Acta Technica Napocensis – Electronics and Telecommunications – Abstracts

Number 1,1997

A COMPRESSION TECHNIQUE FOR GRAY SCALE FINGERPRINT IMAGES
Serban LUNGU, Marius TICO
Abstract: The paper deals with a JPEG implementation for fingerprint image 
compression. A minimum value of signal to noise ratio for fingerprint images was 
established Using a special quantization table and LZW encoder, the implemented 
compression scheme can achieve about 10:1 compression ratio for fingerprint images.
Keywords - fingerprint, image compression, JPEG, discrete cosine transform, 
quantization, bit allocation, peak signal-to-noise ratio.
 
A FUZZY METHOD FOR MANUAL, FAST ANALYSIS OF STATIC HAZARD IN DIGITAL CIRCUITS
C. MIRON, L. MICLEA, M. SABAU, G. OLTEAN, M. GORDAN
Abstract: One type of faults in digital circuits that is undetectable by the means 
of binary logic analysis is the static hazard due to the non-zero time of the 
circuit signal transitions. This type of static hazard is a design fault, and it is 
dependent of the specific circuit realization that implements the desired binary 
logic function.
In this paper, we propose a method for fast detection of static hazard, based on 
the fuzzy modeling of the digital circuit, represented at the gate level. Each gate 
is represented by a fuzzy algebraic relation, which gives its output as a function 
of its inputs. Based on the gate level structure of the circuit, we combine the 
fuzzy algebraic relations for the gates and obtain a fuzzy function which gives the 
output of the circuit as a function of its primary inputs; we call this function 
the fuzzy function of the circuit, and it characterize completely the circuit 
behavior regarding the presence of the static hazard on its output.
Keywords: fuzzy sets, fuzzy algebra, digital circuits, static hazard, logic gate, 
gate modeling.
 
INFLUENCE OF CIRCUITE PARAMETERS AND PERFORMANCE MODELING OF SIGMA-DELTA CONVERTERS
A. RUSU,  S. LUNGU 
Abstract: In this article we have used the results of the theoretical analysis and 
the facilities of the MATLAB program for modeling, analyzing and simulating a wide 
range of Sigma-Delta analog-to-digital converters. The results of simulation using 
MATLAB Toolboxes for Signal Processing, show the follows: the higher-order sigma-
delta modulator provides more quantization noise suppression over the low frequency 
signal band and more amplification of the noise outside the signal band. Sigma - 
Delta A-to-D conversion allows the use of relatively imprecise analog circuits to 
perform high resolution conversion, using only a 1-bit A-to-D converter. Stability 
is the main problem for standard high-order sigma-delta modulators.
 
TESTING SYSTEM FOR ANALOGOUS BOARDS
Dan PITICA, Călin GHERGHE, Mihaela RADU, Mircea DABACAN
Abstract: The paper describes an 'In Circuit' testing method for boards equipped 
with analogue circuits, inspired from the technique used for digital systems. This 
method allows functional tots for a large range: of circuits, keeping the board 
plugged. There are presented not only the principle of the method and some 
theoretical considerations, but also aspects and performances regarding the 
implemented supporting hardware. Moreover, a final application example is provided 
in order to emphasize the advantages offered by this testing system. Key-words: 
functional testing, signature analysis, correlation
 
A SIMULATION METHOD FOR INDUCTION MOTORS
S. LUNGU , D. COSMA 
Abstract: The paper presents a variable step method, which can be used for 
induction motors simulation. The model used for the induction motor is the two 
phase model ; the axes are related to the stator. The nonlinearities due to core 
saturation and to nonsinusoidal flux distribution are not taken into consideration. 
The integration method has a variable step and provides a short computation time, 
which is comparable with MATLAB integration routines. Experimental results are 
compared to the simulated results at the end of the paper.
 
METHODS FOR SYNTHESIS OF DISCRETE LFNEAR TIME VARIANT SYSTEMS
Daniela TARNICERU , Valeriu MUNTEANU
Abstract : This paper develops a framework for the synthesis of linear time variant 
(LTV) discrete systems. In the proposed method, one determines the structure of a 
discrete controller, which Is cascaded with the known time varying process, so that 
the ensemble behaves as a linear time invariant (LTT) system satisfying certain 
design requirements. The obtained controller is very simple, consisting of a 
parallel bank of ITT filters of the fiist older followed by amplifiers with time 
variable gain. Simulation results are included to illustrate the theoretical 
discussions.
 
CEPSTRAL ANALYSIS TECHNIQUE FOR SPEECH SIGNAL
Eugen LUPU, Gavril TODEREAN
Abstract: Since their introduction, homomorphic signal processing techniques have 
been of great interest in speech/speaker recognition. There are two methods to 
compute the cepstral coefficients: Fourier Transform - derived cepstral 
coefficients and IPC (Linear Prediction Coding)-derived cepstral coefficients. 
Homomorphic systems were considered useful for speech processing because they 
offered a method for separating the excitation signal from the vocal tract shape so 
it can be obtained the pitch value and the cepstral smoothing of the spectrum
 
DESIGN OF A POWER FACTOR CORRECTION PREREGULATOR WITH ZERO CURRENT SWITCHING
Dorin PETREUS, Serban LUNGU
Abstract: This paper presents a design of a power factor correction preregulator with zero 
current switching. Design equations relating the boost inductor current, and 
switching times are developed and verified with measurements. Design guidelines for 
low-frequency feedback network are also presented.
 
NORTON AMPLIFIER LM 3900 - A BUILDING BLOCK FOR A TRIANGLE-SQUARE WAVE VCO
Andi CARCOTA
Abstract: Norton Amplifiers are less used than operational amplifiers, although 
their particularities may lead to simplified, efficient designs. A triangle-square 
VCO, designed around a quadruple amplifier, with a possibility to adjust mark/space 
ratio by altering only two resistors' ratio, is presented here.
 
A MODIFIED WIEN OSCILLATOR
Andi CARCOTA
Abstract: High quality, variable frequency RC oscillators are designed with dual 
gang variable capacitors as tuning elements. All successful designs have a major 
drawback : capacitors' common plate is not grounded, leading to unwanted coupling. 
This paper presents a new topology, that avoids this problem.
 
---------------------------------------------------------------------------------
 
Number 2, 1997
 
SEGMENTATION OF FINGERPRINT IMAGES USING THE RIDGE DIRECTION
Marius TICO , Serban LUNGU
Abstract: Fingerprint image segmentation which consist in noise elimination and 
ridge restoration algorithms should be incorporated in the preprocessing stage of 
any fingerprint identification system. A robust segmentation method is very 
important in order to develop further efficient procedure of fingerprint encoding, 
storage and recognition. This paper deals with a new unsupervised fingerprint image 
segmentation method. Central to proposed segmentation method is the computation of 
the ridge directions. In order to avoid the occurrence of false details in the 
fingerprint image a ridge gray levels regularization procedure based on directional 
information is used. The last stage of the segmentation method consist in the ridge 
pixels extraction by means of a local directional threshold. The proposed method 
was compared with other segmentation algorithms and the results show a significant 
improvement in identification performance.
 
SPICE MODEL FOR ZERO CURRENT SWITCHING QUASI-RESONANT BUCK CONVERTER
Serban LUNGU, Ovidiu POP, Silviu PLESA
Abstract : This paper is intended to explore in significant details of the quasi-
resonant converter topology. Voltage and current waveforms and transferred charge 
are analyzed and simulated as function of time and input/output conditions. 
Specific and generalized design equations are given, which arc also applicable to 
other topologies.
Keywords: power supply, quasi-resonant converter, zero current switching, 
conversion frequency, SPICE simulating
 
DIGITAL IMPLEMENTATION OF THE NYQUIST FILTERING CHARACTERISTIC IN HIGH SPEED DATA 
COMMUNICATIONS
POLGAR Zsolt, TODEREAN Gavril
Abstract: The paper studies the digital implementation of the Raised Cosine (RC) 
and Root-Raised Cosine (RRC) Nyquist filtering characteristics, employing FIR 
filters with imposed linear phase characteristics. By computer simulations, the 
authors establish the minimum order of the employed FIR filter and the minimum 
sampling frequency required, in order to secure an acceptable approximation error 
of the RC and RRC characteristics, implemented by FIR filters, for different roll-
off factors.
 
A NEW SPICE MODEL OF THE BAND - GAP VOLTAGE REFERENCES
Serban LUNGU, Silviu PLESA, Ovidiu POP
Abstract: Because the temperature DC analysis of the band - gap voltage references 
achieved by SPICE simulators provides errors, the components library file of the 
most known simulator (MicroSim 7.1. - Design Lab), doesn't contain any band-gap 
voltage reference. There are cases when we wish to know the variation law of the 
reference output voltage with temperature. This paper presents a detailed analysis 
of the effective temperature coefficient of band - gap voltage reference. A new 
SPICE subcircuit is given for the band - gap voltage reference analysis.
 
ANALYSING CURRENT FEEDBACK OP-AMP BASED VOLTAGE AMPLIFIER AND INTEGRATOR USING A 
TWO-POLES MODEL 
Marius NEAG , Lelia FESTILA
Abstract: Most theoretical analysis of CFB-OA were based on simple first order 
models. Recently, Mahattanakul and Toumazou emphasised the importance of 
considering a more accurate model, especially the second pole de- j termining the 
stability and maximum operating frequency conditions. To illustrate this, they 
analysed the j noninverting voltage amplifier and the inverting voltage integrator 
with CFB-OA as basic amplifier. Mainly the classical feedback analysis was used, 
with its two usual approximations - both the basic amplifier and the i feedback 
network were considered unilateral - considered in order to simplify the algebra; 
unfortunately, this led to unnecessary approximations and even inaccuracies.
The same CFB-OA applications are analysed in this paper using an extension of the 
classical feedback theory developed by the authors; the method is straightforward 
and provides an estimation of its errors, when comparing with an accurate analysis, 
at Kirchhoff 's laws level. An ordinary CFB-OA and a high-transimpedance CFB-OA 
design, with current mirrors with gain, have been considered; both one-pole and 
two-pole models have been used, providing for direct comparison. Besides the 
advantage of a more intuitive and simple approach, there are some differences 
between the results obtained here and the previously reported ones: first, the idea 
proposed by Vanisri and Toumazou, to increase the CFB-OA based noninverting voltage 
amplifier bandwidth by using the high-transimpedance CFB-OA, with current mirrors 
with gain, is reconsidered; second, the effects of the output resistances of input 
and output buffers within the CFB-OA structure, over the inverting voltage 
integrator parameters, are revised. Also, it is found that the high- transimpedance 
CFB-OA considerably extends the operational frequency range of the voltage 
integrator. The theoretical analysis is validated by simulations.
 
AN IMPEDANCE CONVERSION SINE OSCILLATOR
GARNIER Eliane, MARCHEGAY Philippe, CIUGUDEAN Mircea
Abstract: The work develops the theory of an impedance converter quadrature sine 
oscillator, whose scheme was previously carried out [3, 4, 5]. One establishes the 
oscillation conditions and an analysis is performed on the output voltage 
quadrature precision. The resonant circuits and the very large quality factor are 
studied. Two realized schemes and the experiment's results are presented. The 
oscillator study shows its remarkable specifications as stability and selectivity.
 
DEVICE MODELLING WITH EVOLUTION STRATEGIES
Mihaela CIRLUGEA
Abstract: This paper describes an approximation method for experimentally obtained 
functions using Evolution Strategies. The results can be used in circuit simulation 
and modeling. The program in this paper is a development of Escapade 1.2 |1|. It 
has to find an appropriate codification as a basis for the new goat-function 
initially known as couples of (x-y) values. The obtained function is then refined 
with the mathematical method of the gradient and, finally, both functions, the 
experimental and the approximated, are plotted. The Evolution Strategies use the 
principles of organic evolution processes as rules for optimum seeking procedures. 
For the studied functions, a polynomial codification was chosen based on an 
alphabet of mathematical operations.
 
-----------------------------------------------------------------------------------
Number 1, 1998
 
MODELING AND QUEUING ANALYSIS OF VARIABLE- BIT-RATE MPEG VIDEO STREAMS
Anastasios DOULAMIS, Nikolaos DOULAMIS , Stefanos KOLLIAS
Abstract: In this paper modeling and queuing analysis of variable-bit-rate (VBR) 
MPEG coded video sources transmitted over B-ISDN networks are studied!. A low pass 
version of the aggregate MPEG sequence is examined so that the network resources 
are analytically predicted. Based on this simplified signal, a discrete-state 
continuous-time Markov chain is proposed to approximate the MPEG video traffic. 
Then, queuing analysis is performed for obtaining analytical solution of cell 
losses and average buffer size. The overall analysis is relied on a fluid flow 
approximation of MPEG video traffic, in which the queue can be viewed as a 
reservoir of water that is fed from water supplies of time varying rates and that 
empties through a fixed rate sink. Experimental results, indicating the good 
performance of the proposed model, are finally presented using MPEG coded video 
streams. Keywords: VBR MPEG video sources, Modeling, Markov Models and Queuing 
Analysis.
 
DEFINING THE AUDIO-CONFERENCE TELEPHONE SERVICE IN A LOW CAPACITY ELECTRONIC 
TELEPHONE EXCHANGE BASED ON SPATIAL SWITCHING
Petrut DUMA, Dumitru Nicolae ALEXANDRU, Luminița SCRIPCARIU
Abstract -The work presents the software implementation of the audio-conference 
telephone service for three subscribers, in a low capacity electronic telephone 
exchange with spatial switching network. There are also described the stages of 
realizing this service, the graph of states structure, as well as the structure of 
the multi-proce as system with uniform time processing division which implements 
this service.
 
STUDY AND DESIGN OF THE PFC SINGLE-STAGE SWITCHING REGULATORS
Serban LUNGU, Ovidiu POP, Silviu PLESA, Gabriel CHINDRIS
Abstract: This paper studies a particular single-stage power-factor-correction 
(PFC) switching regulator employing a discontinuous-conduction-mode (DCM) boost-
input cell and a continuous-conduction-mode (CCM) forward output cell. It also 
presents the design method for boost-input cell and forward output cell and SPICE 
simulation of this scheme. As final results it's presents the values of total 
harmonic distortion and power factor for open-loop circuit.
Keywords: Power factor correction, switch-mode power supply, SPICE simulation
 
PROTECTION OF THE HIGH VOLTAGE BIPOLAR TRANSISTOR
Niculaie PALAGHITA, Cristian FARCAS, Cristina VLAD
Abstract: The switching behaviour of the high voltage bipolar transistor and his 
protection protection are treated in this paper. In the steady regime, no 
signifiant problem occurs, because v(;E and ic are nearly zero (the dissipated power 
isn't important), while the turning off process subbmite the power device to an 
important electrical stress. The effect of parasitic inductances will be analyze in 
order to give solution how to protect the power device. Snubber circuits arc used 
as protection network on turn process. On the other hands, RLD circuits while 
assist the turn on process. Driving protection method is described in this paper.
KEYWORDS: HVBT, turn on, turn off, protection, Snubber circuits
 
ANALYSIS OF RF SIGNAL ENVELOPE DISTORTIONS IN AM RADIO RECEIVERS IN COMBINED AM 
AUDIO AND FM DATA TRANSMISSION
Vlad CEHAN, Nicolae D. ALEXANDRU 
Abstract: Data transmission on a subcarrier that frequency modulates RF carrier is 
a feasible technique for superimposing data on AM signals. Limits of variation for 
subcarrier frequency and frequency deviation in order to produce a quasi-stationary 
state are determined. Also, envelope distortions of the RF signal that undergoes 
both AM and FM modulation, produced by its passage through narrow band circuits 
found in the front-end stage of the AM receivers are analyzed. Keywords: AM-FM, 
Data Transmission, Radio Communication
 
IDDQ TESTING
Dan PITICA, Călin GHERGHE , Mircea DABACAN
Abstract: In CMOS circuits, the fault detection problem becomes more complex 
because bridging faults create indeterminate logic outputs. Apart from the 
difficulty of sensitising, these faults present severe difficulties in fault-effect 
propagation. Designing a circuit under scan rules or BIST methods does not solve 
the problem. To overcome this problem, testing methods based on power supply 
current have been developed. These methods have been found capable of detecting 
bridging faults, open faults and some parametric faults such as gate oxide leakage 
in CMOS circuits. This technique is popularly known as lDDQ testing {1], In the 
paper it is described the basic concept behind Id(>q testing, its advantages, 
limitations and available implementation methods.
Key words: IDDO current testing, CMOS gate, fault-free current.

HIGH LEVEL DESIGN AND IMPLEMENTATION OF THE M-D NONRECURSIVE FILTERS
Adrian BURIAN, Mihaela CARLUGEA, Sorin HINTEA, Lelia FESTILA
Abstract: This paper presents high-level synthesis techniques for implementing from 
1-D to multidimensional M-D nonrecursive digital filters. As the complexity of 
ASICs increases, needs for new design methods are obvious. New methods are needed 
to manage large and complex designs. Hardware description languages and especially 
VHDL serve as a flexible and efficient tools to achieve these goals.. The 
architectures presented in the paper are very regular and support single chip 
implementation in VLSI, as well as multiple chip implementations. The proposed 
systolic arrays, used in implementation of these algorithms, are optimal with 
respect to time.

----------------------------------------------------------------------------------
Number 2, 1998

MATHEMATICAL ASPECTS IN SWITCHING ANALYSIS OF THE POWER MOS TRANSISTORS
Niculaie PALAGHITA, Doru CREANGA, Cristina VLAD, Cristian FARCAS
Abstract: In this paper is present mathematical aspect in switching analysis of the 
power MOS transistors. We calculated and represented graphic the following signals: 
IB. ics. iGi, vGS, ds . At the end, we make a simulation on MATLAB, with precise 
values. The simulated waveform could be compared with catalogue papers.
KEYWORDS: switching times, gate charge circuit, load current.
 
A NEW METHOD FOR IMPEDANCE MEASURING VIA RF COUPLED COILS
Sorin HINTEA
Abstract: Some medical applications need to transmit information or energy to 
implanted devices placed inside the human body via RF coupled-coils. In these 
situations it is difficult to control the position of implanted coil and, as a 
consequence, the coupling coefficient k. This paper describes a particular 
application which permits Li insensitive transmission link. There are presented the 
theoretical relations and also the simulation and experimental results. It is 
introduced a class D amplifier scheme in order to improve the circuit's 
performances. I present here a new application of signal transmission via RF 
coupled-coils: how to measure impedance without electric contact by implanted 
devices. This application is based on the insensitive transmission gain in this 
kind of circuits. The method is theoretically demonstrated and it is simulated by 
PSPICE. The restrictions are discussed and the measuring range is presented. I 
discussed the possibility of using class D amplifiers and the consequences of this.
Key words: coupled-coils, insensitive gain, impedance measuring, implanted device, 
class D amplifier
 
DIFFERENTIAL IMAGE COMPRESSION
Luminita SCRIPCARIU, Petrut DUMA
Abstract: In this paper a binary image coding technique is proposed. Its purpose 
is to ensure a better image compression rate, maximizing the white pixel area. 
Performances of image compression using partitioning into rectangular regions [1] 
may be improved using the proposed image coding method. It also be used for more 
efficient image segmentation, edge detection and for artificial or natural textures 
processing [2]. The binary image matrix is processed in two orthogonal directions, 
along rows and columns, from left-to-right and from top-to-bottom, running 
exclusive-or function of a pair of pixels separated by some other values. A simple 
algorithm reconstruct the original image. Test images of different types, sizes and 
complexities are used to prove the efficiency of the presented method.
 
STABILITY ANALYSIS OF HIGH-ORDER SIGMA-DELTA MODULATORS
Ana RUSU, Serban LUNGU
Abstract: A stability analysis method for delta-sigma modulators using a linear 
model for modulator is described. This method draws out the dependence of stability 
on system parameters such as integrators gains or quantizer gain. Furthermore, we 
show that a stable third-order modulator will become unstable as the input signal 
amplitude is increased beyond a certain threshold. In this paper we have used the 
results of the theoretical analysis and the facilities of the MATLAB software for 
analyzing the stability of the sigma-delta modulators. The method can be extended 
to high-order sigma-delta modulators.
Keywords: analysis, stability, model, modulator, delta-sigma
 
SOME METHODS FOR SUPERIMPOSING DATA ON AMPLITUDE MODULATED RADIO SIGNALS
Vlad CEHAN, Nicolae D. ALEXANDRU 
Abstract: Data transmission using existing broadcasting AM transmitters must 
satisfy some essential requirements. A few possible solutions are translating the 
spectrum of the data bearing signal or exponential modulation of the RF carrier by 
the data signal or by a subcarrier. Translating the data spectrum is not a feasible 
solution. The possibility of data transmission using either amplitude or frequency 
modulation on a harmonic subcarrier that in turn frequency modulates the RF carrier 
of an existing AM broadcast radio transmitter is proved. The essential requirements 
are further investigated for this case. Using a complex modulation (both ASK and 
FFSK) one can ensure both phase and frequency of the RF signal.
 
MAXIMUM LIKELIHOOD DECODER USING HOPFIELD NEURAL NETWORKS
Sorina ZAHAN
Abstract: In this paper a neural approach to the decoding of error correction block 
codes will be presented. The decoder uses continuous Hopfield memories, pertaining 
thus to the class of soft decoders. Nevertheless it can be used as a hard decoder, 
if necessary. Due to the properties and behaviour of the Hopfleld networks, the 
decoder acts on maximum likelihood bases. For this reason we named it neural 
maximum likelihood decoder (NMLD). It offers important advantages compared to 
classical hard and soft decoders but also compared to the neural hard decoder we 
have proposed in a previous work. The experiments revealed good j decoding 
performances, underlining the interesting behaviour of the Hoplield networks, 
which, once again, proved to be not enough exploited.
Keywords: Hopfield neural networks, error correction, soft decoding, hard decoding
 
-----------------------------------------------------------------------------------

Number 1, 1999

QRS DETECTION WITH A CREATED TEMPLATE FROM THE ECG SIGNAL TROUGH INTER-
CORRELATION
I. P. MIHU
ABSTRACT. In purpose of the identification of the QRS complexes, the encouraging 
practical results obtained through the correlation of the ECG signal with 
template matching shows that the performances are better as the matching of the 
template with the real form of a QRS complex is better. That's why for the 
monitoring in real time of a cardiac patient, the playback of a program sequence 
is possible, which in the first moments of the monitoring must have as a target 
to determine a template taken exactly from the ECG signal, came from the 
patient. After a remake, this template will be used during the entire effort 
test The trade of the solution proposed in this article, consists in the fact 
that the digital processing algorithm conceives the "self correlation" of the 
ECG signal. In the same time it removes the isoelectrical line derivation 
effect. Although it treats a precise signal (ECG), the solution could be 
extrapolated and used to identify specific shapes within other signals.


EXPERIMENTAL RESULTS OF TRAFFIC MODELS FOR BURST DATA AND VOICE SOURCES IN ATM 
NETWORKS
Virgil DOBROTA, Daniel ZINCA
Abstract: This paper presents an overview of the existing traffic models for 
burst data and voice sources in ATM, as well as some experimental results and 
comments about their implementations. The first paragraph is devoted to burst 
traffic generated by ON/OFF sources of constant throughput. The Matlab-based 
scheduler is able to determine the number of ON cells to the number of OFF cells 
ratio, for every burst, until the transmission process is completed. In the 
cases of multiple constant bit rate sources or a single source of variable bit 
rate, the superposition of several independent models previously discussed can 
be applied. This is method is not useful for voice traffic, possible solutions 
being offered by Markov-Modulated Poisson process model or by fluid source 
model. Experiments carried out gave the Matlab-based schedule of ATM cells 
departure, generated by 5 multiplexed sources.

SYSTEM LEVEL MODELING OF THYRISTOR RECTIFIERS
Liviu MODRAN 
Abstract: Some simple and effective thyristor rectifier dynamic models, suitable 
for the study of high power rectifier systems within the MATIAB/SIMULINK 
environment, is presented. This non linear, fast and accurate model, which 
accepts, as inputs, the pulse number and the trigger angle, is adequate for any 
pulse number greater than one, for all rectifier associations and reproduces the 
static and dynamic features of a real rectifier. The overlap phenomenon can also 
be included at the expense of some simplicity and simulation speed.


DATA ACQUISITION SYSTEM ON PARALLEL PORT
Ioan CIASCAI
Abstract - In very many experiments or laboratory measures it arises the problem 
of measuring analog signals with high precision which must be process on a 
compatible IBM-PC computer. The realization of a good acquisition system is 
generally expensive, and several problems appear, especially in making PCB, 
supply sources decoupling and noises induces by switching supply sources. The 
system presented in this paper has a very low price, is very simple and can be 
made very easy. Also the system can be made on an experimental board without 
affecting the precision, if the decoupling of power sources and reference is 
done like it's presented in the data book


TECHNOLOGY MAPPING FOR MULTIPLE-VALUED CIRCUITS
Dorin SIMA 
Abstract: The task of logic synthesis is to convert the logic description of set 
function into a netlist of gates that implements the functions. This paper 
describes the possibility of implementing some combinational and sequential 
circuits with multiple-valued PLAs (MVPLA) or by multiple-valued multiplexers 
(MVLMUX). The algorithms are based on multiple-valued decision diagrams (MDD) 
representation of the functions

OPTIMISATION OF THE "FAN" COMPRESSION ALGORITHM
Ioan P. MIHU
ABSTRACT. This article presents a compression algorithm through the fan method, 
applied and optimised in a specific way to the signal obtained from ECG and 
having as a purpose, the conservation of data of cardiac patients under 
observance. The support of the method, just as the mathematical apparatus that 
stood at the base of implementation, and finally the original criteria for 
optimisation of the specific implementation for each patient separately, are 
presented.

------------------------------------------------------------------------------

Number 2, 1999

PERFORMANCE AND AVAILABILITY CRITERIA FOR DIGITAL RADIO-RELAY AND SATELLITE 
SYSTEMS
Tudor PALADE
Abstract: This paper deals with the study and comparison between the G.821 and 
G.826 ITU-T Recommendations. The error performance and availability objectives 
have been pointed. The G.821 parameters are based on bit errors while G.826 
parameters are based on block errors. The relationship between block-error and 
BER has been analyzed and a method has been described that makes it possible to 
convert old bit error measurements into block-errors which can be evaluated 
according to G.826 parameters. The error performance objectives according to 
G.826 are known to be more stringent than G.821 objectives. This fact has to be 
taken into account when predicting outage probabilities. In the concept of 
unavailability all causes of quality degradation should be taken into account 
not only the propagation effects.
Key words: digital radio-relay performance and availability, bit errors, block 
errors.


THE REFLECTIONS PRODUCED BY LUMPED LOAD OF DIGITAL SIGNAL LINE
Virgil GOLUMBEANU, Paul SVASTA
Abstract: In interconnection technique of digital circuits in general the 
matched lines method for reducing the influence of reflections in a view to a 
good working of interconnected circuits is used. Parallel terminated 
transmission line is often used. Due to the input capacitance of digital 
circuits appears a reactance load. Although the matching is very good realized 
from the viewpoint of the resistive load, there is a reactive mismatching that 
can produce reflections at high frequency. Using Laplace transformation the 
reflections produced by parasitic capacitance load will be determined. These 
reflections will be analyzed depending on the type of the digital circuits and 
the parameters of the interconnection line. For some types of digital circuits 
the maximum fanout from this point of view, will be determined.


A SIMPLE METHOD FOR THD-COMPUTING
Gavril TODEREAN , Andras BALOGH, Balazs SZEKELY, Dezso DENES
Abstract; This work presents a time-domain method for Total Harmonic Distortion 
(THD) computing, avoiding the classical iiarmonic analysis (FFT). The proposed 
formulas are intended for microcontroller-based AC-Power measuring instruments, 
using the digital signal processing of the voltage and current vaweform. We 
assume a simplified waveform aquisition, without anti-aliasing filtering before 
sampling. However the theoretical aspects of undersampling are well known, we 
try in this work a cantitative evaluation of the error implied in THD-computing. 
We compute the THD in both time- and frequency-domain, using as test signals 
four analythically known, periodical waveforms. The signals are produced by 
Fourier-synthesis, resulting a high resolution digital representation of the 
test waveforms. The THD-value obtained with our method applied over the test 
signals is compared to the true THD-value, obtained by summing the first 100 
(more then all relevant), exactly known harmonic components. The simulation is 
done over four waveforms, taking as variable the number of samples/period and 
the number nuber of significant harmonic components present in the test-signal, 
trying out two numerical integration methods. The relative errors obtained are 
collected in tables. The simulation does not cope with the finite word-length 
effect, which occurs in a real implementation.
Keywords: THD, errors, FFT, simulation, computational cost


REGION-BASED FRACTAL IMAGE COMPRESSION USING DETERMINISTIC SEARCH
Macarie BREAZU, Gavril TODEREAN 
Abstract: The paper introduces a new method based on deterministic searcii to 
fractal image compression. In order to find a good region-based partitioning, we 
propose a deterministic search method for finding the blocks to be merged. For 
each range, a list of best N domains is maintained. When two ranges are to be 
merged and their common edge disappears, for the new range the best N domains 
are selected only from the 2xN domain extension of the two ranges. At each step 
the edge with minimum collage error increase is deterministically selected and 
the two corresponding ranges are merged. The process starts with atomic blocks 
as ranges and ends when the desired number of ranges is achieved. In order to 
reduce the encoding time, a suboptimal initialization method is also considered. 
Experimental results prove that our method yields a better rate-distortion curve 
than the classic quad-tree partitioning scheme.


Speeding up the fractal image encoding by six-class statistical classification
Mihaly MILIAN , Denes PAZMANY
Abstract: Lossy image encoding by Partitioned Iterated Function Systems, also 
known as Fractal Image Compression has recently became an attractive research 
area. As a result of the collage theorem (Banach's fixed point theorem) in 
metric space theory, with a few constraints a complete set of contractive 
transformations is guaranteed to produce an approximation to an original image 
when is iteratively applied to any initial image with the same H/V ratio.
While quick decompression algorithms exist, the compression process due to an 
exhaustive search is extremely time consuming. The most common solutions to this 
problem involves classification methods of domain and range blocks (the inputs 
and the targets of the transformations) according to some criteria, such as the 
existence of sharp edges or smooth areas, after which matches across class 
boundaries are excluded.
We propose a six-class classification of image blocks based on the uniformity of 
it's quadrants. The decision is made depending on the dispersion to mean value 
ratio comparing the value computed for a block with those resulted for the four 
quadrants on a logarithmic scale. This approach results in an important speed up 
of the encoding process while the quality of the reconstructed image remains the 
same.

------------------------------------------------------------------------------

Number 1, 2000

MIDI VIOLIN IMPLEMENTED ON TMS320C31 DSP STARTER KIT
Adrian FRATILA, Eugen LUPU
Abstract: This paper will discuss the technical design of a MIDI Violin, an 
interface between a classical violin and a synthesiser, more exactly the 
possibilities to transform in real time an audio signal into a MIDI signal, the 
last used as protocol for the digital transmission of audio events between 
synthesisers. To transmit music as MIDI messages we need to know the values of 
amplitude and fundamental frequency of violin sound. The proposed solution for 
the fundamental frequency estimation is based on direct methods performing first 
a rough fundamental frequency estimation using DFT and then around that 
fundamental frequency to calculate the Chirp Z Transform which can deliver a 
fine frequency resolution, approached to the human ear perception. The 
application has been implemented using a TMS320C31 DSP Starter Kit with an 
extension board designed by the author that board contains 128Kwords (32 bits) 
SRAM , 32Kbytes EEPROM for the boot operation and a MDDI IN/OUT interface.



QUALITY AND AVAILABILITY IN MOBILE SATELLITE SYSTEMS
Tudor P.PALADE, Ioan CHISALITA, Nicolae CRISAN, Mircea ARUNCUTEAN
ABSTRACT:   Fade margins for mobile satellite systems with forward error control 
operating in Ricean fading channels is considered. The fade margin is calculated 
according to combined criteria where quality requirements on both the average 
bit error rate and the instantaneous bit error rate is imposed during the time 
the system is available. This approach result in considerable less calculated 
fade margins compared to the traditional approach based purely on the channel 
fade statistics. 


LAYER 4 SWITCHING IN A TCP/IP ENVIRONMENT FOR THE ATM SOURCES
Virgil DOBROTA, Daniel ZINCA, Cristian Mihai VANCEA, Aurel VLAICU
Abstract: This paper is focused on the results of a Layer 4 switching 
experiment, aiming to evaluate the performances at the interface between the 
applications and the nonblocking stream-oriented sockets in TCP/IP. One major 
objective is to apply the traffic models for burst traffic and video sources, 
initially designed for ATM sources, to user applications requesting transport 
layer services.


A NEW THREE-PHASE MULTILEVEL VOLTAGE SOURCE INVERTER
M. Lucanu, N. Cozma-Popescu, N. Lucanu
Abstract: The paper presents a new topologies of three-phase multilevel 
inverter. The correct operation of the inverter was verified by simulation.
                   
                   
ANALOG MODELS FOR ANALYSIS AND DESIGN OF SOME SWITCH CAPACITOR CIRCUITS
Lelia Festila, Mihaela Cirlugea, Calin Coltea, Corneliu Rusu
Abstract : Unitary and systematical analog models for analysis and design of some 
switched capacitor (SC) circuits are presented. These models permit a direct 
connection to s domain and a simple determination of the SC circuit function. 
They also permit the reconfiguration of a RC-OA circuit, so that it should be 
implemented in the SC mode. The proposed analog models adapted for the SC circuits 
avoid the confusions introduced by usual analog models used in literature in the same 
purpose. We analyzed and designed some circuits taken from literature as example 
and proved the simplicity and efficiency of this method.


ELECTRONIC DEVICES MODELING BY FUZZY LOGIC INTERPOLATOR
Gabriel OLTEAN, Mihaela GORDAN 
Abstract: An important application of fuzzy logic systems is the approximation 
of non-linear curves. This application has been successfully used in device 
modeling based on their response to one input stimulus; an example is the I(VD) 
characteristic modeling for a diode [3]. In this paper we propose a 
generalisation of the fuzzy modeling from [3] to the modeling of a three-
dimensional surface y(x1, x2) and an application of this fuzzy modeling 
principle to describe the electronic devices behaviour, when this behaviour 
depends on two input variables. As an example we consider the current dependence 
on voltage and temperature for the rectifier diode, I(VD, T), and the transistor 
characteristic Ic (VBE, VBC).

--------------------------------------------------------------------------------

Number 2, 2000

TWO CAN BUSES REAL-TIME COMPLEX SYSTEM
Laurentiu DIMITRIU, Liliana VORNICU
Abstract: In this paper we present a two CAN buses real-time complex system. 
Using an original approach of system's timing we propose a powerful structure of 
internal data transfer. This is based on the features of Control Area Network 
(CAN) protocol, developed by Bosch. Considering a multi-processor structure, we 
have grouped the systems on functional criteria and use an external bus to 
interconnect the systems. Each system has an internal bus for its data transfer 
purposes. To amplify the system's possibilities, we propose the use of a 
multiplexing circuit to multiplex the internal serial buses.


ASSOCIATIVE MEMORIES ON CELLULAR NEURAL NETWORKS USING THE EIGEN VALUES 
DECOMPOSITION METHOD
Corina BOTOCA
Abstract: It is presented and theoretically justified the eigenvalue 
decomposition design method for an associative memory, or a cellular neural 
network . Introducing the condition of symmetric interconnections it is 
developed a new desigu method. This method generates a completely stable 
associative memory. Some simulation results are exposed and the performance of 
the obtained associative memory is compared with other differently designed CNN 
memories.


A WATERSHED ALGORITHM AND IT'S APPLICATION IN IMAGE SEGMENTATION
Arpad Zsolt BODO, Romulus TEREBES, Monica BORDA
Abstract: One of the most interesting and useful outcomes of mathematical 
morphology is the watershed operation. Algorithms, which result this operation, 
are numerous. This work describes one of these sequential algorithms, and also 
the operations needed for its implementation. The watershed operation or 
transform is a choice of image segmentation method in the field of mathematical 
morphology. Various examples are given, which shows the largeness of the 
applicability of this watershed algorithm in segmentation.


ON GENERALIZED WELCH MULTISETS SEQUENCES FOR SYNCHRONOUS-CDMA CHANNELS
Paul COTAE, Felix DIACONU
Abstract: In this paper we describe the properties of those sequences (matrix) 
that meet equality in the Welch's lower bound (WBE) on total squared 
correlation. The generalized WBE sequences (matrix) are defined. A new algorithm 
for generating WBE spreading sequence multisets is proposed. We apply the 
proposed algorithm when all users have different average-channel-input energy 
constraint in the case of synchronous COMA channels. We obtain real optimum 
signature sequences that maximize the sum capacity of S-CDMA channels.


A NEW METHOD TO DEDUCE THE VOLTAGE TRANSFER CHARACTERISTIC FOR SOME TWO-PORT 
NETWORKS
Gabriel OLTEAN, Mihaela GORDAN, Ioana OLTEAN
Abstract: A new method to deduce the voltage transfer characteristic for two-
port networks containing bias sources is presented. The method allows the 
deduction of the transfer characteristic only by translating a known transfer 
characteristic in the input voltage-output voltage plane; this way the analysis 
of the whole circuit is no longer necessary. Also the method can help to 
redesign a circuit in order to obtain a circuit that keep unchanged the shape of 
the transfer characteristic, while changing its location in the input voltage-
output voltage plane. The method is important from the didactical point of view, 
giving to the students new insights about some electronic circuit analysis and 
design.


SURFACES RENDERING OF MEDICAL IMAGES
Aurel VLAICU, Nicolae CRISAN, Radu LUCIAN
Abstract: The paper presents a three dimensional rendering method which 
accomplishes very good results in the field and needs only the 3D surface 
information and not the whole volume of data, as other algorithms in the 
literature. Thus, only the shape information is used for the reconstruction, 
using less voxels that have to be read from the 3D volume and, thus, less time 
for the display of the object on the screen. The main disadvantage is that no 
information is available for volume visualization. The matter is questionable, 
though, since a volumetric algorithm can solve this problem, in which case no 
shape rendering is necessary. The method presented in the paper assumes that the 
edge information is available and stored already on disk. No interpolation 
algorithm is needed before this, because all the operations are done in the same 
time as the algorithm itself. The main idea is to implement an algorithm of 3D 
surface rendering for medical images of human organs, acquired from CT's or MRI 
systems, using the wire frame method, which is usual in 3D graphical 
visualizations.


PROGRAMMABLE PSEUDO-RANDOM GENERATOR
Viorel OLARIU, Sorin OLARIU
Abstract: Pseudo-Random (PR) sequences are used as spectrum-spreading (SS) 
modulation for direct sequence SS design, as hopping pattern sources in 
frequency and/or time hopping systems, and as filter section controllers in 
matched filter SS systems. The basics of linear feedback shift register (LFSR) 
are described for short. A complex programmable PR generator is proposed and 
investigated. The paper also presents the on board high-resolution frequency 
synthesizer and the quadrature modulator.


A REAL TIME APPLICATIONS AND NON-REAL TIME APPLICATIONS EXECUTING ALGORITHM
Liliana VORNICU, Laurentiu DIMITRIU
Abstact: A complex system contains real time applications and also non-real time 
applications. In distributed complex systems working on high speed processors, 
the system's applications executing becomes difficult because the non-real time 
applications have deadlines high values, theoretically infinite and the timing 
constraints are absent. By choosing a proper executing algorithm, the real time 
applications and the non-real time applications can be forced to respect their 
timing constraints, periods and deadlines. So, the executing problems in a 
complex system are solved concerning these constraints.


STUDY OF THE PERFORMANCES OF DIGITAL AUTOMATIC GAIN CONTROL METHODS BASED ON THE 
ENERGY OF THE RECEIVED SIGNAL
Zsolt A. POLGAR,  Gavril TODEREAN
Abstract: The paper presents in details a certain automatic gain control (AGC) 
method frequently used in digital modems, based on the computation of the 
received signal energy, and a study concerning the performances of this method. 
Such a study is important especially in the case of the modems which do not use 
adaptive equalizers, due to two motifs: the level variations introduced by the 
transmission line increase the error probability after demodulation, and they 
can affect other circuits of the modem, increasing also indirectly the error 
probability.
The performances of the AGC circuits, as the performances of other circuits of 
the modem, are affected in a crucial manner by the intersymbol interference due 
to the filtering of the data signal. One of the main purposes of the present 
study is to emphasize the effects of the intersymbol interference on the 
performances of the considered AGC circuit and to present some to these 
problems. The obtained results could be generalized to other AGC circuits, with 
some approximations.


A FAST AND SIMPLE APPROACH TO CHARACTER RECOGNITION BY FUZZY IMAGE MODELING
Mihaela GORDAN, Eugeniu MECIU
Abstract: A simple class of techniques for character recognition is the one 
based on searchable databases, when the character described by some features is 
compared to a number of models described by the same features, and the character 
is labeled according to the best matching. The recognition speed mainly depends 
on the number of operations required for the comparison. The character 
recognition system must be insensitive to character rotation, distortion and 
additive noise. Usually both requirements: speed/robustness are not easy to 
attain. Here we propose a fuzzy logic-based system for character recognition, 
faster than some existing solutions, with reasonable invariance performances.


A METHOD OF SINE ROM COMPRESSION IN DDS
Viorel OLARIU, Sorin OLARIU 
Abstract - The spectral purity of the direct digital synthesizer (DDS) is 
determined by the resolution of the values stored In the sine table ROM. 
Therefore, it is desirable to increase the resolution of the ROM, Unfortunately, 
larger ROM storage means higher power consumption, lower reliability, lower 
speed, and greatly increased cost The basic memory compression techniques are 
briefly investigated. A new memory compression technique is proposed and 
investigated in detail. Computer programs have been created to generate the 
lookup table and to simulate the effects of the memory compression.
Keywords: sine ROM compression, lookup table, direct digital synthesis, 
frequency synthesizer


SOME OBSERVATIONS REGARDING THE REFLECTION LOSSES
D, Pitica, C. Farcas 
Abstract: This paper presents some observations regarding the reflection losses. 
The paper presents reflection mechanism for plane-waves. The attenuation factor 
by reflection is calculated in the far field region and in the near field region 
(for E-field and H-field). Also, is presented a comparison between attenuation 
factor for plane wave from calculus and the real attenuation factor for plane 
wave. 


DEDICATED ASSISTED DIAGNOSIS USING AN IMAGE ANALYSIS APPLICATION
Mircea VAIDA, Arghir SUCIU, Titus MOLDOVAN
Abstract: Image Analysis can offer useful results in the process of clinical 
diagnosis. Interactive measurements offer good results especially in research 
activity. Frequently, in clinical diagnosis we need to make automated 
measurements with and adequate visualization of the analyzed particles or 
regions of interest These measurements depends on the image quality and the type 
of measurements that must be done. Our application offers some facilities 
referring to image processing and pattern recognition and some dedicated 
elements of stereological, AgNOR and DNA measurements for assisted diagnosis. 
Results visualizations are very important to consider in the automated process 
of measurements. The DNA measurements process will consider an intelligent Look 
Up Table that preserve the intensity of the image. New aspects and features that 
can offer better results were also considered. New considerations referring to 
multi-sensorial pattern recognition techniques are also considered

--------------------------------------------------------------------------------

Number 1, 2001

SECURE COMMUNICATION SYSTEMS USING THE TMS320C5xDSP
Irinel Valentin PLETEA, Ion BOGDAN, Nicolae Dumitru ALEXANDRU
Abstract: This paper describes and characterizes secure communication systems 
over the Public Switched Telephone Network (PSTN). Global solutions are 
presented concerning the secure transmission of speech, data and facsimile group 
3 signals. The areas covered by this work are source coding/decoding and 
ciphering/deciphering.

A METHOD FOR PREDICTING ERROR PERFORMANCE AND UNAVAILABILITY DUE TO RAIN 
ATTENUATION
Tudor PALADE
Abstract: In order to predict error performance and unavailability according to 
ITU-T Rec. G.821 (and later on Rec. G.826) a distinction must be drawn between 
available and unavailable time. Error performance is specified for available 
time with a recommended measuring time of one month. For DRRS (Digital Radio 
Relay System) the worst month concept is used. The term unavailable time is 
defined in G.821. Unavailability is specified for a measuring time of one year 
or more. Rain attenuation is considered as a flat fade within the occupied 
bandwidth, which means that there is an obvious relationship between BER and 
input signal levels. For a given system, rain attenuation is easily convert to 
input signal level by subtracting it from the clear air reference level. Taken 
this into account, and using predicted annual and worst month rain attenuation 
distributions, the suggested method can be used to predict error performance and 
unavailability due to rain attenuation.

PERCEPTUAL SPEECH ANALYSIS
Pop G. Petre
Abstract : Perceptual analysis emulate human ear non-linear frequency response 
by creating a set of filters on non-linearly spaced frequency bands. Bark and 
Mel scales are the most used perceptual scales. PLP (Perceptual Linear 
Predictive) analysis is a combination between FFT method and LPC method which 
modify the short-term power spectrum by means of a series ol transformations 
based on physiological and psychological reasons and finally compute cepstral 
coefficients. Mel cepstral analysis use a cepstral smoothing in order to get the 
final smoothed spectrum.

SOME ASPECTS REGARDING THE DIGITAL GENERATION OF THE (CO)SINUSOIDAL SIGNALS 
EMPLOYING THE TAYLOR SERIES DECOMPOSITION
Vasile BOTA, Zsolt A. POLGAR
Abstract: The paper analyses the approximation error and the total harmonic 
distortion factor of the sinusoidal and cosinusoidal signals generated digitally 
using the approximation of the (co)sine functions by Taylor series 
decomposition. It presents the effects of the number of series terms retained 
after the series truncation, upon the two parameters of the generated signals, 
for different decomposition points and intervals. The effects of the finite 
format representation of the series coefficients, within fixed point DSPs, on 
the generated signals are analysed, as well.

DSP SOLUTIONS FOR BLDC MOTORS
Irinel Valentin PLETEA, Dimitrie ALEXA
ABSTRACT: This paper present the control of Brushless Permanent Magnet DC motors 
using the TMS320C24x. This new family of DSPs enables single chip, cost 
effective, a modular and increased performance solution for BLDC drives. A 
solution is presented below: control structures, power hardware topology, shaft 
position sensors, control hardware and remarks on energy conversion efficiency 
can be found in this document. In addition, this report deals with sensorless 
algorithms as an alternative to position sensors for speed control.

PERFORMANCE EVALUATION OF LAYER 4 SWITCHING IN IPv6 VERSUS IPv4
Daniel ZINCA, Virgil DOBROTA, Cristian Mihai VANCEA
Abstract: This paper is focused on a Layer 4 switching experiments of IPv6 over 
Fast Ethernet, running under Windows 2000 Professional. Prior to our studies on 
IPv4 over ATM with TCP relaying, firstly introduced at IEEE LANMAN'99, we are 
trying to evaluate the performances at the interface between the applications 
and the nonblocking stream-oriented sockets in TCP/IP. The first major objective 
is to get consistent results for a IPv6 versus IPv4 debate, even the 
implementation phase of this new version of the Internet Protocol is under 
progress. The second objective is to propose a method for measuring the round-
trip-time RTT at the Application Layer. It will provide 1 nanosecond - accuracy, 
that is requested by the new technologies such as Gigabit Ethernet.


FUZZY COMPUTING OF INITIAL SOLUTION IN ANALOG CIRCUIT DESIGN OPTIMIZATION
Gabriel OLTEAN
Abstract: A new fuzzy based method to compute the initial parameter values for 
the optimization design phase of electronic circuits is presented. A Mamdani 
type fuzzy system models the relations expressing the dependencies between 
circuit parameters and circuit functions. This fuzzy system is able to compute 
the values of design parameters for every set of design requirements. This way, 
the optimization phase has a good, automatic computed, starting point. Using 
this approach for a CMOS operational amplifier, initial solutions were computed 
for several sets of design requirements. Some of our results are much better or 
comparable than the ones presented in the literature.

AN IMPROVED SYNTHESIS METHOD BASED ON STATE VARIABLES OF Gm-C FILTERS
Doris LUPEA
Abstract: This article presents an improved synthesis method based on state 
variables of Gm-C filters. Our method presents the advantage of a reduced number 
of transconductors for the implementation of the Gm-C filters, unlike the case 
when the usually known method with state variables is applied. This can be put 
into practice if some transformations are being made to the passive prototype 
expected to be functionally simulated by the Gm-C filter.
This synthesis method has been used for the implementation of various types of 
filters. By PSpice simulations the correctness of the design has been checked 
out.

"PARALLEL" DRIVING
Niculaie PALAGHITA, Dorin PETREUS, Cristian FARCAS, Alexandra BORCA
Abstract: This paper makes some remarks referring to the power bipolar junction 
transistor driving. The paper presents Philips strategy of power bipolar 
junction transistor driving and proposes another strategy most efficiently. The 
proposed strategy is called "parallel" driving. Parallel driving supposes two 
phases. In first moment, in base of power bipolar junction transistor is applied 
a small negative current, when the stored charges from structure of transistor 
are reduced by recombination. After 0,2-KJu.s, it is applied a strongly negative 
current to obtain a quickly fall time of collector current and decrease the 
dissipated power to turn off. Also, the power bipolar junction transistor works 
in profound saturation, where has a small signal dynamic resistance.

THE INVERTER INPUT CURRENT RIPPLES USING DELTA MODULATION
Niculaie PALAGHITA, Dorin PETREUS, Alexandra BORCA, Cristian FARCAS
Abstract: This paper presents an analysis of the brushless DC motor functioning 
using two points adjustment (delta modulation) current control. It's presented 
the study of the motor current conduction transfer process from one phase to 
other phase, and the experimental results of the study are given in this paper. 
From experimental results we draw some conclusions.

SPICE MODEL FOR QUASI-RESONANT CONVERTERS CONTROL CIRCUIT
Ovidiu POP, Gabriel CHINDRIS, Silviu PLESA
Abstract: The zero voltage or current switching is one of the modern power 
conversion techniques. Starting from the quasi-resonant converters control 
circuit UC 3861 data sheets, this paper present a method to design a SPICE model 
for this control circuit. Are also presented the models for each functional 
block of the control circuits and a close-loop simulation for a zero voltage 
switching quasi-resonant converter.

THE HYPERBOLIC APPROXIMATION OF SOME NONLINEAR MAGNETIZATION CHARACTERISTICS
Victor POPESCU, Ioana SARACUT
ABSTRACT: This paper presents the possibility of approximating of some nonlinear 
magnetization characteristics in weak fields. The modified hyperbolic 
approximation will allow the control of the characteristic's curvature without 
affecting the slope in origin nor the position of the asymptote. The proposed 
approximation can be utilized in realizing some PSpice models of nonlinear 
inductance.

PSpice MODELS OF NONLINEAR INDUCTANCE AT LOW CURRENTS
Victor POPESCU, Ioana SARACUT
ABSTRACT: Modeling of weak electromagnetic waves propagation in ferromagnetic 
environments needs models of nonlinear inductance at low currents. Simulation 
programs offer only polynomial approximations of the magnetization 
characteristics. This paper presents models based on cubic and modified 
hyperbolic approximation. Comparative results are presented, from which results 
that the hyperbolic approximation based model is more flexible in adjusting the 
curvature of non-linear characteristic.

NON-PARAMETRIC CURVATURE ESTIMATION FOR GRAY SCALE IMAGES
Romulus TEREBES
Abstract: A digitised curve is viewed as a surface over the xy-plane. The level 
curves of this surface provide information about edge directions and feature 
locations. This paper presents a method for the extraction of tangent directions 
and curvatures of these level curves. The tangent direction is determined by 
performing a Principal Component Analysis. The curvature calculation is then 
performed, unlike most previous work, on non-parameterised curves; it works 
instead on the tangents across adjacent level curves. Examples of real images 
are given which show application of the proposed method in feature point 
detection

----------------------------------------------------------------------------

Number 2, 2001

NEW TECHNIQUE IN 3D MEDICAL CT IMAGES RENDERING USING OPEN GL OR DIRECTX
TECHNOLOGIES
Nicolae CRISAN
Abstract: The most effective 3D technologies, as Open GL and DirectX, are 
already used in virtual reality with good results. VR world has gone up last 
years, so the VR method as Open GL and DirectX could be the best solution for 
future 3D medical applications. The paper presents a new method, which prepare 
data, in order to be suited for 3D rendering, using DirectX or OpenGL 
technologies. The presented method could be considered a complete solution 
including capturing, edge detection, data interpolation, error decreasing and 
data rendering.

INTERACTIVE INTERNET-INTRANET DOCUMENTS WITH IMAGE PROCESSING FACILITIES
Mircea-FIorin VAIDA, Ovidiu BUZA
ABSTRACT: Image processing and WEB design are considered to develop medical 
image documents. We developed an integrated software application that allows 
image processing facilities with WEB set up documents considering HTML format. 
The process is interactive and automated and the resulted document can be 
accessed on a local Intranet or in Internet if the WEB document is located on a 
WEB Server. The domain of interest can be larger from medical application, to 
educational, marketing and industrial domain too.

INTELLIGENT ANTENNAS FOR FUTURE WIRELESS COMMUNICATIONS
Tudor PALADE
Abstract: The first part of this paper will present an overview of the impact of 
intelligent antennas for next generation radio including a review of major 
challenges for the realization of such wireless systems, followed by the key 
issues and the state of the art of marked potential and expected impact and an 
assessment of real world performance gains. The second part will be a 
description of intelligent antenna systems. This section emphasizes that the 
integration of basic antenna characteristics with fast, highly efficient signal 
processing solutions implemented in a Digital Signal Processor (DSP) provides a 
flexible system whose characteristics can respond to changes in the radio 
environment to provide efficient communications. The third part of this paper 
will address tht RF aspects of intelligent antennas. Major factors relevant to 
radio frequency (RF) integrated circuit design will be discussed, including the 
main directions in the antenna modeling and CAD development, antenna 
technologies and microwave techniques. The fourth part will be dedicated to 
practical applications of smart antennas to broadband wireless communications.

SOME CONSIDERATIONS REGARDING THE DIGITAL SYNTHESYS OF THE SINUSOIDAL SIGNALS 
EMPLOYING RECURENCE RELATIONS AND TABLE LOOK-UP
Zsolt A. POLGAR, Vasile BOTA
Abstract: The paper discusses the performances provided by two methods of 
digital synthesis of the sine waves, the first employing the recurrence, 
specific to sine-waves and the second based on topic reading. It presents the 
way the sinusoidal signals can be generated and the performances of these 
synthesis methods. Performances are evaluated by means of the root mean square 
error (erp) and total harmonic distortion coefficient (THD). As for the THD 
coefficient, the paper proposes a more accurate computation method. The 
distortions induced by the digital-analog conversion of the synthesized signal 
are also considered.

INDUCTION HARDENING COMPUTER AIDED DESIGN
Petre G. POP*, Marius VLAD**
Abstract : Many attempts of induction hardening is possible to avoid by 
predicting the technological parameters which are necessary to obtain the 
desired depth of the hardened layer. This impose the precalculation of the 
temperature fields inside the induction heated piece, based on a mathematical 
model which must consider all the phenomcnons which really happen : the 
temperature dependence of the material caracteristics, the great length of the 
piece with respect to that of the inductor, etc. We used an algorithm based on 
the numerical treatment of the differential ecuation of the unstationary thermic 
conduction, by means of the United differences method. The software, realized in 
Delphi 3.0, based on this method determine the technological parameters for the 
induction machine, editing the technological card, used directly by the 
operator.

THE INFLUENCE OF NETWORK FACTORS ON PROCESS MONITORING
Pavel HOROVCAK
Abstract: The contribution deals with the process monitoring based on www 
technologies and research the influence of clients' number and network transfer 
rate on process monitoring quality. The process monitoring is realized by 
distributed application. Server part of application (written in Delphi) acquires 
actual measured data and through the socket communication channel sends them to 
the client. The client part of application is realized as applet (written in 
Java), which receives data from server and executes their processing. There are 
some factors influencing the quality of client-server communication on server 
and client side- as number of running tasks, exploitation of system recourses, 
number of connected clients and network rate.

-----------------------------------------------------------------------------

Number 1, 2002

NEW EVALUATION CRITERIA FOR EDGE DETECTION ALGORITHMS SELECTION
Nicolae CRISAN 
Abstract: The most important in edge detection algorithms is a good evaluation 
of contour errors. Contour error depends on differences between original contour 
and detected one. Not a good definition of contour error and error formula is 
important but the contour reference definition. One very well known method is to 
apply one edge detector gradient, but these result contours can not be 
considered as reference, because of contour errors. In his article proposes one 
method, which offers this image reference without edge errors. An evaluation 
criterion and classification is shown for each well-known edge detection 
algorithm.

DRIVE CIRCUIT FOR INDUCTION MOTOR
Dorin PETREUS, Niculae PALAGHITA Cristian FARCAS
Abstract: This paper presents a drive circuit for induction motor. The circuit 
is design to (hive an IGBT inverter and is built around the 10 HEF 4752 from 
Philips. Theoretical and experimental results are presented. Circuit waveforms 
and a complete scheme of a drive circuit for induction motor are given.

ON THE ERROR PERFORMANCES OF THE TRANSMISSION SYSTEMS EMPLOYING ERROR CORRECTING 
CODES AND PARTIAL RESPONSE FILTERING
Zsolt A. POLGAR. Gavril TODEREAN
Abstract: One of the facilities provided by the Partial Response (PR) filtering 
is the insertion of some spectral zeros, where pilot signals, employed for the 
synchronization, can be transmitted. However, the systems employing PR filtering 
provide poorer signal/noise ratio performances. The paper analyses the 
employment of the high rate convolutional error correcting codes (CEC C) and of 
a combined CECC-PR soft-decoding method, in order to secure S/N ratio 
performances similar to the ones of the non-PR filtered systems, in about the 
same bandwidth.

PERIOD - DEADLINE SINGLE PAIR SELECTION METHOD IN REAL -TIME COMPLEX SYSTEMS
Liliana VORNICU
Abstact: In real time complex systems working on high-speed processors, it is 
very important for the designer to know the timing constraints actioning in the 
real time systems. Generally, these timing constraints have variable values (for 
example, periods and the deadlines, which are standard timing constraints) and 
the designer has to chose a suitable single pair period - deadline from the set 
of pairs. To do this, he must use an algorithm that offers a maximum utilization 
of the real time system. In the following, we will present a system utilization 
calculation method that permits the selection of a suitable pair.

ANALOG DESIGN: MULTIOBJECTIVE OPTIMIZATION METHOD BASED ON FUZZY LOGIC
Gabriel OLTEAN, Costin MIRON
Abstract: The paper presents a new method for optimizing the analog circuits 
design. The method performs a multiobjective optimization, the parameter 
modification taking into account the unfulfiliement degrees of all the 
requirements. The method uses fuzzy sets to define fuzzy objectives and fuzzy 
systems to compute new parameter values. The solution of a multiobjective 
optimization is a set of Pareto optimal points. We can obtain the set of Pareto 
optimal points using the idea of population of solutions. We select as the final 
optimal solution the one with smallest mean of unfulfiliement degrees of 
requirements (zero if possible). The strategy to compute new parameter values 
uses local gradient information and encapsulates human expert thinking: to 
improve a performance we modify mostly the parameter with more influence. This 
way every Pareto optimal point can be found with accuracy.
After introducing our optimization method, we optimize the design of a common-
emitter amplifier. As expected the results prove that the method is efficient. 
For different sets of requirements we found a set of Pareto optimal points and 
for the sets with non-conflicting requirements the final solution has the mean 
of unfulfiliement degrees equal to zero.

SECURE CLIENT-SERVER CONNECTING PROTOCOL USING ARAKI-UEHARA-IMAMURA DIGITAL 
SIGNATURES
Valentin DEAC, Monica BORDA
Abstract: A cryptographic system for a secure Client-Server communication which 
uses symmetrical keys needs a connecting protocol which deals with important 
problems such as: the authentication of the two parts involved, the 
confidentiality, the key exchange etc. In this paper we propose a connecting 
protocol based on public key cryptography which tries to solve the above 
mentioned problems. A Certificate Authority which respects the ITU-T X.509 - 
Directory Authentication Service standard [5| is used for the public key 
certificates. For the digital signatures involved in the protocol we use the 
Araki-Uehara-Imamura method [1].

ANALYSIS METHOD OF LOG-DOMAIN CIRCUITS BASED ON BEHAVIOURAL F-F ] MODELS
Mihaela CIRLUGEA, FAZAKAS Albert, Lelia FE§TILA
Abstract: We propose a simple method to analyze some log-domain filters by using 
behavioral models based on inverse F-F"1 functions, leased on intuitive symbols 
and models, the method is easily to be applied and understood and can also be 
used in the educational process.

DIGITAL WATERMARKING BASICS
Monica BORDA 
Abstract: This paper presents a brief overview of a relatively new research area 
(the digital watermarking) with great implication in the copyright protection 
for digital multimedia technologies. This research direction, started two years 
ago, is one of the most important opened in the Research Center in Information 
Science and Signal Processing of TUC-N, guided by prof. dr. eng. Monica Borda. 
This work is sustained by contracts funded by the National Agency for Science, 
Technology and Innovation (ANSTI) and by the World Bank. Research agreements in 
this domain are done with Plymouth University (UK), the Kyushu Institute of 
Technology (Japan) and Northern Jiaotong University Beijing (China).

DIGITAL WATERMARKING USING HASH FUNCTIONS
Reka MAJOR, Valentin DEAC, Monica BORDA, Dr. Graham WADE, Cristian SERDEAN
Abstract: The digital watermarking is used to protect the intellectual property 
rights in the multimedia field, It consists of algorithms which embed in digital 
multimedia data such as video, image or audio, an invisible information related 
to its owner and its current users. In order to achieve a proper frequency 
spreading and to protect the watermarked data, a cryptographically secure 
pseudo-noise signal is needed. In this paper we propose a technique for 
generating a watermark signal which depends on the original data and resists at 
some intentional and unintentional attacks. In the generating process we use 
hash functions and the DES algorithm.

-----------------------------------------------------------------------------

Number 2, 2002

AN APPROXIMATION OF SATURATION CHARACTERISTICS
Victor POPESCU
Abstract. Based on the modified hyperbolic approximation for nonlinear 
magnetization characteristics for low currents [9], this paper presents an 
extension of this approximation's application to the saturation characteristics. 
The approximation thus obtained has the same three qualities as the one used for 
low currents; (1) a single description for both linear and nonlinear portions of 
the characteristic, (2) independence of asymptote and curvature adjustment and 
(3) smoothness of the characteristic as well as of its slope. There are also 
some examples given in the paper. The characteristic's shape can be adjusted by 
four parameters; the parameter L controlling the slope in origin, the 
nonlinearity factor k, controlling the asymptotic slope, the breaking current 
(the abscissa of the asymptote and the slope in origin intersection point) and 
the nonlinearity factor k2 controlling the characteristic's curvature. The four 
parameters controlling the approximation can be easily determined for a given 
(experimental) nonlinear characteristic. The flexibility of the hyperbolic 
approximation allows it to match the given characteristic.

REPEATING SPACE TIME CODES FOR TURBO TRELLIS CODED MODULATION
Osman N. UCAN, Onur OSMAN, Omer ERKAN
Abstract: In this paper, we introduce a new type of Space Time codes denoted as 
Repeating Space Time codes (RST). Antenna diversity and using of Space Time 
block codes are effective ways to combat multi-fading and improve the error 
propagation over Rician especially Rayleigh fading environments. To evaluate 
this new antenna diversity scheme, we use Turbo Trellis Coded Modulation (TTCM) 
and we also compare Repeating Space Time codes with Alamouti's Space Time codes. 
For whole of the simulation results, four iterations are made for different 
Rician environments as K~<*>, 10, 0 dB and also we use two bit streams and each 
one has 1024 bits as an input.

SOME EXPONENTIAL STRUCTURES USED IN LINEAR APPLICATIONS
Lelia FESTILA, Albert FAZAKAS, Mihaela CIRLUGEA
Abstract: We analysed in this paper some exponential cells in order to put into 
evidence their connection with other linear universal building blocks based on 
the same basic schematic (current conveyor CCII+, Diamond transistor, Diamond 
current source, OTA). Some limits and requirements for the linear operation have 
been deduced. We also concluded that the sh cell is a general universal 
structure that can be used in both linear and nonlinear applications. Examples 
and simulations proved the validity of our analysis.

EVALUATION OF FAIR BANDWIDTH ALLOCATION POLICIES FOR ELASTIC TRAFFIC THROUGH A 
LINEAR PROGRAMMING MODEL
Ioannis D. MOSCHOLIOS, Michael D. LOGOTHETTS, George K. KOKK1NAKIS
Abstract: Fair Bandwidth Allocation (FBA) for elastic traffic is based on the 
Max-Min Fairness policy (MMK), initially proposed for non.eelastic traffic. We 
present a Linear Programming (LP) model, which describes the MMF policy and 
incorporates, in a parametric way, both weighted and un-weighted FBA policies 
for elastic traffic. We propose new FBA policies based on the MMF and 
considering either the number of connections or of Virtual Circuits (VCs) 
accommodated in a link, together with their traffic parameters. The number of 
links of a call-connection can be used as weight. Moreover, we evaluate the 
van«*,s FBA policies through the LP model and reach sound conclusions.

ORIENTATION DRIVEN DIFFUSION
Romulus TEREBES, Olivier LAVIALLE, Pierre BAYLOU, Monica BORDA
Abstract: In this paper we propose a new discrete version of the classical 
anisotropic diffusion filter. Our filter uses supplementary pattern orientation 
information, the diffusion process rest scalar. The efficacy of the proposed 
method is shown through application samples involving synthetic and ancient 
engravings real images. We address a recently mentioned undesired pinhole effect 
of the classical diffusion, and through experimental results we show that our 
method greatly reduces this effect.

WEB DESIGN FACILITIES FOR DISTANCE LEARNING PERSPECTIVE 
Mircea-Florin VAIDA, Dinu CHIRA
Abstract: Considering the perspective learning with web facilities we developed 
a dedicated application for distance learning considering the PHP language. The 
application will emulate a virtual university and is compatible under Unix and 
Windows operating systems. Some multimedia documents arc also considered in the 
learning process and will be also available with the chat facility implemented 
to improve the educational process. A comparison with other distance learning 
products is also made. Security elements are implemented with the SSL protocol.

ON THE APPROXIMATE DIGITAL GENERATION OF UNIFORM AND GAUSSIAN DISTRIBUTED NOISE 
SIGNALS
Vasile BOTA, Zsolt A. POLGAR
Abstract: The paper presents a digital method for generating the uniform and 
Gaussian distributed noise signals. The noise signals are approximated by means 
of a pseudo-random sequence (PRS) and an approximation step s. The paper 
analyzes the influences of the PRS length and s upon the accuracy of the 
Gaussian law observance, upon the amplitude range of the output signal and gives 
some brief considerations regarding the spectra of the generated signals. Some 
applications of the proposed method that were used by the authors in digital 
implementations are briefly described as well.

A MODULAR APPROACH TO STATE VARIABLE Gm-C FILTER DESIGN
Doris LUPEA, Gabor CS1PKES
Abstract: This paper presents a generalized, modular approach to Gm-C filter 
design. The synthesis method is based on state variables. For a modular design 
of these filters some basic building blocs must be designed. These basic 
building blocs are based on impedance inversion implemented with gyrators. The 
proposed synthesis method can be used to design any type of filter for a given 
approximation. Ail the experimental circuits have been simulated using PSpice.

COMPARISON OF FEATURE PARAMETERS USED IN SPEAKER RECOGNITION
Pop G. Petre, Toderean Gavril
Abstract: Speaker identity is correlated with the physiological and behavioral 
characteristics of the speaker, both encoded in the spectral envelope and in the 
supra-segmental features (voice source characteristics and dynamic feature 
spanning over several segments). Speaker recognition requires speaker 
characteristic features, independent of the particular spoken word if possible. 
Features based on short-term spectral estimate have a strong dependece on 
individual speakers and consequently are used in speaker recognition. In 
addition, pitch information is sometimes used if it can be estimated reliably. 
We study the influence of some features (LPC, LPC cepstrum, MFCC, delta- and 
delta-delta-) on speaker recognition using DTW and VQ methods.

CONCORDE: VOCATIONAL TRAINING THROUGH OPEN AND DISTANCE EDUCATION
Mircea GIURGIU, Lynn CONNAUQHTQN, Luis LIZAMA, Manolis CHRISOSTALIS, Simona 
COSCIA, Dana BENEA, Daniel DON 
Abstract: The paper presents intermediary results obtained in the frame of the 
EC funded Leonardo da Vinci pilot project with the title "Acquisition of 
Complementary Competencies through Open and Distance Education" (CONCORDE). The 
partnership is formed by seven institutions, three from Romania, with the aim to 
create a training network and develop training materials in the form of Open 
Distance Learning (ODL). In the first stage of the project a survey of training 
needs as complementray competencies was done by each partner and according to 
this, the partnership is in the phase of designing and development of ODL 
modules. These modules are going to be in the areas of: information technologies 
(basic and high level-telecommunications), Internet communication, injection 
moulding for plastic materials and business administration. Some of the 
intermediary results are disseminated in this communication.

OCTOPUS: AN APPROACH FOR A TRANSNATIONAL ON-LINE RESOURCE CENTRE
Vito CARIOCA, Rui GAIBINO, Antonio MEDINA, Tiberio F. MURIAS, Mircea GIURGIU, 
Joao C. CHOURICO, Luis SANTOS, George VAVIZOS
Abstract; This communication presents the main ideas behind the Socrates/Minerva 
project: "Transnational on-line resource centre" (OCTOPUS) dedicated to the 
development of a web resource centre in the specific field of environmental 
education, for this pilot phase. This resource centre will implement an 
inovative mechanism of information retrieval through a website which will allow 
the access of the users from different regions with common interests in the area 
of environmental education using modern electronic tools. The core of the 
resource centre is a very well organized database which is uploaded and searched 
using metadata descriptors. Apart of technical and scientific development, this 
site will offer open training activities which are the subject of this paper, 
too.

-----------------------------------------------------------------------------------

Number l, 2003

CHANNEL CAPACITY DESIGN IN INTEGRATED SELF-HEALING NETWORKS
Michael LOGOTHETIS, Ioanis NIKOLAOU , George KOKKINAKIS 
Abstract: Integrated self-healing networks combine more than one survivability
techniques and can achieve higher network reliability and cost-effective design. `We
present three pure heuristics algorithms for the design of spare and working channel
capacities of integrated self-healing networks. All algorithms are based on the
shortest path principle. Algorithm l and 2, contrarily to algorithm 3 do not take
into account possible restrictions in node capacities. In comparison to other
shortest-path based algorithms they are more homogeneous, perform better and achieve
considerable savings in total channel capacity of the backbone network. Algorithm 2
performs best under specific communication demand requirements, while algorithm 1 is
not sensitive to the communication demands.

FUZZY IMAGE ENHANCEMENT USING THE LOGARITHMIC MODEL
Vasile Patrascu
Abstract: The logarithmic model of image representations supplies a great number of
tools for analyzing and processing both the color images and gray level ones. Simple
and efficient methods for image enhancement can be obtained through affine
transforms, defined by the logarithmic operations: the addition and scalar
multiplication. The parameters of an affine transform are determined taking into
account the statistical features of the image. Most of the times, it is calculated a
single affine transform for the whole image, then the enhancement uses a point
transform. One can get a surplus of quality if a partition is defined on the image
support and, then, one determines an affine transform for each element of the
partition. Using a classical partition may lead to appearance of some
discontinuities when passing from one element of the partition to another one. The
necessity of eliminating all these drawbacks guides us naturally to the utilization
of the fuzzy partitions and from now on, to the calculus of the affine transforms' 
parameters in the context of the fuzzy sets theory. In this way, we come to calculate 
the fuzzy mean and the fuzzy variance within the logarithmic model.

A MODEL FOR ADAPTIVE HYPERMEDIA SYSTEMS
Rodica Baciu
Abstract: In this paper, we describe an abstract model of adaptive hypermedia
systems through extended conceptual base of general hypermedia model. The major
difference between unadaptive and adaptive hypermedia systems lies in the way nodes
are traversed and the rules that govern such traversals. Whereas traversal in
unadaptive systems is, generally, referential, traversal in adaptive system can also
be contextual, i.e. the link to follow is determined by the user model. In proposal
model we conceptualised additional facilities not available in the hypermedia model
and necessary in modelling of adaptive hypermedia systems.

SOME ASPECTS REGARDING THE ERROR PERFORMANCES OF THE LDPC CODES DECODED WITH THE
MESSAGE PASSING ALGORITHM
Vasile BOTA, Zsolt POLGAR
Abstract: The Message-Passing algorithm (MPA) is the algorithm employed for decoding
the Low Density Parity Check (LDPC) codes. Basically, this algorithm decodes every
bit of a codeword by using the Bayes criterion and uses the syndrome to check the
decoded codeword. Due to its adaptive and probabilistic nature, this algorithm
exhibits a particular behaviour, compared to other decoding algorithms. This paper
analyzes three aspects regarding this algorithm: the non-convergence of the
algorithm in terms of the LDPC code parameters, the influence of the LDPC code rate
upon the bit-error performance and a comparison between the performances of the MPA
and those of an ideal Maximum Likelihood (ML) decoding algorithm.

A NOVEL ENERGY RECOVERY SNUBBER
Niculaie PALAGHITA, Dan MICU. Cristian FARCAS, Dorin PETREUS
Abstract: In this paper, we propose a novel topology of energy recovery snubber. The
proposed topology of energy recovery snubber is obtained by modifying the Williams'
snubber. For this, we introduce a transformer to obtain a second path for energy
recovery. The mathematical equations that describe turn-on and turn -off transitions
are presented. The behavior of the Williams' snubber and the proposed energy
recovery snubber was experimentally analyzed.

ON THE STUDY OF THE MOTION WITH DOPPLER SYSTEMS WITH ACTIVE FIXED REFERENTIAL
Dan POPA, Corneliu TOMA
Abstract: This paper includes a brief introduction to the bases of the study of
motion in different environments with the aid of Doppler systems with
electromagnetic and ultrasound waves. There are given limits up to which they can be
used in an elastic environment of infinite dimensions. The paper sets the basis of
the study of the motion with Doppler systems with active fixed referential, offering
a solution for converging the Doppler shift in frequency and another solution for
the structure of the electronic equipment used for such a system. It also presents
the perspectives of using these systems in the field of scientific and technical
development.

TESPAR ALPHABET GENERATION USING STANDARD VQ AND COHONEN NN
Petre POP, Eugen LUPU, Gavril TODEREAN
Abstract. TESPAR (Time Encoding Signal Processing and Recognition) is a processing
and recognition method in the time domain, proposed by [1]. The key problem in
TESPAR is to define the alphabet used for the coding process, generated by a
quantization process. This paper presents two ways for TESPAR alphabet generation,
standard VQ and Kohonen neural networks.

APPLICATION WITH ATMEL AVR RISC MICROCONTROLERS
Radu Balan, Alin Gliga
Abstract: This paper presents a few applications of ATMEL AVR RISC microcontrollers
in mechatronics. This family of microcontrollers can be used to control different
types of actuators (DC motors, AC motors, stepper motors, shape memory alloy SMA,
etc.), different processes, automotive instrumentation, onboard computers, security
with code lock or guarding perimeter / entrances with IR on LASER beams, wireless
devices. The programming software is very easy to use even for non-specialists.
Different types of software can be used to develop applications: assembly or high
level programming languages such as C or Basic.

CLOSED-LOOP IDENTIFICATION IN ADAPTIVE CONTROL: SIMULATIONS AND EXPERIMENTS
Radu BALAN, Alin GLIGA
Abstract: This paper describes a few examples of closed loop identification in the
case of implementing of a model based predictive control which uses an line
simulations and rule based control. The most important step in adaptive control is
the model update (identification of the model parameters). The degree to which the
model is able to represent the process and automatically adjust to the process
changes, determines the accuracy of the calculated control action and the resulting
control performance. A Delphi application is used both for, simulations and
experiments. To connect to the process it is used a system with 80C552
microcontroller.

MODULAR DESIGN OF Gin-C STATE VARIABLE FILTERS FOR APPROXIMA-TIONS WITH ARBITRARY
TRANSMISSION ZEROS
Doris LUPEA, Gabor CSIPKES
Abstract: This paper presents a generalized, modular approach to Gm-C filter design
for approximations with arbitrary transmission zeros. The synthesis method is based
on. state variables. For a modular design a transconductance amplifier based
structure is defined, which is the same for all filter types and for all
approximations. The desired transfer function can be realized by some building
blocks based on impedance inversion implemented with gyrators. These blocks can be
derived from the passive prototype of the filter. All the experimental circuits have
been simulated using Pspice.

-----------------------------------------------------------------------------------
Number 2, 2003

REPEATING SPACE TIME CODES FOR TURBO TRELLIS; CODED MODULATION
Osman N. UCAN, Onur OSMAN, Omer ERKAN
Abstract: In this paper, we introduce a new type of Space Time codes denoted as
Repeating Space Time codes (RST). Antenna diversity and using Space Time block codes
are effective ways to combat multi-fading and improve the error propagation over
Rician especially Rayleigh fading environments. To evaluate this new antenna
diversity scheme, we use Turbo Trellis Coded Modulation (TTCM) and we also compare
Repeating Space Time codes with Alamouti's Space Time codes. For whole of the
simulation results, four iterations are used for different Rician environments as
K=8, 10, 0 dB and also we use two bit streams and each one has 1024 bits as input.

SECURITY ISSUES REGARDING WEB DISTRIBUTED APPLICATIONS
Cosmin STRILETCHI, Mircea VAIDA
Abstract: The aim of the paper is to present the security elements involved in web
applications. E-commerce applicataans and data access such as university's
information database may be protected using different security techniques.
Considering an e-commerce application, we developed some security elements that will
provide an accurate and safe data transfer mechanism. We adapted the application for
allowing our university's students to access their exam results and to view their
overall situation or some private data using web facilities.

TRAFFIC CONTROL MECHANISMS IN LARGE-SCALE IP NETWORKS
Evi Tsolakou, Anastasios Doulamis and Nikolaos Doulamis
Abstract: The DiffServ architecture has been proposed as a scalable solution for
providing service differentiation among flows. Towards the enhancement of this
architecture a new set of network services are proposed in this paper. Each network
service is appropriate for a specific type of traffic and is realized through its
own network mechanisms, which are the Traffic Classes. Traffic Classes are composed
of a set of admission control rules, a set of traffic conditioning rules and a
per-hop behavior (PHB). A large-scale network is considered as the reference
topology for studying the performance and effectiveness of the proposed network
services with the use of the OPNET simulation tool.

ADAPTIVE INTERPOLATION BETWEEN BIOMEDICAL SLICES, USING INTER-CORRELATION BETWEEN
VOXELS
Nicolae Crisan
Abstract: Medical slices, such as CT or MRI, are used for 3D reconstruction of human
organs like kidney and bones. The anisotropy along z-axis is the reason, for the
medical interpolation between 2D sections. This article, it is not focused on a new
interpolation techniques used to diminish the z-axis anisotropy, but an adaptive
interpolation method, taken into account in ader to optimized the performance of the
interpolation. No matter what algorithm of interpolation is performed on a medical
image, the idea presented in this article is to find the best direction throughout
the 3D medical scene along the interpolation will be done, taking into account the
maximum voxels inter-correlation.

AUTOMATIC REMOVAL OF NOISY DATA FROM DATA SETS. APPLICATION IN CIRCUIT FUNCTIONS
MODELING
Gabriel OLTEAN, Mihai CRASI 
Abstract: The problem to build a model of an unknown complex function from data sets
is a fundamental issue in a variety of scientific and engineering fields. Even in
the case of a large amount of data, there can be some situations for that same
representative features of the function are altered due to the noisy measurement. In
order to obtain an accurate model we have developed and implemented a method to
automatically find and eliminate the data points affected by noise. The interaction
between the user and the computer program is facilitated by a friendly graphical
user interface. We validated the method by using it to model two bi-dimensional
mathematical function: a linear one and a nonlinear one. As an application of the
method we use it to model a circuit function for a simple operational
transconductance amplifier..

GRAY LEVEL IMAGE ENHANCEMENT METHOD USING THE LOGARITHMIC MODEL
Vasile PATRASCU
Abstract: This paper presents a method for image enhancement using affine transforms
within the logarithmic model. The parameters calculus of the affine transform was
made after a short analysis of the well-known image enhancement method based on the
histogram equalization.

HIGH FREQUENCY, NARROW BANDWIDTH BAND-PASS FILTER SYNTHESIS BASED ON Gm-C BIQUADS
USING Q-ENHANCEMENT
Gabor CSIPKES, Doris LUPEA
Abstract: This paper presents a new approach to high frequency, narrow bandwidth
band-pass filter design. The proposed method has been developed starting from the
classical synthesis method with biquads and can be used to design filters for new
mobile transceiver architectures.

NITRIDE AND POLYSILICON ETCHING IN PARALLEL PLATE PLASMA
Florin BABARADA, Cornelia DUNARE, Marcel PROFIRESCU, Claudiu AMZA, Eugen LAKATOS
Abstract: Nitride and Polysilicon are usually layers for microelectronic fabrication
technology. Dray etching of nitride and polysilicon films is very important for
integrated circuits fabrication and now also for Mieroelectromechanicals Systems
(MEMS) [1]. The increase of the etching uniformity is very strong required by MEMS
because the dimensional configuration and precision of mechanical structures has a
direct influence on the mechanical properties. It was found that the uniformity
could be improved by decreasing the power and pressure or by increasing the flow
rate.

-----------------------------------------------------------------------------------

Number 1, 2004

DEVELOPING A CONTROL SYSTEM FOR AN ELEVATOR USING XC 4000XL FAMILY OF XILINX FPGA
Sorin Hintea, Albert Fazakas, Mihaela Cirlugea
Abstract: This paper describes a design procedure of a two chamber elevator control
system using FPGA circuits. The control system memorizes and controls calls received
regarding the current speed and actual position of the chambers, sensor's operation,
damage and alarm, and possibilities of technical support to the supervisor in the
hot points of the system function. The modular structure of the design and software
solutions of the XILINX FOUNDATION package allow changes in controller functions
with small corrections in hardware.

MUSICAL INSTRUMENT CLASS IDENTIFICATION USING CEPSTRAL COEFFICIENTS DERIVED FROM A 
CONSTANT Q TRANSFORM
Annamaria MESAROS, Eugen LUPU
Abstract: Musical signals analysis and synthesis has found its place within digital
signal processing in the last decades. A method of musical instrument class
identification by analyzing the musical signal is presented in this paper. The
instrument class model is based on the cepstral coefficients derived from the
coefficients of a constant Q transform. Usually in speech/speaker recognition
applications, the cepstral coefficients proved to be a robust set of features. Based
on the cepstral coefficients values, more identification methods were developed: one
employs a calculation of a distance measure and two implementations based on neural
networks. The results are satisfactory, especially in case of neural networks.

AVERAGED MODELING OF CONVERTERS OPERATING IN CONTINUOUS AND DISCONTINUOUS CONDUCTION
MOD - REDUCED ORDER MODEL
URSARU OVIDIU, DIORDIEV ALEXANDRU, AGHION CRISTIAN, TIGAERU LIVIU
Abstract: The difference between the converters de-dc, which operate in continuous
and discontinuous modes is that the latter has three functioning stages for a cycle
of switching - switch ON, switch OFF and stage of zero current. Moreover, there are
new restrictions to the functioning of the circuit, namely the momentary value of
the current through the inductance where the energy accumulated begins to grow above
zero at the beginning of the commutation cycle and reaches zero before the
commutation cycle ends.

OPERATIONAL TRANSCONDUCTANCE AMPLIFIER FOR FILTERING IN HIGH FREQUENCY IF SAMPLING
RECEIVERS
Gabor CSIPKES, Doris CSIPKES
Abstract: This paper presents an operational transconductance amplifier that is
suitable for high frequency band-pass filters design. The improvement is done using
a voltage mode-current mode combined approach that allows an extension of the
variation range far the transconductance parameters compared to classical OTA-s
while a relatively low consumption is maintained. The OTA structure is based on a
combination between two classical circuits: a second generation current conveyor and
the cascode output stage.

AN EXTERNAL SOFTWARE FOR CONTROLLING THE PSPICE SIMULATOR
Liviu NEDELEA. Albert FAZAKAS, Lelia FESTILA
Abstract: This paper presents a method to externally control the PSpice simulator
8.x - 9.x. implemented in a software application running on windows platforms. It
also contains a working demonstration an an analog low pass Elliptic filter. The
obtained data are processed on an external spreadsheet calculation in Microsoft
Excel.

IMAGE REJECT RECEIVER ARCHITECTURES FOR. WIRELESS COMMUNICATIONS - PERFORMANCE ANALYSIS
Doris CSIPKES, Gabor CSIPKES
Abstract: This paper presents a detailed analysis of some image reject architectures
that are frequently employed in mobile communication receivers. The behaviour of the
described architectures is characterized using the parameter called Image Rejection
Ratio (IRR). The spectrum of the wanted and unwanted signal components in different
intermediate nodes of the chain is also presented. The calculations have been
verified using Matlab.

THE INFLUENCE OF FEATURES COEFFICIENTS SEQUENCES IN SPEAKER RECOGNITION
Petre G. POP, Eugen LUPU
Abstract: Speaker recognition requires speaker characteristic features, independent
of the particular spoken word if possible. Features based on short-term spectral
estimate exhibits a strong dependence on individual speakers and consequently are
frequently used in speaker recognition. However, these features also contain
information about the lexical content of the utterance. In this paper we present the
results of a study about the influence of eliminating individual or sequences of
coefficients from speech features (LPC cepstrum, MFCC) on speaker recognition using
VQ.

PERFORMANCE ANALYSIS OF THE WCDMA DOWNLINK CHANNEL
Martin SZEKELY, Tudor PALADE, Emanuel PUSCHITA
Abstract: The paper presents the experimental evaluation of the WCDMA downlink
channel. With that end in view, it was built a simulator to investigate the
performances of DPCH (Downlink Physical Channel). In the first part of the paper, we
have a brief description of the simulator and the parameters involved. The second
part of the paper is dedicated to performance analysis: BER vs. EbNO and BER vs.
Number of Users.

LOW-COMPLEXITY SYSTEM FOR SPEAKER RECOGNITION
Eugen LUPU, Petre G. POP, Ervin KRAINICS
Abstract. The speaker recognition has many applications in telecommunications,
"on-site" and forensic fields. The usual approaches require important computational
resources generally necessitating complex signal processing in the frequency domain.
This work proposes a low-complexity implementation of TESPAR coding method, on an
ATMEL microcontrotler system. The computational requirements for this method are two
orders of magnitude less than that required by other usual methods.

DIELECTRIC PROPERTIES ANALYSIS OF A NEMATIC LIQUID CRYSTAL AT MICROWAVE FREQUENCIES
Daniela Ionescu, Nicolae Dumitru Alexandru
A nematic liquid crystal was tested in order to be used for electronic displays. A
study was carried out using microwave simulations methods, in frequency range
1-10GHz, using specific simulation program, HFSS (Ansoft Technologies). The aim of
this paper is to find the crystal birefringence, using testing methods in electric
fields orientated parallel or transverse to the nematic unit vectror of the crystal.

IMPLEMENTATION OF A MEDICAL MULTIMEDIA DATABASE CONSIDERING ORACLE AND J2EE PLATFORM
Mircea-Florin VAIDA, Jozsef DOMOKOS
Abstract: Oracle9i Application Server is a scalable, secure, middle-tier application
server. It enables delivering Web content; host Web applications, connected to
back-office applications, and access data on wireless devices. Using Oracle9i
facilitieswe developed a dedicated application for medical image and multimedia
content storage and retrieval considering the J2EE platform and the Oracle9i
Database Server and Application Server. The application stores multimedia content
such as images, audio and video data, in an Oracle9i database and the access to the
content is guaranteed by an Oracle9i server application. The clients logs into
system from a LAN (or Internet) having different types of security roles on the
content. The application interface is presented to its users through a Web site or a
file transfer client and a customer interacts with the application using a Web
browser or a client application for file transfer. The application will allow
information querying and searching, user management and maintenance.

ON USING SIMULATED DATA FOR SPEECH RECOGNITION OVER TELEPHONE CHANNELS
Dragos BURILEANU, Cristian NEGRESCU, Victor CROITORU, Mihai SIMA
Abstract: Robust speech recognition is nowadays an active investigation field for
the speech research community: Increasing the accuracy of speech recognition over
telephone lines has become a necessity, as there is a clear increasing demand for
more spoken language processing applications involving telephone speech. This paper
advocates the use of simulated data during the development of robust speech
recognition systems and presents a rigorous method to model and simulate a real
telephone channel. Experiments using simulated telephone data for a
speaker-independent isolated-spoken word recognition system are described. System's
recognition performance in different training and testing conditions are also
discussed.

SOME ASPECTS OF TESTING PROCESS FOR TRANSPORT STREAMS IN DIGITAL VIDEO BROADCASTING
Radu ARSINTE, Ciprian ILIOAEI
Abstract: This paper presents some aspects related to the DVB (Digital Video
Broadcasting) investigation. The basic aspects of DVB are presented, with an
emphasis on DVB-T version of standard. The main purpose of this research is to
analyze the way that the transmission of the transport streams is realized in case
of the Terrestrial Digital Video Broadcasting (DVB-T). To accomplish this, first,
Digital Video Broadcasting standard is presented, and then the main aspects of DVB
testing and analysis of the transport streams are investigated. The paper presents
also the results obtained using two programs designed for DVB analysis: Mosalina and
TSA.

DISTRIBUTED IMAGE PROCESSING APPLICATION CONSIDERING CORBA AND XML TECHNOLOGY
Silviu Titus MOLDOVAN, Mircea-Florin VAIDA, Gavrila TODEREAN
Abstract: This paper will present a distributed image processing application that
has the objective to create an efficient and powerful instrument for image analysis
in varied domains like medical applications, industrial applications and others. For
the implementation we have used the CORBA technology and as programming language we
have used C++ and Python. The information (texts, images, sounds) is stored in XML
files.

TRAINING MATERIALS AND THE PEDAGOGICAL MODEL IMPLEMENTED ON THE "CONCORDE" WEB
PLATFORM FOR OPEN AND DISTANCE EDUCATION
Mircea GIURGIU, Nikos BOGONIKOLOS, Luis LIZAMA
Abstract: The paper presents some of the final results obtained in the frame of the
EC funded Leonardo da Vinci pilot project with the title "Acquisition of
Complementary Competencies through Open and Distance Education" (CONCORDE) and it
complements other dissemination papers. Mainly, we focus here on the technical
implementation of an innovative pedagogical model, based on spiral evolution and on
the final training materials available on the web site, CD or printed form. These
materials respond to the training needs identified in the first stage of the project
and they are aimed for competence development through Open and Distance Education,
particularly web-based training, in the following areas: information technologies at
basic level, Internet programming, plastic materials and business administration. A
web platform developed at Technical University of Cluj supports on-line training
material organisation via web interfaces and open access to training resources
through personalized interfaces. Evaluation tools, additional training resources and
communication facilities, such as forum, are an integral part of this electronic
platform. Some of these final results are outlined and disseminated in this
communication.

ON-LINE RESOURCES AND TRAINING MATERIALS IN ALFA VERSION OF OCTOPUS PROJECT
Mircea GIURGIU, Vito CARIOCA
Abstract: This informative article aims to present the current stage of the
Socrates/Minerva project: "Transnational on-line resource centre" (OCTOPUS)
dedicated to the development of a web resource centre in the specific field of
environmental education and to introduce the new activities that will take place in
the near future. After a one year of activity, an evaluation instrument of the alfa
version of the web prototype has been developed and applied at the end of year 2002
in order to investigate and to conceptualise technical and pedagogical requirements
for the next phase dedicated to the implementation of the beta version. The
heuristic evaluation model and the epistemological assertions of the method are
briefly presented. Technologically, some relevant implementation issues will be
demonstrated in this paper. With these, requirements for the beta version have been
established and they are under development in order to allow expert evaluation for
the mid year 2003. Meanwhile, resources uptoading is a continuous process and the 
designing of self learning materials has passed the conceptualisation phase.

------------------------------------------------------------------------------------

Number 2, 2004

FORECASTING ANN MODELS BASED ON TIME SERIES ANALYSIS USING NON-UNIFORM DISTRIBUTION
OF TRAINING SAMPLE NUMBER
Denis E. ARTEMKIN
Abstract: This article presents some ways which allow to raise efficiency of
forecasting artificial neural networks (ANN) functioning. These ways are based on
exponential and so-called adaptive distributions of training sample number. The
article illustrates with a practical.example that using proposed ways allows to
raise ANN efficiency expressed as less mean square error of forecast and greater
number of correctly predicted time series values, in comparition with both classical
methods of ANN training and methods of math statistics, such as auto-regression
model in the given example.

IMPLEMENTING THE DISCRETE COSINE TRANSFORM AS A NEW COMPONENT OF IMAQ VISION LIBRARY
Bogdan ORZA, Mircea GIVAN, A. VLAICU
Abstract: IMAQ Vision is a library of functions used to implement digital image
processing applications in LabView. The paper presents a modality of implementing
new image processing algorithms that are not included in the basic package of IMAQ
Vision. Taking into account that the image compression field is less represented and
developed within IMAQ, the implementation of the discrete cosine transform - one of
the most used transforms in the image compression - was considered useful.

IMPLEMENTING A MEDIA RELAY SCHEME FOR REAL TIME STREAMING AT THE APPLICATION LAYER.
Ch. Z. Patrikakis, N. Minogiannis, Y. Despotopoulos, P. Fafali
Abstract: In this paper, the description of an open architecture for supporting real
time media streaming is presented. The architecture is based on media relay nodes
that can be deployed transparently to any existing media distribution scheme, which
can support any type of media streamed using the RTP and RTSP protocols. The
architecture is based on overlay networks and is deployed at the application level.
In parallel, a prototype of the relay node that has been developed and deployed in
desktop PC, laptop and PDA is presented, together with performance tests that have
been conducted.

TREE CONTENT BASED SEARCH ALGORITHM FOR AN IMAGE DATABASES 
M. MOCOFAN, C. CALEANU, V. MARANESCU
Abstract: For the indexing of an image data bases can be used many features. In this
paper are used features like histograms for the colour components, stohastic
moments, few parameters of the co-occurrences matrix and the Gabor filters
decomposition. Using all of them we obtain good performance in the field of
content-based image indexing and retrieval. There are some speed problems because
are many features used in the search process. Here, I present an algorithm for
indexing and retrieval, which use a tree algorithm for speed increasing. The area of
applications is very wide: multimedia documents, transaction systems, medical
application, query and browsing.

A NEW METHOD FOR ESTABLISHING THE TRANSFER FUNCTION OF OPTIMUM FILTERS
Valeriu MUNTEANU, Daniela TARNICERIU
Abstract: A new method to compute the transfer function for an optimum filter, both
for discrete and continuous domain, is proposed. The optimum filter characterized by
the transfer function obtained based on the proposed method assures the minimum mean
square error and its performances are the same as those of filters derived on the
base of classical methods developed by Wiener and Kalman.

MAXIMUM RANGE PLANNING FOR 2.4 GHZ WLAN NETWORKS
Nicolae CRISAN
Abstract: The wireless networks have had a huge impact over the communications data
systems in the beginning of 2000's Due to this fact it is more and more important to
look closely, the radio planning approach, in order to take maximum advantages over
the radio channels and the terrain. The article is focused on a very well known
technique, which shows the way in planning radio links, using a simplified
mathematical method. This method is based on the measurements we made in our
laboratory, using a radio system consist of one radio access point and more mobile
clients, working on Cluj-Napoca city.

IMAGE ENHANCEMENT METHODS USING HIGH ORDER STATISTICS
Vasile PATRASCU
Abstract: This paper presents an image enhancement method that uses piecewise linear
transforms. The determination of this transform is done using a new measure for the
mean dynamic range. Firstly the method is defined for monochrome images and secondly
is extended for the color ones. Also it is presented the way for using of this
method in the context created by the image support fuzzification. Using of the
support fuzzification leads to the mean dynamic range calculus for fuzzy sets.

BENCHMARKING FEEDFORWARD NEURAL NETWORKS TRAINING ALGORITHMS
Catalin-Daniel CALEANU, Mugur MOCOFAN, Valentin MARANESCU 
Abstract: The aim of this paper is to provide some guidance in choosing among
different types of feedforward neural networks training algorithms. More
specifically, experimental results are presented, using gradient descendent,
conjugate gradient, Newton and quasi-Newton training type methods, when neural
network topology is varied as the total number of weights and biases is ranging form
N x 10 to N x 106.

-------------------------------------------------------------------------------------

Number 1, 2005

A STATISTICAL APPROACH FOR EFFICIENT WEB DATA MINING
Nikolaos PAPADAKIS, Dimitrios SKOUTAS, Konstantinos RAFTOPOULOS, Theodora 
VARVARIGOU
Abstract: We present a fully automated system (wrapper) for extracting information from 
semistructured web pages. The emerging need for such systems occurs due to the need for 
going beyond the concept of "human browsing" by automating the process of information 
retrieval enabling further utilization by targeted applications. The key idea in our novel 
system is to exploit the format of the information contained in the web pages discovering the 
underlying structure and finally map it to semantic relationships. In doing this we identify one 
section of the web page as the one containing the useful information and we proceed in 
extracting semantic tokens contained in this section by using clustering techniques and other 
tools of statistical origin. Our innovation consists in building a system that can operate 
without human intervention or training and yet achieving excellent extraction precision and 
recall.

MULTIMEDIA PRESENTATIONS GENERATOR FOR STREAMING AUDIO-VIDEO 
CONTENT OVER INTERNET
Ovidiu BUZA, Kalman PUSZTAI, Mircea VAIDA
Abstract: A new technique has been developed in the last decade in the field of multimedia 
producing. It involves data streaming technology that is capable to provide audio-video high 
rate encoding and real-time deploying of multimedia content over a network. In this paper we 
present main concepts of data streaming and also the VPG application we have developed. 
VPG application is a graphical interactive platform that allows development of data streaming 
multimedia presentations that can be watched live through Internet. The difference between 
VPG and other international products of this category consists of modularity and efficiency in 
using multiple interactive capabilities.

CORNER FREQUENCY PROGRAMMING STRATEGY FOR LOW PASS FILTERS IN 
SOFTWARE DEFINED RADIO
RUS Cristian Matei, HINTEA Sorin
Abstract: A new method to change the comer frequency of a low pass filter is proposed. 
Traditionally, arrays of resistors and capacitors are used to select the desired frequency; the 
proposed method deals with the concept of "virtual scaling" a capacitor to achieve the change 
of the corner frequency. Thus, the need for resistor and capacitor arrays is eliminated and the 
IC size of the filter is minimized.

BEHAVIOURAL BLOCKS FOR A MORE PRECISE CALCULUS OF  PSPICE FUNCTIONS
Mihaela CARLUGEA
Abstract: I created some specialized behavioral blocks that compute fast and accurately the 
power, the rms and average values of  a signal. These values can be obtained in regular 
simulations with functions using specific PSpice techniques (products, filters) and thus the 
results have significant errors or a long simulation time. This measuring method overcomes 
this problem. The signals can be seen through an adjustable "examination window". This 
offer a better precision than the one using the PSpice method. In the paper there are examples 
that prove this.

APPROXIMATION OF THE FREQUENCY CHARATERISTIC OF STANDARDIZED 
VOCAL TELEPHONE CHANNEL USING DIGITAL FILTERS WITH FINITE IMPULSE 
RESPONSE
Zsolt Alfred POLGAR, Vasile BOTA, Mihaly VARGA
Abstract: The implementation of a vocal telephone channel simulator demands, among other 
processing, the simulation of the frequency characteristics of standardized telephone channels. 
A possible approach is the employment of digital filters with finite impulse response (FIR - 
filters), which must have imposed amplitude and phase characteristics. The paper shows that 
the direct approach to the design of these filters can't be used and proposes an algorithm based 
on the method of vector space projections (MVSP), by applying this method to imposed 
characteristics defined in a finite number of points. This algorithm ensures very good results 
for all the telephone channels standardized by ITU- T V.56 Recommendation.

RICH CLIENT CONCEPT IN INTERACTIVE LEARNING
Aurel VLAICU, Cosmin PORUMB, Sanda PORUMB
Abstract: This article consists of a set of methodologies used for designing interactive 
learning solutions. The approach is dedicated not only to the development process but it 
proposes an open architecture, Client-Side Manager for Interactive Learning, and an entire 
API, CSMIL API, that can be extended within the development process. It also highlights an 
interesting strategy to Integrate the rich client concept with the multimedia servers for 
improving the educational capabilities and transforming a classic eLearning platform in a 
powerful interactive learning system. CSMIL architecture provides an innovative presentation 
of data in a multimedia format, 2D graphics and speech-based user interaction capabilities are 
described during the article.

GENOMIC SIGNAL PROCESSING (I)
Petre G. POP, Victor I. POP, Eugen LUPU 
Abstract: The purpose of this paper is to present the biological background used in genomic 
signal processing, then to survey some of these techniques and how they are currently being 
applied in the rapidly expanding area of genomic signal processing.

BROADBAND SERVICES FOR EVERYONE OVER FIXED WIRELESS ACCESS NETWORKS
Tudor Palade, Nicolae CRISAN, Emanuel Puschita, Ligia CHIRA
Abstract: Availability of broadband access to everybody has become a high priority nowadays 
and the 3G or the upcoming 4G growth comes to approve this assumption. Necessity to 
develop the information society expressed in documents like e-Europe plan it is already a 
European Union political goal.

----------------------------------------------------------------------------------------

Number 2, 2005

ADAPTIVE RADIO TECHNIQUES AT 5 GHz
Ligia CHIRA, Tudor PALADE
Abstract: This paper discusses the necessity and advantages of using adaptive radio 
techniques in Current 5 GHz wireless communication systems. Specific propagation issues 
are addressed here and an overview of adaptive techniques is presented with a focus on 
multiple-antenna systems. Most recent and near future wireless standards are presented from a 
comparative perspective when analyzing when and where adaptive systems should be used. 
The paper presents the results, interpretations, and conclusions of our 5GHz channel 
simulations on an 802.11 a PHY scheme.

GENOMIC SIGNAL PROCESSING (II). FOURIER ANALYSIS, A SHORT REVIEW
Petre G. POP, Eugen LUPU
Abstract: The first part of this paper presented biological background and some of signal 
processing techniques and how they are currently being applied in the rapidly expanding area 
of genomic signal processing. This second part focuses on spectral analysis of biological 
signals. We considered the indicator sequence approach, digital filters used for exon 
prediction and DNA spectrograms.

IMPLEMENTING A TEST STRATEGY FOR AN ADVANCED VIDEO ACQUISITION 
AND PROCESSING ARCHITECTURE
Radu ARSINTE
Abstract: This paper presents some aspects related to test process of an advanced, non-
standard architecture, video board, used in the IP surveillance system. The system is based on 
a Pentium compatible architecture using the industrial standard PCI04+. First the overall 
architecture of the system is presented, involving both hardware or software aspects. The 
acquisition board that is developed in a special, nonstandard architecture is also briefly 
presented. The main purpose of this research was to set a coherent set of procedures in order 
to test all the aspects of the video acquisition board. To accomplish this, it was necessary to 
set-up a procedure in two steps: stand alone video board test (functional test) and an in-
system test procedure verifying the compatibility with both OS: Linux and Windows. The 
paper presents also results obtained using this procedure.

ADJUSTABLE FROM ZERO VOLTAGE REGULATOR USING ONLY ONE 723 INTEGRATED CIRCUIT
Gabriel OLTEAN, Emilia SIPOS
Abstract: A new circuit configuration to build an adjustable from zero voltage regulator with 
723IC is proposed. Our circuit has the advantage of using only one 723IC (not two as for 
existing solution). Also then is not necessary to use an auxiliary negative voltage source. The 
final circuit configuration is deduced in three steps. We start from an initial configuration and 
using SPICE simulation we improved the circuit to the final configuration. The resulting 
output voltage is adjustable in the range [0V; VREF].
In order to check and validate our circuit we compare its regulation performances (line and 
load regulation and temperature coefficient) with the regulation performances of the existing 
solution with two 723ICs.

REVERBERATION ALGORITHMS
Norbert TOMA, Marina Dana TOPA, Erwin SZOPOS
Abstract: Artificial reverberation is used in different domains from cinematography to 
psychology for enhancement of room's acoustical properties. The paper starts with the 
presentation of the reverberation phenomenon followed by a description of some early and 
late reverberation algorithms. We describe the structure of some classical reverberation 
algorithms, where no frequency dependence for the reverberation time can be defined and no 
precise rules for design can be set up. Then some improved reverberators as well as their 
design and performance are presented. Several reverberation algorithms can be found in the 
literature, but very few design hints are given. Our paper includes a detailed design of the 
improved reverberators. The simulations, which were carried out, prove the better quality of 
the designed reverberators.

REPRESENTATION AND COMPARISON OF BJT AC PARAMETERS USING PSPICE
Albert A. FAZAKAS 
Abstract: Bipolar junction transistor AC parameter representation can be simply obtained 
from the OrCAD Model editor, Parts. However, the represented characteristics do not offer 
data navigation, nor a possibility to compare two different BJT models. In this paper, 
alternative BJT AC parameter representation methods, using OrCAD Pspice, are presented 
The parameter representations allow both data navigation and parameter representation of 
more than one model on the same graphic, easing up model comparisons. A method for 
measuring and representing collector-current dependent transistor parameters such as the 
forward DC beta, or the transit frequency fT at a specific collector current, is also shown. 
Starting from the fT versus the collector current characteristics of a transistor model, an 
application example that concerns voltage buffer speed optimization versus the bias current is 
also presented.

A VISION-BASED FUZZY EXPERT SYSTEM FOR THE SURVEILLANCE AND DIAGNOSIS OF HYDRODAMS 
USING UNDERWATER COLOR IMAGE ANALYSIS
Mihaela GORDAN, Costin MIRON, Ioan STOIAN, Odysseas TSATOS, Apostolos 
GEORGAKIS, Ovidiu DANCEA 
Abstract: Surveillance of hydro plants represents a serious environmental problem. Limited 
access to upstream areas of hydro-dams makes their visual inspection difficult. However few 
attempts to build computer vision systems for upstream dam walls monitoring and diagnosis 
are currently reported; this artificial vision application area is on its beginning. In this paper 
we propose a novel solution towards the automation of underwater hydro-dams monitoring 
and diagnosis, based on the visual examination of color underwater images of the upstream 
dam walls acquired with the use of an underwater robot vehicle. Whereas the traditional 
8ethods require the human expert to visually examine the underwater images (8 time-
consuming and tiring task), the solution presented here reduces significantly the need for 8 
human expert examination. This goal has been achieved by developing an image analysis 
subsystem specific to underwater dam wall examination, designing a fuzzy expert subsystem 
for the upstream dam wall diagnosis and combining the two subsystems into a complete 
vision-based fuzzy expert system for a specific diagnosis task of the dam walls.

A SYSTEMATIC DESIGN OF LOG-DOMAIN FILTERS
Lelia FESTILA, R. GROZA, M. CARLUGEA, A. FAZAKAS
Abstract: A systematic and general design method for exponentially state - space filters is 
presented. We focus in this paper on applications in log - domain, by simulating RLC ladder 
prototypes. After describing the RLC prototype by a block diagram or a signal flow-graph, the 
method can be applied directly without other intermediate operators as other modular design 
methods use [2, 3, 4, 7]. The examples include not only first order nonlinear basic building 
blocks (integrators, lossy integrators) but also second order nonlinear cells (BPF, high Q 
BPF).

MILLIMETRE RADIOWAVE RAIN FADE SLOPE AND FADE DURATION STATISTICS
Mihai V. BUD, Tudor P. PALADE
Abstract: The ITALSAT propagation experiment started in 1991, allowed an extensive 
program of measurements at 18.7, 39.6 and 48.5 GHz. This paper reports some statistical 
results of the second order millimeter radiowave rain fade slope and fade duration parameters. 
These parameters, are necessary in order to develop propagation impairment mitigation 
techniques. Statistics of second order parameters, fade slope and fade duration, are presented. 
The measurements available were carried out at Spino d'Adda, in the northern Italy, for a 
period of 7 years (from 1993 to 2000).

---------------------------------------------------------------------------------------------

Number 1, 2006

A SOFTWARE APPROACH TO THE SECURED DISTRIBUTION OF MULTIMEDIA 
DATA (VIDEO ON DEMAND) 
Cosmin STRILETCHI 
Abstract: The aim of the paper is to present the security elements involved in the distribution 
process of multimedia content over the Internet. The software application implements a 
secured solution for a Video on Demand system destined for on-line watching and for home 
distribution of video data. The multimedia content for on-line watching is displayed at the 
client side using web pages embedded ActiveX objects that encapsulate the Windows Media 
Player and the Real Player movie controllers. The data meant for downloading is encoded 
and contains additional information about the expiration date. A downloaded video material 
can be viewed at home for a specified amount of time using a dedicated movie player offered 
online, capable of displaying the encoded data.

A NEW CONTROL AND HARMONICS ELIMINATION METHODS FOR A STATIC VAR COMPENSATOR 
USING A THREE LEVEL (N.P.C) INVERTER 
Mustapha BENGHANEM, Azeddine DRAOU (IEEE Senior member)
Abstract: In this paper the use of harmonics elimination method applied to a three level 
inverter is reported. The method used to calculate the switching angles is clearly shown and 
discussed. Simulations results using Pspice program and experimental results are carried out 
to validate the mathematical model. Finally, we applied this harmonics elimination method to 
control an advanced static VAR compensator (ASVC) which uses a three level voltage source 
inverter.
In this paper a neutral point clamped inverter having harmonics contents output voltage far 
less than those of two-level inverter at the same switching frequency is introduced. We also 
present the modelling and analysis of this new type of inverter used for advanced static var 
compensation. This ASVC uses three-level voltage source inverter (VSI) transforming a DC 
component to AC through a set of capacitors, which are used, as a power storage device. 
Furthermore, a simplified mathematical model of the ASVC is derived, and various 
simulation results presented using MATLAB.

A FRAMEWORK FOR DESIGNING CASCADED ANALOG FILTERS TAILORED TO MULTIPLE REQUIREMENTS
Liviu Nedelea, Marina Topa, Marius Neag, Lelia Festila
Abstract: The paper presents a computer-aided method and software for designing analog 
active filters considering an extended set of requirements. It covers all design steps, from the 
generation of the transfer function to filter synthesis, circuit implementation and verification 
through simulation. The generation of the transfer function starts with a choice of classical 
approximations, which can be optimized in respect to an extended set of behavioral and 
implementation criteria. The filter synthesis is based on the cascaded of first- and second-
stages approach; an extensive library of such filter stages offers the designer a wide range of 
circuit implementation options. The circuit element values are calculated and the complete, 
Spice-compatible, netlist of the filter is generated. Finally, simulations can be run within the 
proposed framework. The CAD tool proposed here is an effective, powerful and user-friendly 
tool for optimized design of complex analog filters, tailored to very specific requirements.

LDPC CODES FOR HIGH BIT-RATE MOBILE COMMUNICATIONS
Zsolt Alfred POLGAR, Florin ARDELEAN, Vasile BOTA, Mihaly VARGA
Abstract: The paper analyzes the HER vs. SINR performances of three classes of LDPC 
codes, namely the L(m,q) codes, the Kirkman codes and codes generated with a geometrical 
construction. It presents briefly the construction methods of the control matrices H and 
emphasizes the code parameters (girth, girth distribution, order of the bit and check nodes in 
the Tanner graph, code word length) that affect significantly the performances of a LDPC 
code decoded with the Sum-Product algorithm. The throughput vs. SINR performances of 
some selected codes are evaluated in an access scheme proposed for the downlink connection 
of the 4G systems. The throughputs provided by the LDPC-coded adaptive modulation 
scheme are significantly higher than the ones provided by a non-coded adaptive modulation 
scheme in the same environment.

A NOVEL HIGH-SPEED VOLTAGE BUFFER WITH LOW DC OFFSET AND OPTIMIZED STEP RESPONSE
A. FAZAKAS, M. NEAG 
Abstract: This paper presents a novel high-speed voltage buffer and an effective method for 
optimizing the step response of this class of circuits. The buffer combines a low offset 
structure with a signal-dependent biasing technique that results in significant slew-rate 
enhancement. A set of RC networks provide an effective way of controlling the step-response 
parameters for a wide range of loading conditions. The compensation elements are sized using 
the PSpice Optimizer, directly in the time domain; a self adjusting scheme can be 
implemented, as well. A design example validates the proposed structure and the optimization 
method.

EFFECTIVE METHODS TO ANALYZE SATELLITE LINK QUALITY USING THE BUILT-IN FEATURES 
OF THE DVB-S CARD
Radu ARSINTE 
Abstract: This paper presents aspects of a simple methodology to evaluate satellite link 
quality (especially for downlink segment) using accessible technologies and Satellite PC 
boards (DVB-S). Traditional methods are based on high end testing communication 
equipment, allowing better results in precision, but with significantly higher costs, not entirely 
justified in small communications companies or educational environment. The described 
methodology uses the BER delivered by the source decoder (a standard component in DVB-S 
receivers), allowing to estimate the C/N (carrier to noise) ratio, and by consequence, of all the 
main parameters of the satellite link. A limitation is the fact that only the downlink segment 
could be accurately evaluated, this comparative estimation in most application is more than 
needed (ex. Satellite TV reception).

A PARTIALLY SEQUENTIAL-PARTIALLY PARALLEL ANALOG IMPLEMENTATION OF A SVM CLASSIFIER
Gabriel OLTEAN, Mihaela GORDAN, Sorin HINTEA
Abstract: The classification task in large data spaces can now relay on newly developed 
Support Vector Machine technique. A trade-off between the hardware complexity and 
processing time is a partially sequential- partially parallel implementation of the processing 
issues based on a description of the vectors as time-continuous signals. We present an analog 
implementation of a SVM classifier that uses an analog multiplier to compute the dot product 
between each support vector and the test vector, in a sequential manner. All the dot products 
necessary to classify a test vector are implemented in parallel. The correct operation and the 
accuracy of our analog SVM classifier is proven on the IRIS dataset.

---------------------------------------------------------------------------------------------

Number 2, 2006

DNA SECURITY BETWEEN TECHNICAL AND SPIRITUAL CONCEPTS
Mircea-Florin VAIDA, Tatiana HODOROGEA
As modern encryption algorithms are broken step by step, the world of information security 
looks in new directions to protect the data it transmits. The concept of using DNA computing 
in the fields of cryptography has been identified as a possible technology that may bring 
forward a new hope for un-breakable algorithms. We need better information security because 
the people of Earth are in a period of general moral decline. A strong connection exists 
between material and spiritual life. As the universal maxim goes, "Information is power." 
Therefore a strong link exists between material and spiritual life. In this work we consider a 
technical process for protecting medical information and other data assets using a technique of 
deriving DNA public keys from blood analysis. A DNA encryption technique is further 
developed here in which a person's medical data is encrypted in DNA strands based on the 
central dogma of molecular biology. Protection is enhanced by using the patient's own blood 
mineral levels as a seed for selecting, transmitting, and recovering that person's public key. A 
lot of spiritual concepts are in connection with the new biometrics and quantum physics 
discoveries. These concepts are able to offer a deeper mode to understand the human structure 
and also to consider possible new directions for research in the security domain.

EFFECTIVE VIDEO PROCESSING ARCHITECTURES FOR FACE RECOGNITION 
ALGORITHMS IMPLEMENTATION
Radu ARSINTE 
Abstract: The paper presents a study of autonomous face recognition systems based on high 
performance DSP, so called Media Processors and on Field Programmable Gate Array 
(FPGA) devices. An overview of the most powerful Media Processors and FPGAs available 
today is followed by an analysis of the common features of the face recognition algorithms 
used to implement an application. The paper presents also a generic system implemented 
using DSP Media processors, both from a hardware or software perspective. The presentation 
is focused on the possibility of implementation for face recognition algorithms in embedded 
systems, taking as examples databases with a limited number of faces. Results, estimations 
and benchmarks for the generic systems are also presented at the end of the paper.

VOICE BIOMETRICS
Petre G. POP, Eugen LUPU
Abstract: Speaker recognition is the process of automatically recognizing who is speaking 
based on individual information included in speech waves. This technique makes it possible 
to use the speaker's voice to verify their identity and control access to services such as voice 
dialing, banking by telephone, telephone shopping, database access services, information 
services, voice mail, security control for confidential information areas, and remote access to 
computers. This paper is a survey of speaker recognition problematic and includes speech 
signal production, speech signal analysis, speaker recognition principles, speaker recognition 
methods and performance evaluation.

COMPARING METHODS FOR EXTRACTION OF RELEVANT INFORMATION OF IRIS
Laura Kovacs, Anca Discant, Carol Rus, Sorina Demea
Abstract: Starting with images of real human eye containing information about iris and pupil, 
images taken from patience, studied a few extraction methods of useful iris information and a 
classification of these irises by color. The purpose of the paper is to analyze the methods used 
until now and try new and suitable ideas in the future. In our work it was made an 
identification of the useful information in the iris circular zone and removed the unuseful 
information about pupil and external zone of the iris. We focus the work on image analyze 
and image enhancement, by image filtering, edge detection, dilation and filling. An automatic 
detection for the center of the iris and pupil zone and a principle of iris classification by the 
color of them is made also.

FINGERPRINT BASED BIOMETRIC METHODS
Marius TICO
Abstract: Fingerprints are graphical ridge patterns present on human fingers, which, due to 
their uniqueness and permanence, are among the most reliable human characteristics that can 
be used for people identification. This paper presents a general overview of the fingerprint-
based biometric systems including: image acquisition, image enhancement, feature extraction 
and fingerprint matching. Some of the most representative algorithms proposed in the open 
literature are reviewed. A brief description of some algorithms previously proposed by the 
author is also included.

AN OVERVIEW OF BIOMETRICS
Eugen LUPU, Petre G. POP
Abstract: Biometric systems allow automatic person recognition based on physical 
(fingerprint, face shape, iris, palm, DNA) or behavioral (voice, writing style, gait) features 
which belong to a certain person and cannot be forgotten, lost, copied or stolen like 
passwords, cards, keys etc. Security biometric methods are more reliable than those based on 
passwords or PIN. An overview of the biometric systems is presented including history, 
requests, taxonomy and standardizations. Also, a brief introduction for the most important 
biometric verifiers and a future view for the multimodal biometrics are presented.

TESPAR A BIOMETRIC TIME DOMAIN APPROACH TO SPEAKER RECOGNITION
Vasile Vlad MOCA
Abstract: Voice recognition is one of the less intrusive biometric methods available today and 
one of the few that can be performed online trough existing telephone infrastructure. TESPAR 
coding (Time Encoded Signal Processing And Recognition) has been proved to be a powerful 
biometric able to extract highly informative features from a large variety of biological 
bandwidth limited signals [1,2,11,12] like voice, electroencephalogram (EEG), magneto 
encephalogram (MEG) and electrocardiogram (RCG). In respect to voice and speaker 
recognition the novelty of the TESPAR method consists in the time domain approach, unlike 
classical methods (dynamic time warping (DTW), vector quantization (VQ), hidden Markov 
models (HMM)) mostly based on frequency analysis. The introductory part presents a few 
practical applications of the TESPAR method in monitoring and classification of human 
biological signals. Further on the simple TESPAR model is presented along whit a short 
discussion on classification. The last part is focused on speaker recognition applications and 
tries to motivate why TESPAR can be regarded as a biometric method.

-----------------------------------------------------------------------------------

Number 3, 2006

FEC PERFORMANCE EVALUATION FOR 2-11 GHZ IEEE 802.16a SINGLE CARRIER SYSTEM
Nicolae CRISAN
Abstract: Additional physical layer specifications related to the IEEE 802.16a 
standard are focused on enhancing the medium access control by providing 
additional specifications in order to support broadband wireless access at 
frequencies in-between 2-11GHz. Non license 5GHz radio band, it looks very 
interesting for the beginning for the most important wireless technologies 
providers, in their effort to diminish the cost of the new wireless equipments. 
The standard 802.16 is not yet complete and more simulations are necessary, 
followed probably by new amendments to the existing ones in the very next years. 
The article is focused on the FEC block performances evaluation (Forward error 
correction block) of the 802.16a system, from bit error rate point of 
view, using System View simulator by Elanix.


GENOMIC SIGNAL PROCESSING (III): FINDING DNA APPROXIMATE TANDEM REPEATS
Petre G. POP
Abstract: Local repetitions in genomes are called tandem repeats. A tandem 
repeat contains multiple, but slightly different copies of a repeated unit. It 
changes over time as the copies are altered by mutations, when additional copies 
are created by amplification of an existing copy, or when a copy is removed by 
contraction. Theses changes let tandem repeats evolve dynamically. The detection 
of tandem repeats is important in biology and medicine as it can be used for 
phylogenic studies and disease diagnosis. Though extensive work has been done in 
order to identify tandem repeats, there are inherent limitations with the 
current approaches in terms of the number of pattern sizes that can be searched 
or the size of the input length. The purpose of this paper is to present the 
biological and technical aspects of tandem repeats with focus on signal 
processing techniques.


ADVANCED CONCEPTS FOR ELEARNING PLATFORMS
Sanda Porumb, Cosmin Porumb, Bogdan Orza, Aurel Vlaicu
Abstract: The most important advantages the information technology provides in 
the Open and Distance Learning area are dedicated not only to the Virtual 
Universities but also to the Special Education Programmes. One of the situations 
can be reached in the govern institutions where the sacked persons want to find 
a new job in other domain and they need a lot of training. This article 
describes new concepts used during the development process of the interactive 
learning service based on a set of applications designed to provide an 
infrastructure for webcasting, web conferencing capabilities, synchronized video 
presentation and automated creation of navigable and searchable multimedia 
archives.

INTEGRATED ENVIRONMENT FOR DISTANT MANAGEMENT OF TUMORS
Bogdan ORZA, Aurel VLAICU, Laura GRINDEI
Abstract: TeleOralTum software is intended to collect on one server specific 
data from departments that work on the facial cancer diagnosis. Classical 
medical services imply the existence of a direct link between medical staff in 
different departments and the patient. Telemedicine is an old concept, which 
arise debates for more then 30 years. The application that we developed proves 
the need for a diagnose and management system that will allow a rapid gathering 
of data, a rapid elaboration of the diagnose, but also elaboration of reports 
and estimates that are so much needed in this field of medical research. The 
application works on a three layer distributed architecture, thus taking 
advantage of a high security for the patient data, easiness in splitting the 
work among software development teams, and also the easiness in which other 
medical departments can be added to the application. The user administration 
section is used to divide the accessibility domain for different types of client 
users. The architecture of TeleOralTum is an innovative one and has at its 
origin open-software tools.


DESIGNING ON LINE EVALUATION TESTS FOR eLEARNING
Laura V. GRINDEI    Aurel VLAICU    Bogdan ORZA
Abstract: On line tests, along with other evaluation activities, offer students 
an opportunity to apply the concepts, skills, and attitudes they have learned 
within eLearning programs. Therefore, well designed on line tests provide a 
reliable way to measure progress objectively. This paper present general 
strategies that can enhance professors/ instructors ability to design on-line 
tests that are effective in motivating, measuring, and reinforcing eLearning.

-----------------------------------------------------------------------------

Number 4, 2006


MATLAB MODELING AND FPGA IMPLEMENTATION OF NEURONAL ALGORITHMS FOR BLIND AUDIO 
SIGNAL SEPARATION
B. S. KIREI   A.FAZAKAS    Marina TOPA
Abstract: A number of applications in a variety of areas comprise the task of 
obtaining certain signals called sources, which are not directly accessible, but 
have to be extracted from a set of measurable signals regarded as mixtures of 
sources. As neither the source signals nor the mixing structure are known, this 
is referred to as the blind source separation problem. This paper deals with 
the FPGA implementation of neural networks controlled by adaptive learning 
algorithms for blind source separation. The register transfer level model of the 
neural network is achieved in Simulink, the FPGA implementation is done using 
advanced design techniques as MicroBlaze microcontroller and creation of custom 
peripherals. To ease the test of the algorithms a graphical user interface in 
the MATLAB environment has been developed. 


MODELING THE RAYLEIGH-FADED MOBILE RADIO CHANNEL
Zsolt Alfred POLGAR, Vasile BOTA, Mihaly VARGA
Abstract: The paper proposes two joint mathematical characterizations of the 
mobile multipath Rayleigh-faded radio channels and of two possible OFDMA-type 
multi-user access techniques appropriate for high bit rate mobile transmissions. 
These mathematical models are used to compute the probability density functions 
(p.d.f.) of the received SNR and its state probabilities and to build Markov-
chain models that describe jointly the channel and the multi-user access 
technique. The paper also discusses the effects of the number of active users 
and their speed upon the p.d.f. of the received SNR and upon the state and state 
transition probabilities for the access methods considered.


ECHOGRAPHIC MEASUREMENTS USING AN IMAGE PROCESSING APPLICATION
Mircea-Florin VAIDA, Valeriu TODICA
Abstract- The aim of the paper is to present a flexible dedicated application, 
Healthlmag, able to integrate different medical facilities. Representative 
quantitative parameters are used for echographic measurements. Important 
facilities concerning the preprocessing, specific quantitative parameter 
measurements, data visualization are considered for medical investigations.


PERFORMANCE EVALUATION OF H-ARQ ADAPTIVE CODED QAM TRANSMISSIONS OVER MULTIPATH 
MOBILE CHANNELS
Vasile BOTA, Zsolt Alfred POLGAR, Mihaly VARGA
Abstract: This paper presents the evaluation of the average spectral 
efficiencies provided by the adaptive employment of a set of LDPC-coded QAM 
modulations in an OFDMA downlink scheme governed by an H-ARQ protocol over 
mobile radio channels. It discusses the selection of the set of coded 
modulations, the setting of the SNR domains where they should be employed and 
shows the computation of the average spectral efficiency provided by this 
approach in non-ARQ or H-ARQ applications. The paper also includes some 
considerations about the joint modeling of the channel and user-access method 
employed. Finally, some conclusions about the influences of various system 
parameters upon the average spectral efficiency and error rates are also 
presented.


THE MANAGEMENT INFRASTRUCTURE OF A NETWORK MEASUREMENT SYSTEM FOR QOS PARAMETERS
Alexandra BIKFALVI, Paul PATRAS, Cristian Mihai VANCEA, Virgil DOBROTA 
Abstract: The paper presents the designing principles of a management 
infrastructure for monitoring the capabilities of different network interface 
cards (Gigabit Ethernet, Endace), both for traffic generation and capturing at 
reception. This evaluation is useful to asses the goodness of captured traffic 
analysis and QoS performance measurements, using PC based platforms. The idea 
was to implement the communication of the management information between an 
administration console and a set of distributed SNMP-based software measurement 
agents for GNU/Linux platforms that enable to perform QoS measurement sessions.