Advances in Electrical and Computer Engineering
ISSN 1582-7445 - No 1 / 2005

EDITOR

  Adrian GRAUR   - Faculty of Electrical Engineering
 
 

ASSOCIATE  EDITORS

  Ştefan Gh. PENTIUC  - Faculty of Electrical Engineering
  Leon MANDICI   -  Faculty of Electrical Engineering
  Dorel CERNOMAZU  - Faculty of Electrical Engineering
 
 

EDITORIAL  BOARD

 Tudor AMBROS   - Technical University of Chişinău - MOLDAVIA
 Dumitru N. ALEXANDRU   - Technical University Iasi
 Oleg BREKHOV   - Moscow Aviation Institute - RUSSIA
 Mihai CERNAT - "Transilvania" University Brasov
 Cristophe CHAILLOU   - Universite des Sciences et Technologies de Lille - FRANCE
 Mihai CRETU   - Technical University Iaşi
 Luc DE BACKER   - Katholieke Hogeschool Sint-Lieven Gent - BELGIUM
 Timothy HALL   - University of Limerick - IRELAND
 Ileana HAMBURG   -Institute for Technology Gelsenkirchen - GERMANY
 Stefan HOLBAN   - "Politehnica" University Timisoara
 Maria G. IOANNIDES   - National Technical University of Athens - GREECE
 Ioan JURCA   - "Politehnica" University Timisoara
Vladimir KATIC  - University of Novi Sad - SERBIA AND MONTENEGRO
 Eleftherios A. KAYAFAS   - National Technical University of Athens - GREECE
 Mykola KIRICHENKO   - Institute of Cybernetics Kiev - UKRAINE
 Valeriu KRAPIVIN   - Moscow Institute of Radioengineering and ELectronics - RUSSIA
 Ioan LETIA   - Technical University Cluj-Napoca
 Volodymyr MESYURA   - The State Technical University of Vinnitsa - UKRAINE
 John MILNER   - City University London - GREAT BRITAIN
 Radu MUNTEANU  - Technical University Cluj-Napoca
 Jose OLIVEIRA e SA  - Instituto Politecnico do Porto - PORTUGAL
 Costică NITU   - "Politehnica" University Bucureşti
 Adrian PETRESCU   - "Politehnica" University Bucureşti
 Mircea PETRESCU  - "Politehnica" University Bucureşti
 Nicola PITRONE  - Universita degli Studi Catania - ITALIA
 Radu Emil PRECUP - "Politehnica" University Timisoara
Jean Paul SIX   - Université des Sciences et Technologies de Lille - FRANCE
 Fedir SOPRONIUC   - University of Cernăuţi - UKRAINE
Mircea STRATULAT  - "Politehnica" University Timisoara
 Gavril TODEREAN - Technical University Cluj-Napoca
 Petru TODOS   - Technical University of Chisinau - MOLDAVIA
 Bernard TOURSEL   - Université des Sciences et Technologies de Lille - FRANCE
 
 

  Back to top

FEE INFO

16 – 21 May 2005
 

The Faculty of Electrical Engineering organized the 12th Annual International Students’ Computers Contest “Hard & Soft”. The theme of the contest was a Lift Control System based on remote instrumentation, analysis / control and feedback.

The contestants were required to design and build an electronic system linked to a PC where data manipulation had to lead to system feedback and user display. Some elements of the system could be remote. Three implementations were demanded: hardware, software, hardware-software.

Participants: teams from Beijing (China), Brasov, Bucuresti, Cernautzi (Ukraine), Chisinau (Moldavia), Craiova, Iasi, Suceava, Timisoara.

President of the jury: Professor Timothy HALL from University of Limerick, the Republic of Ireland. His team included Cathal McHUGO, Kevin JOHNSON from the same university, and Claudine LECOCQ from Ecole Politechnique Universitaire de Lille, France.

The prizes were awarded to:
• the “Politehnica” University of Timisoara  -  1st prize
• the University of Technology in Beijing   -  2nd  prize
• the Technical University of Moldavia in Chisinau -  3rd  prize

An International Robotics Seminar was held the same week. It brought together scientists, engineers, manufacturers, graduate students and postdoctoral scholars in order to provide and exchange information on recent developments and future trends in Robotics Theory, Artificial Intelligence Focused on Robotics, Electrical Drives Control, Pattern Recognition.
 
 
 
 
 
 

Professor Adrian GRAUR

Abstract. The robot Manura has been designed for cleaning barns and equipped with short-range sensors. The localization of this operational robot is accomplished by a neural network. We focus on the learning techniques applied in this project. The training is improved by adding some self-generated exemplars
Keywords: robotics, localization,  hybrid neural approach regression, classification, prediction, Manura robot
 
 

Abstract. Efficient river inflow forecast is essential for various purposes such as for flood control, distribution of water for irrigation purposes, etc. It is also a highly useful tool for the efficient operation and maintenance of hydro power plants as the economics of the hydro electricity depend largely on the reservoir height as well as the inflow rate into the dam. Through effective inflow prediction mechanism, proper maintenance and operation schedule of the generating machines can be planned, thereby reducing the forced outages and increasing the generation output.
In this paper, a neural network based approach is presented to forecast the daily river inflow for a Wangchu River of Bhutan since on this river basin lie the two biggest hydro plants, Chukha Hydropower Corporation (CHPC) and Tala Hydro Project Authority (THPA), with generating capacities of 336 MW and 1020 MW, respectively. For a small Himalayan kingdom where half the total national revenue comes from the sale of hydro power alone, with almost around 80 percent exported to India, it is essential to have a proper mechanism to accurately forecast the inflow as it would help implement the optimal utilization of water resources as well as plan efficient load scheduling. The latter is particularly important for power export as prior electric generation anticipation based on river inflow is critical in transacting the power sale in advance, thereby significantly improving the revenue earned.
In this paper, two types of neural networks are designed and their performances compared in predicting the next day inflow. Both the networks are extensively tested using the inflow and other weather related data from the year 1999 to 2003.
Keywords: reservoir inflow, neural network, hydro power.
 
 

Abstract. In this paper work are proposing some original solutions particular to the data transfer mean and protocols corresponding. The bus imagined has some specific features that conferring originally behaviors. The protocol involved is thinking in this application to working in logical mean with the hardware-designed structure. In this sense is performed entire hardware in relation with the signal type generated by the various kinds of sensors. Was be given attention to the adapting electronic signal structures like as an interface between sensors involved, from one the side, and the physical bus system (with all its conflict problems which can appear). The protocol will be considering these features and will be thinking in so manner to increasing the capabilities of the system.
The authors contributions are located both electronic hardware part (through the physical line transmitting and the adapter electronic modules) and software part (through the implemented protocol routine system). The electronic hardware part is performing with an 8-bit PCB80C553 Philips microcontroller board. The software part consists in three different routines type: a program to communication with the PC system (involved in the loading procedure), an interfacing program between the microcontroller system and the user, and a protocol communication with the designed bus program. For the PC communication is used the C language and for the user interface and the protocol communication on the bus is used the microcontroller's assembler language.
For PC/PCB communication can be used a wireless channel data transfer and so is increased the handling system capabilities.
Keywords: I2C bus, communication, self-arbitration, bus adapting, data signal. clock signal.
 
 

Abstract. The paper is dealing with some electromagnetic compatibility (EMC) problems of converters and inverters, which are utilized for feeding of electric machines and its analyzing by PSPICE program. The main attention is focused on implications of parasitic capacitance and capacitive coupling existence.
Keywords: EMC, converter, PSPICE simulation, parasitic capacitances, capacitive coupling.
 
 

Abstract. Fault detection is a crucial and challenging task in the automatic control of complex system, e.g. in flexible manufacturing systems, and robotic systems as representative classes of discrete event systems.
In this paper, we consider the discrete event systems (DES) modeled via a state – space representation. The control objective is the avoidance of a given set of states. We achieve this goal by verifying that certain predicates, specified in terms of states, are always false. We model the DES involved in physical systems with a class of controlled state machines, respectively the formalism of the controlled Petri nets. A model of a railway system illustrates the theoretical approach. Possible extensions of the work are also discussed.
Keywords: Controlled Petri nets, discrete event systems, railway systems.
 
 

Abstract. Synthesis and implementation of FSM machines will be more easy with complex circuits, like MSI and LSI circuits. As the realized functions by these circuits are more complex, the combinational logic network around the  finite automata ‘kernel’ is too simple, [1],[2]. Usually, heuristics methods can be used  for synthesis of the additional combinational logic network,  [3],[4]. It will be deducted the transition matrix.
Keywords: Finite State Machine, Transition Matrix, Counter, Logic Primitive, Medium Scale Integration,    Lower Scale Integration.
 
 

Abstract. The aim of this paper is to present a solution based on a Multiagent System for a flexible manufacturing cell. The system is composed of three agents controlling a conveyor band, a stoking area, and a robot arm that takes pieces from the conveyor and put them in the stocking area. The manufacturing cell is supervised by video cameras connected to a computer.
Keywords: intelligent agents, multiagent system, flexible manufacturing cell, control algorithms
 
 

Abstract. A fault diagnosis system based on neuro-fuzzy model is proposed. The wall hang boiler is modeled as a second order fuzzy model in a dynamic configuration. The identification model methos is used as basic tool for estimation of the parameters of neuro-fuzzy system. The signals used in fault diagnosis need various preprocessing phases due to many component tolerances, noises and environment disturbances. The experimental results along with theoretical support are presented.
Keywords: neuro-fuzzy model, fault identification, wall hang boilers, model based fault detection.
 
 

Abstract. In computer systems, there is an urgent need for accurate authentication techniques to prevent unauthorized access. Authentication is the process of confirming the correctness of the claimed identity Many computers that store critical information are vulnerable to unauthorized access because of weak authentication. In some cases, the safety of the public can be at risk, such as in the case of a multi-billion dollar passenger-screening system defenceless against terrorists with forged security badges. Traditional authentication techniques such as the ubiquitous username / password method are inadequate for personal identity since they can only provide proof of possession and/or proof of knowledge. Only biometrics, the authentication of individuals using biological identifiers, can offer true proof of identity. Current research suggests that multimodal biometric systems, those that use more than one biological identifier, can improve the accuracy of biometric systems. This improvement in accuracy depends on critical factors in design, implementation and security. This paper explains each of these critical factors so the increase in accuracy observed in current research can be achieved in real-world applications.
Keywords: biometric, identification, verification, recognition, image acquisition, iris, fingerprint, face detection
 
 

Abstract. The paper presents some aspects regarding the representation and processing of the data and knowledge within an information system for vegetal genetics, dedicated to the educational and research institutions in this field. The data and knowledge that are to be stored and processed in a documentation and assisted research system for vegetal genetics, refers resources of the vegetal genetics representing all the vegetal life forms: wild plants, varieties and local populations, lines, hybrids, weeds, improved forms, etc., all of these being subject to the genetic erosion phenomena, pathogenic elements aggression (phytopatology) and environment factors. The plant description (the phenotype), the genetic hereditary heritage (the genotype) as well as the results of the experiments and researches performed in the vegetal field are fed into the system’s database within a predefined evolutionary self developing structure. Among the specific research problems in this field one can mention: population recognition and classification, plant disease diagnostication. The basis concept was inspired by the activity of Suceava Genebank - Romania.

Keywords: assisted research,, vegetal genetics, phenotype, genotype, data and knowledge base, evolutionary  structure, artificial intelligence, pattern  recognition.
 
 

Abstract. NeuroShell is one of the many neural networks simulators available. In order to build a correct model to solve a specific problem, some steps must be taken into considerations. The paper presents an overview of the NeuroShell 2 program and some practical results obtained for the chosen problems.
Keywords: neural networks, NeuroShell, simulation, prediction.
 
 

Abstract. This article presents efficient algorithms for the planning of some vehicles that have the mission to detect and extinct a fire. The machines work in a varied land, with hills and valleys, with holes, cliffs, mined land and with other obstacles they may run into their way (this obstacles are from the beginning in the studied land or they can appear instantly), and whom the machine must avoid. The machines can depart all at the same moment or they can leave one at a moment. The machines can meet each other on their way but they must avoid the collisions by communicating with the server. The machines are equipped with complex sensorial systems like video cameras, transducers, proximity sensors (thermics, smoke, gases, etc.), and neuro-fuzzy logic. There are used rapid sequential algorithms of the Strassen and Winograd for the matrices multiplication (using parallel calculus) and the Dijkstra’s algorithm to compute the optimum way between two points using parallel programming.
Keywords: fire, extinction, intelligent machines, computer vision, parallel programming.
 
 

Abstract. Algorithms developed by the professor John Daugman from Cambridge University for recognizing persons by their iris patterns have now been tested in six field and laboratory trials, producing no false matches in several million comparison tests. The recognition principle is the failure of a test of statistical independence on iris phase structure encoded by multi-scale quadrature wavelets. The combinatorial complexity of this phase information across different persons spans about 249 degrees of freedom and generates a discrimination entropy of about 3.2 bits/mm2 over the iris, enabling real-time decisions about personal identity with extremely high confidence. The high confidence levels are important because they allow very large databases to be searched exhaustively (one-to-many .identification mode.) without making false matches, despite so many chances. Biometrics that lack this property can only survive one-to-one (.verification.) or few comparisons. This paper presents an efficient software product that recognizes a person by its iris pattern by searching in a large iris database.

Keywords: software product, iris, image, recognition.
 
 

Valeriu LUPU
"Stefan cel Mare" University of Suceava
str.Universitatii nr.9, RO-720225 Suceava
valeriul@seap.usv.ro
 

Abstract. This article presents neuro-fuzzy logic and complex sensorial systems used for the planning of some intelligent vehicles that have the mission to transport persons with handicaps inside of a building (or in a room), transport of the person with disabilities inside their home without companion ,baggage transport in airports (without driver), container transport in harbours, marrow places (without crew), transport of dangerous waste products, transport of persons in winter conditions, cleaning of waters, mountains (cleaning of the waters from oil waste), detection and extinction of fire. The machines work in a varied land, with hills and valleys, with holes, cliffs, mined land and with other obstacles they may run into their way (this obstacles are from the beginning in the studied land or they can appear instantly), and whom the machine must avoid.  The cars can leave all at one moment or individually. They can meet on their way, but they must avoid the collisions by communicating between them through the server. The machines are equipped with complex sensorial systems like video cameras, transducers, proximity sensors, neuro-fuzzy logic. Sequential rapid algorithms – like Strassen’s and Winograd’s – are used for the matrices multiplication (by using parallel calculus) and Dijkstra’s algorithm for the calculus of the optimum road between two points by using parallel programming.
Keywords: transportation, neuro-fuzzy logic, parallel programming, intelligent machines
 
 

Abstract. The control system for a robot arm is an ongoing image processing project for which many enhancements have been done during the last years. The paper introduces a new control system for a robot arm, developed in the Electrical Engineering Department. The control is represented by the robot’s movements of grabbing and moving objects located on an operational table to user selected destination points. The enhancements involve robot visual-based automatic calibration with respect to the working field axes and scale, fast and robust object detection. A self-calibration algorithm, based on video motion and color detection, is used in order to provide a good conversion of the objects’ coordinates from the image space into the working area coordinates system. The algorithm proves to be general enough to be employed in other similar image processing based scenarios where self-calibration might be useful
Keywords: robot control, image processing, automatic self calibration, video motion detection, edge detection
 
 

Abstract.  This paper introduces the concept of an autonomous machine designed to search a candle inside a maze. The shape and dimensions of the maze, as well as the height of the candle are given. The robot has three proximity sensors: left, right and front. Using these sensors, the robot adjusts its position relative to the walls. The electric compass and the sensors for reading the white tape delimiting the room exits are used to determine the starting room. Once this phase is completed, the robot follows a pre-established pattern and searches the candle in every room. Once the candle is found the robot switches on an extinguishing fan.
Keywords: robot, maze, candle, proximity, sensors.
 
 

Abstract. An adaptive approach for skin color detection is presented based on image filtering with dynamically computed thresholds. Two dimensional hue/saturation histograms are analysed and partitioned into rectangular blocks. For each block, simple parameters are computed such as: 1st order moments, standard deviations and simple ratios. A multilayered perceptron is trained using a set of images containing both skin and non-skin blocks. For each frame in a video sequence, each block is assigned a probability of being skin and the parameters of the winning block, if any, are used for image thresholding.
Keywords: skin color, two dimensional histogram, thresholding, multilayered perceptron.