods of financial analytics;
development of technologies in area of intelectulazation of information-analytical processes and creation of financial analytics knowledge bases in governmental management;
the conducted activity enabled to provide functional completeness of solutions to the defined research tasks and creation of documentation as per the Requirements Specification;
research and creation of the output documentation were performed on the basis of a systematic approach, conceptual completeness of results and consistency;
the conducted work follows the principal of minimal implementation costs for the proposed solutions.

Team:

Anatoliy Sachenko
Taras Lendyuk

Published results:

Palagin A., Rippa S. and Sachenko A. Conceptualization and problems of ontologies // Journal of Artificial Intelligence, 2008 Vol. 3, pp 374-379.

Development of Effective GRID-technologies for Ecology Monitoring Using Satellite Data

Principal investigator from ICS: Prof. Anatoliy Sachenko

Principal investigator NSAU: Prof. Nataliya Kussul

Collaborative project of Scientific-Technologic Centre in Ukraine and National Sciences Academy of Ukraine has been performed together with the Space Research Institute of National Sciences Academy of Ukraine and National Aerospace Agency of Ukraine, Kyiv.

Grant STCU #3872

Duration: 12.2005 – 12.2007

Objectives: Development of an effective distributed computations techniques that provide simple and transparent solutions to the computationally-complicated tasks in different areas, especially associated with space data processing.

Main tasks:

developing methodology for constructing temporal interpolation earth atmosphere photographs;
developing methodology for predicting solar activity and corresponding algorithms for holding parallel computations;
developing parallel implementation modeling methods algorithms for dynamics of main processes in multi-component ground environments with the corresponding cluster.
developing GRID-service for monitoring and control solutions process in systems;
developing GRID-service for balancing system loading;
developing GRID-service for visualization of computational results;
developing GRID-service for granting users’ access to system;
developing service for system security purposes;
combining some clusters or computational networks into one complex for searching solution to the same task.

Team:

Anatoliy Sachenko
Volodymyr Turchenko
Viktor Demchuk

Published results:

Turchenko V., Demchuk V., Sachenko A. Interplanetary Shock Arrival Time Prediction Using Multi-Layer Perceptron // Proceedings of the 4^th IEEE Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications IDAACS’2007. – 2007. – Dortmund (Germany). – pp. 185-190.
Turchenko V. An Approach to IP Shock Arrival Time Prediction Using Approximating Neural Network // International Journal of Information Technology and Intelligent Computing. – 2007. – No. 4. – Vol. 1.
Turchenko, V. Demchuk, A. Sachenko, Y. Veremeyenko. An Approach to Interplanetary Shocks Prediction Using Single ACE/EPAM Channel Data // Proceedings of the Fourth International Conference on Neural Networks and Artificial Intelligence ICNNAI`2006. – 2006. – Brest (Belarus). – pp. 140-144.

Development of Web Ontologies as Data Exchange and Decision Support Tools to Facilitate Economic Cooperation between Ukraine and USA

Foreign partner: New Jersey Institute of Technology, USA

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from USA: Dr. Yefim Kats

Project had been performed according to Research program of the US National Science Foundation.

Grant # NSF-04-12

Duration: 2004 – 2007

Objectives: develop Web-ontologies as data exchange and decision making instrument for promotion of economic partnership between Ukraine and USA.

Main tasks:

Standard ontology dictionary used in economic interchange, including dictionaries for typical e-commerce models, identification.
Identifying objects as classes or relations with adequate limiting interpretation.
Identifying specific ontology relations for (intelligence) agents based on automated processing.
Developing Windows object library compatible apparatus for measuring possible ontology errors.

Team:

Anatoliy Sachenko
Roman Pasichnyk
Yuriy Pihovsky
Andrii Melnyk

Published results:

Pasichnyk R., Sachenko A. Semantic WEB-Search Developing by Problem-Oriented Ontology Means // Proceedings of the IEEE International Workshop IDAACS’2007. – 2007. – Dortmund (Germany). – pp. 445-448.
Hrusha V. Specifics of Ontologies Design and Application in proceedings of the 11^th scientific conference of Ternopil State Technical University. – 2007. – Ternopil: TSTU. – pp. 78.
Pasichnyk, А. Sachenko, А. Melnyk “Formalization of ontology creation process using base classes” in proceedings of the 13^th national conference “Modern problems of applied mathematics and informatics”, Lviv, October 3-5 2006, P.162-163.
Master thesis by Andrii Melnyk was defended in 2006.
Course thesis by Andrii Melnyk was defended in 2005.
Master thesis by Vitaliy Kharchuk was defended in 2004.

Dynamically Reprogrammable Network Capable Application Processor with Internet Capability

Principal investigator: Prof. Anatoliy Sachenko

The project is funded under the Ministry of Education and Science of Ukraine

Grant #0107U005985.

Duration: 08.2007 – 12.2007

Objectives: to enter the US smart sensors market with the Network Capable Application Processor (NCAP) developed within the project CRDF #UE2-2534-TE-03 – device aimed at software data processing in smart distributed measurement and control systems which uses adaptive software reconfiguration for intelligent functions execution (self-adapting and self-training). The developed NCAP will have the following features:

ability to work in distributed measurement control systems utilizing the Internet;
online remote reprogramming of user application software;
support of a wide set of network interfaces.

Main tasks:

a minimal set of the design documentation sufficient for production of a prototype NCAP had been developed, which allowed to choose its elemental basis and embodiment;
there was developed a package of structural documentation;
there was developed software for interface microcontroller providing software support of hardware drivers for supported interfaces – data link layer, IР protocol (Internet Protocol) – network layer, TCP protocol (Transport Control Protocol) – transport layer, НТТР protocol (Hypertext Transfer Protocol) – session layer, dynamical HTML-page, where the data is presented and received by all supported interfaces and can be read – presentation layer;
two prototype NCAP devices had been developed and underwent testing that allows to debug application software of its microcontrollers and their interaction between each other, as well as with the server and measuring-control modules in real time.

Team:

Anatoliy Sachenko
Volodymyr Kochan
Roman Kochan
Andrew Stepanenko
Ihor Maykiv
Pavlo Bykovyy

Published results:

Maykiv I., Stepanenko, Wobschall D., Kochan R., Kochan V., Sachenko A., Vasylkiv N. Remote Reprogrammable NCAPs: Issues and Approaches // Proc. Of the IEEE International Workshop on Intelligent Data Acquisition and Advancing Computing Systems (IDAACS’2007). – 2007. – Dortmund (Germany). – pp. 109-113.
Stepanenko A., Maykiv I., Wobschall D., Kochan R., Kochan V., Sachenko A, Multi-port Serial NCAP Using IEEE1451 Smart Transducer Standard // Proceedings of the IEEE Sensor Application Simposium SAS’2009, 17-19 February, 2009, New Orleans, USA, pp. 293-297.

Investigation of the Intelligent Properties of Re-Configurable Network Capable Application Processor in Adaptive Distributed Instrumentation and Control Systems

Foreign partner: Sensors Development and Applications Group, National Institute Standards and Technologies, USA

Principal investigator from Ukraine: Dr. Volodymyr Kochan

Principal investigator from USA: Kang Lee

This project has been performed within US Civilian Research and Development Foundation (Cooperative Grant Program).

Grant # CRDF.CGP. UE2-2534-TE-03

Duration: 2005 – 2006

Objectives: Development of the IEEE-1451 standard compatible Network Capable Application Processor (NCAP) with dynamic software and hardware reconfiguration and investigation of its self-adaptive and intelligent properties in information-measurement systems.

Main tasks:

Investigation of the NCAP intelligent properties to be used with smart sensors, deployed in distributed information measurement systems with different architectures and functional requirements.
Extension of the NCAP’s functional features compatible with the IEEE1451 standard to support dynamic online reprogramming of software and a set of network interfaces.
Development and investigation of the prototype NCAP and its programming methodology.

Team:

Volodymyr Kochan
Anatoliy Sachenko
Roman Kochan
Oleh Adamiv
Iryna Turchenko
Andriy Stepanenko

Published results:

Kochan V., Lee K., Kochan R., Sachenko A. Approach to Improving Network Capable Application Processor Based on IEEE 1451 Standard // Computer Standards & Interfaces. – 2005. – 28. – Issue2. – pp. 141-149.
Stepanenko A., Lee K., Kochan R., Kochan V., Sachenko A. Development of a Minimal IEEE1451.1 Model for 8051-Compatible Microcontrollers // Proc. Of the 2006 IEEE Sensors Applications Symposium. – 2006. – Houston, Texas (USA). – pp. 88-93.
Kochan R., Kochan V., Sachenko A., Maykiv I., Turchenko V, Markowsky G. Interface and Reprogramming Controller for Dynamically Reprogrammable Network Capable Application Processor (NCAP). // Proc. Of 3-th IEEE International workshop on Intelligent Data Acquisition and Advancing Computing Systems (IDAACS’2005). – 2005. – Sofia (Bulgaria). – pp. 639-642.
Kochan R., Kochan V., Sachenko A., Maykiv I. NCAP Based on FPGA // Proc. Of the IEEE Instrumentation and Measurement Technology Conference IMTC/2005. – 2005. – Ottawa, Ontario (Canada). – pp. 813-817.
Kochan R., Lee K., Kochan V., Sachenko A. Development of a Dynamically Reprogrammable NCAP // Proc. Of the IEEE Instrumentation and Measurement Technology Conference IMTC/2004. – 2004. – Como (Italy). – pp. 1188-1193.
Roman Kochan. Improvement of components of precision distributed information control systems: Ph.D. Theses on speciality 05.11.16 / Ternopil Academy of National economy. – Ternopil, 2005. – 193 p.

Methods and Algorithms for Face Detection and Recognition for Real Time Video Surveillance Systems

Foreign partner: Belarus State University of Informatics and Radio Electronics, Belarus

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Belarus: Prof. Rauf Sadykov

This project has been performed in frames of State fund of fundamental research programs, Ministry of Education and Science of Ukraine order #356 dated to 14.06.05.

Duration: 2005 – 2006

Objectives: Development of algorithms for preliminary processing of images based on segmentations and development of algorithms and software for face detection in static vision conditions.

Main tasks:

Development of effective algorithms and software for capturing face images in video stream;
Development of approximate 3-dimension face models;
Development of algorithms for selection of informative features and classification of images according to modified syntactical discriminator functions;
conducting experimental diagnosis and configuration of proposed algorithms for achieving maximum results of program model;
development of a software system which implements the designed recognition scheme.

Team:

Anatoliy Sachenko
Vasyl Koval
Ihor Paliy
Yuriy Kurylyak
Victor Kapura

Published results:

Kurylyak. System of Face Detection at Static Images. – 2006. – 83p.
Kurylyak, Ihor Paliy, Vasyl Koval, Anatoliy Sachenko. Improved Method of ace Detection Using Color Images // Proceedings of the International Conference “Modern Problems of Radio Engineering, Telecommunications and Computer Science” TCSET’2006. – Feb’28 – Mar’4, 2006. – Lviv-Slavske, Ukraine. – pp. 186-188.
Sachenko, V. Koval, I. Paliy, Y. Kurylyak. Approach to Face Recognition Using Neural Networks // Proceedings of the IEEE Second International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications IDAACS’2005, Sofia, Bulgaria, September 5-7, 2005, pp. 112-115.

Development of Methods and Tools for Improvement of Robot Navigation in a non-Structured Environment

Foreign partner: Kaunas Technical University, Lithuania

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Lithuania: Dr Arunas Raudis

This project has been performed in frames of State Fund for Fundamental Research Programs, Ministry of Education and Science of Ukraine order #174 dated by 23.03.05.

Duration: 2005 – 2006

Objectives: Development of methods and tools for improvement of mobile robot navigation in non-structured environment.

Main tasks:

Development of methodology for creation of a mobile robot management system, which reflects schemes for conforming mobile robot subsystems for ensuring unobstructed navigation in non-structured environment.
Development and implementation of main concepts for processing sensor data and creating environmental local map to improve robot navigation in non-structured environment with the help of artificial neural networks.
Development and implementation of effective and self-adaptive methods for robot navigation and pathway planning.
Research of experimental methods (with the use of imitation modeling and neural network resources).

Team:

Anatoliy Sachenko
Vasyl Koval
Oleh Adamiv
Yuriy Kurylyak
Maxym Lunochkin
Serhiy Maystrenko

Published results:

Koval V., Adamiv O. The Software Structure Development for Mobile Robot Control // Proceedings of the IEEE Second International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications IDAACS’2005. – 2005. – Sofia (Bulgaria). – pp. 120-124.
Oleh Adamiv. Models and Intelligent Means of Autonomous Mobile Robot Adaptive Control: Ph.D. Theses on speciality: 05.13.23 / Ternopil National Economic University. – Ternopil, 2007. – 166 p.

Development of Parallel Neural Networks Training Algorithms on Advanced High Performance Systems

Foreign partner: Parallel Computing Laboratory, Department of Electronics, Computer Science and Systems, University of Calabria, Italy

Principal investigator from Ukraine: Dr. Volodymyr Turchenko

Principal investigator from Italy: Prof. Lucio Grandinetti

Grant # INTAS YSF 03-55-2493

Duration: 2004 – 2006

Main tasks:

Develop a parallel algorithm of enhanced data integration method using C programming language and MPI parallelization technology.
Design and implement in C programming language and MPI parallelization technology two new methods of coarse-grain neural network parallelization which provides high efficiency of parallelization at the certain training parameters of neural networks and dynamic mapping method, which is more universal than static and shows better efficiency at different initial parameters of neural networks and provides parallelization. A series of on-line computational experiments of the above mentioned algorithms of the parallel machines SGI Origin 300, NEC TX-7 is performed and the computational grid consists of the cluster of double-processor Compaq personal computers under Linux operation system and Globus middleware package.
Develop and implement in C programming language using MPI and MPE libraries the fine-grain parallel training algorithm of multilayer perceptron with parallelization of the outputs of hidden layer neurons at the initial stage of information processing inside neural network module.
Compare the advantages and disadvantages of middleware technologies, in particular Globus, in a case of coarse-grain parallelization algorithm of Integration Historical Data Neural Networks with dynamic mapping on the parallel computer Origin 300 without using middleware package and on the computational grid operated by Globus middleware package.

Published results:

Turchenko. Parallel Algorithm of Dynamic Mapping of Integrating Historical Data Neural Networks, Information Technologies and Systems, 2004, Vol. 7, No. 1, pp. 45-52, ISSN: 0135-5465, http://www.tanet.edu.te.ua/iics/vtu/B7.pdf.
Turchenko, V. Demchuk. Efficiency Analysis of Parallel Routine Using Processor Time Visualization, International Scientific Journal of Computing, 2005, Vol. 4, Issue 1, pp. 12-18, ISSN: 1727-6209, http://www.tanet.edu.te.ua/computing/Computing2005Vol4Issue1-12-18.pdf.
Turchenko. Computational Grid vs. Parallel Computer for Coarse-Grain Parallelization of Neural Networks Training, Lecture Notes in Computing Science LNCS 3762, Edited by Robert Meersman, Zahir Tari, Pilar Herrero, Berlin, Heidelberg, New York, Springer-Verlag, 2005, pp. 357-366, ISSN: 0302-9743, http://dx.doi.org/10.1007/11575863_55.
Turchenko, C. Triki, L. Grandinetti, A. Sachenko. Efficiency Estimation of Parallel Algorithm of Enhanced Historical Data Integration on Computational Grid, International Scientific Journal of Computing, 2005, Vol. 4, Issue 3, pp. 9-19, ISSN: 1727-6209, http://www.tanet.edu.te.ua/computing/Computing2005Vol4Issue3-9-19.pdf.
Turchenko. Fine-Grain Approach to Development of Parallel Training Algorithm of Multi-Layer Perceptron, Artificial Intelligence, 2006, Vol. 1, pp. 94-102, ISSN 1561-5359, http://www.tanet.edu.te.ua/iics/vtu/B1.pdf.

Development of a Web-based Measurement System with Distributed Intelligence

Foreign partner: Laboratory of Signal Processing and Information Measurement University of Sannio, Benevento, Italy

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Italy: Prof. Pasquale Daponte

Project was performed under the Ministry of Education and Science of Ukraine order #M/79-2004, state registration #0104U006975.

Duration: 2004 – 2006

Objectives: to create a distributed measurement system (based on Intranet and Internet technologies), that can provide high accuracy sensor data processing by the use of artificial neural networks. The system’s feature is remote units working in real time mode during long delays in data link layer, and costs decrease is achieved by shifting of some intelligent functions to a main server.

Main tasks:

Development of distributed measurement system architecture with either Internet- or Intranet-technologies.
Research and design of networked software structures. Development of software for distributed system using Web-technologies.
Testing and verification of the developed software for distributed measurement system.

Team:

Anatoliy Sachenko
Volodymyr Turchenko
Volodymyr Kochan
Roman Kochan
Iryna Turchenko
Volodymyr Hrusha
Olexandr Osolinskiy

Published results:

Hrusha, O. Osolinskiy, P. Daponte, D. Grimaldi, R. Kochan, A. Sachenko, I. Turchenko. Distributed Web-based Measurement System // IEEE Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications. 5-7 September 2005, Sofia, Bulgaria – pp. 355 -358.
Hrusha, O. Osolinskiy, R. Kochan, G. Sapojnyk Development of Web-based instrumentation, Proc. Of the International Conference “Modern Problems of Radio-Engineering, Telecommunications and Computer Science” TCSET’2006, February 28 – March 4, 2006, Lviv-Slavsko, Ukraine – pp. 199-201.
Hrusha, O. Osolinskiy, P. Daponte, D. Grimaldi, R. Kochan, A. Sachenko, I. Turchenko. Distributed Web-based Measurement System // IEEE Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications. 5-7 September 2005, Sofia, Bulgaria – pp. 355 -358.
Turchenko, V. Kochan, A. Sachenko, R. Kochan, A. Stepanenko, P.Daponte D. Grimaldi “Simulation Modeling of Neural-Based Method of Multi-Sensor Output Signal Recognition” in Proceedings of 2006 IEEE Instrumentation and Measurement Technology Conference IMTC/06. – April 24-27, 2006. – Sorrento (Italy). – pp. 1530-1535.

Design of Distributed Sensor Network for Ayers Island Security Using Value Analysis Technology

Foreign partner: Department of Computer Science, University of Maine, USA

Projects investigator from Ukraine: Prof. Anatoliy Sachenko

Projects investigator from USA: Prof. George Markowsky

Project had been performed within the frames of the First Steps to Market program of the US Civilian Research and Development Foundation.

Grant # CRDF FSTM UM2-5012-TE-03

Duration: 2003 – 2005

Objectives: investigating possibilities for developing distributed sensor network with defined features for providing security Ayers Island, Orono, ME, USA.

Main tasks:

Analyze component and perimeter security systems vendors, examine well-known perimeter security systems.
Propose algorithm for defining key functional indicators for perimeter security distributed systems components that can optimize preparing procedure for CAD, intended for design and optimization according to functional-price characteristics security system. This algorithm usage filled DB with functional-price characteristics for perimeter area security systems components that are unified and eligible for creating CAD.
Morphological matrix method was proposed for optimization according to functional-price characteristics of designed security systems and selecting variants of DSN that create Paret boundaries for all alternative variants according to two key functional characteristics.
CAD software module was developed, functions for all modules were described, and major requirements to perimeter area security systems CAD were established. Proposed CAD allows us to design projects perimeter area security systems, using perimeter area security systems components database.
Demonstrate CAD version that was used for developing perimeter area security systems for Ayers island in Orono, ME according to quality, reliability and price characteristics.

Team:

Anatoliy Sachenko
Volodymyr Turchenko
Volodymyr Kochan
Pavlo Bykovyy

Published results:

Bykovyy P. Choosing of Technical & Economic Indices for Knowledge Base of Perimeter Security Systems // Proceedings of the 2004 IEEE International Conference on Intelligent Systems 3. – 2004. Bulgaria. – pp. 54-57.
Turchenko, V. Turchenko, V. Kochan, P. Bykovyy, A. Sachenko and G. Markowsky. “Database Design for CAD System Optimizing Distributed Sensor Networks for Perimeter Security.” Proceedings of the 8^th IASTED International Conference on Software Engineering and Applications SEA’2004 (2004): 59-64. (USA)
Kochan, V. Kochan, A. Sachenko, I. Maykiv, I. Turchenko and G. Markowsky. “Network Capable Application Processor based on FPGA.” Proceedings of the 22^nd IEEE Instrumentation and Measurement Technology Conference IMTC 2005 II (2005): 813-817. (Canada)
Bykovyy, I. Maykiv, I. Turchenko, O. Kochan, V. Yatskiv and G. Markowsky. “A Low-Cost Network Controller for Security Systems.” Proceedings of the 3^rd IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications IDAACS’05 (2005): 388-391. (Bulgaria)

Development of Intelligent Precision System for Thermal Objects Control

Foreign partner: Department of Automatics, the University of Mons, Belgium

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Belgium: Prof. Marcel Remy

The project had been performed under the NATO (Cooperative Science & Technology Sub-Program).

Grant NATO PST.CLG.977647

Duration: 2002 – 2004

Objectives: development of precision and self-adaptive temperature control system for temperature objects with multi-zone linked control.

Main tasks:

Analysis of precision thermal objects and their control systems;
Analysis of error control system components and ways for reducing their influence on general system error.
Development of constructive-technological and structural-algorithmic methods for improving accuracy of measuring channels and control channels for multi-zone thermal objects.
Development of result processing methods for defining thermal objects parameters.
Adaptation of random small perturbation method for thermal objects with multi-zone linked control.

Team:

Anatoliy Sachenko
Roman Pasichnyk
Volodymyr Kochan
Volodymyr Turchenko
Roman Kochan
Nadia Vasylkiv
Yuriy Pihovsky
Mykola Derlytsya

Published results:

Derlytsya M., Pigovsky Y., Pasichnyk R., Kochan V. Improved Control System of Multi-Zone Thermal Object // Scientific Journal of Khmelnytsky Podillya Technical University. – 2004. – No. 2. – Vol. 1. – pp. 30-33.
Kochan V., Vasylkiv N., Chyrka M. The Error Evaluation of Temperature Measurement in Diffusion Furnace // Proceedings of the VIII International Conference Temperature. – 2003. – Lviv (Ukraine). – pp. 33.
Sachenko A., Kochan V., Pasichnyk R. Development of the Simulation Model of Thermocouples // Proceedings of the IEEE Instrumentation and Measurement Technology Conference IMTC/2003. – 2003. – Vail, CO. – pp. 1673-1677.
Derlytsya M. Improvement of the PC Based System of Optimal Control of Multi-Zone Thermal Object // Master Thesis, Ternopil Academy of National Economy. – 2004.
Pigovsky Y. Simulation Model for Effectivity Control of the Chip Manufacturing Process // Master Thesis, Ternopil Academy of National Economy. – 2004.

Using Multisensor Fusion and Neural Networks Techniques for Robot Control

Foreign partner: Laboratory of Robotics Systems, Unevirsity of La Coruña, Spain

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Ukraine: Prof. Richard Duro

The project had been performed under the NATO (Cooperative Science & Technology Sub-Program).

Grant NATO PST.CLG.978744

Duration: 2002 – 2004

Objectives: development and implementation of main concepts of merging sensor data, using neural networks for controlling mobile robot. It is assumed that robot moves in unknown (dangerous for human) environment. Main purpose is the endpoint reached through obstructions.

Main tasks:

Development of new methods for merging sensor data, using neural networks.
Development of algorithms and software for merging sensor data subsystem.
Hardware implementation of merging methods for sensor data on mobile robot.
Verification and testing procedures of developed engines for merging sensor data on mobile robot.

Team:

Anatoliy Sachenko
Volodymyr Turchenko
Vasyl Koval
Oleh Adamiv

Published results:

Koval V. The Fusion of Structured Light and Video Image for Mobile Robot Control // Scientific and Technical Journal Artificial Intelligence. – 2004. – Donetsk (Ukraine). – No1.
Koval V. The Method of Obstacle Detection Using Fusion Technique of Heterogeneous Sensors // ASU and Automatic Devices. – 2004. – Kharkiv (Ukraine). – pp. 128-135.
Koval V., Turchenko V., Kochan V., Sachenko A., Markowsky G. Smart License Plate Recognition System Based on Image Processing Using Neural Network // Computing. – 2003. – Vol. 2. – Issue 2. – pp. 40-46.
Adamiv O., Koval V., Turchenko I. Predetermined Movement of Mobile Robot Using Neural Networks // International Scientific Journal Computing. – 2003. – Ternopil (Ukraine). – Vol. 2. – Issue 2. – pp. 64-68.
Koval V., Turchenko V., Sachenko A., Becerra J., Duro R., Golovko V. Infrared Sensor Data Correction for Local Area Map Construction by a Mobile Robot // The Lecture Notes in Artificial Intelligence, LNAI2718. – 2003. – pp. 306-315.
Koval V. The Method of Local Area Map Construction for Mobile Robot // Scientific Journal of Ternopil State Technical University I.Pulyuj. – 2002. – Ternopil (Ukraine). – Vol. 8. – No2. – pp. 80-88.
Koval, “Adversary merging sensor data algorithm on ultisensory systems”, // Sensors and systems, #7 (38) Sep. 2002. Pp.39-41.
Vasyl Koval. Methods and Algorithms of Map Development of Mobile Robot Environment Using Sensor Data Fusion: Ph.D. Theses on speciality 05.13.23 / Ternopil Academy of National Economy; NAS of Ukraine; State Research Institute of Information Infrastructure. – Ternopil, 2004. – 208 p.

Development of an Intelligent Sensing Instrumentation Structure

Foreign partners: Electronic Laboratory, Aristotle University, Thessaloniki, Greece, Parallel Computations Laboratory, University of Calabria, Italy, Department of Electronics at Brest Polytechnic Institute, Belarus.

Principal investigator from Ukraine: Prof. Anatoliy Sachenko

Principal investigator from Greece: Prof. Theodore Laopoulos

Principal investigator from Italy: Prof. Lucio Grandinetti

Principal investigator from Belarus: Prof. Volodymyr Golovko

The project had been performed under the “INTAS Open Call” program, grant # INTAS OPEN 97-0606.

Duration: 1999 – 2001

Objective: development of information measurement system for increasing measurement accuracy using automated correction of instrumental compound measurement error.

Research tasks:

Target area analysis and requirements definition for intelligent sensor measurement system;
Development of distributed structure for intelligent sensor measurement system;
Development of methods for evaluating the results of processing with the target objective to increase the system operational characteristics;
Development and testing of the prototype intelligent sensor measurement system.

Team:

Anatoliy Sachenko
Volodymyr Kochan
Volodymyr Turchenko
Roman Kochan

Published results:

Sachenko A., Kochan V., Turchenko V., Tymchyshyn V., Vasylkiv N. Intelligent Nodes for Distributed Sensor Network // Proceedings of the 16^th IEEE Instrumentation and Measurement Technology Conference IMTC/99. – 1999. – Venice (Italy). – Vol. 3. – pp. 1479-1484.
Golovko V., Grandinetti L., Kochan V., Laopoulos T., Sachenko A., Turchenko V. Tymchyshyn V. Approach of an Intelligent sensing Instrumentation Structure Development // Proceedings of the IEEE Intenational Workshop on Intelligent Signal Processing WISP’99? Budapest, Hungary, 4-6 September, 1999. – pp. 336-341.
Sachenko A., Kochan V., Turchenko V., Laopoulos T., Golovko V., Grandinetti L. Features of Intelligent Distributed Sensor Network Higher Level Development // Proceedings of the 17^th IEEE Instrumentation and Measurement Technology Conference IMTC/2000. – 2000. – Baltimore (USA). – pp. 335-340.
Sachenko A., Kochan V., Turchenko V., Golovko V., Savitsky Y., Dunets A., Laopoulos T. Sensor Errors Prediction Using Neural Networks // Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks IJCNN’2000. – 2000. – Como (Italy). – Vol. IV. – pp. 441-446.
Sachenko A., Kochan V., Kochan R., Turchenko V., Tsahouridis K., Laopoulos Th. Error Compensation in an Intelligent Sensing Instrumentation System, 18^th IEEE Instrumentation and Measurement Technology Conference IMTC/2001. – 2001. – Budapest (Hungary). – pp. 869-874.
Turchenko V., Kochan V., Sachenko A., Laopoulos Th. The New Method of Historical Data Integration Using Neural Networks // Proceedings of the International Workshop on Intelligent Data Acquisition and Advanced Computing Systems IDAACS’2001. – 2001. – Foros (Ukraine). – pp. 21-24.
Turchenko V., Kochan V., Sachenko A. Estimation of Computational Complexity of Sensor Accuracy Improvement Algorithm Based on Neural Networks // Lecture Notes in Computing Science, No 2130, Ed. By G.Gooss, J.Hartmanis and J. van Leeuwen, Springer-Verlag, Berlin, Heidelberg, New York. – 2001. – pp. 743-748.
Volodymyr Turchenko. Neural Network Methods and Means of Efficiency Improvement of Distributive Networks of Sensor Data Acquisition and Processing: Ph.D. Theses on speciality 05.13.13 / Lviv National Polytechnical University. – Lviv, 2001. – 188 p.
Volodymyr Tymchychyn. Efficiency Increasing of Specialized Computer System Design on the Base of Typical Microprocessor Platforms: Ph.D. Theses on speciality 05.13.13 / Lviv National Polytechnical University. – Lviv, 1999. – 200 p.
Patent of Ukraine 25609A, MKI G06F 15/00. Two-Wired Local Area Network, Signal Repeater and Invertor for Using in it / V. Kochan, V. Tymchyshyn (Ukraine); Applied 30.10.97 # 97105295; Issued 30.10.98.
Patent of Ukraine 25498A, MKI G06F 11/00. Method of Communication Channel Bandwith Increasing on the Base of Serial Interface and Device for it Realisation / V. Kochan, V. Tymchyshyn (Ukraine); Applied 27.01.98 # 98010432; Issued 30.10.98.

You missed

The value of knowledge increases when it is shared

Meeting of WUNU team with representatives of regional companies as part of pilot training within WORK4CE project

Visit to the Academy of the Greek Army in Athens, Greece

WUNU scientists presented the newest system of environmental data analysis at the MIT conference