|
CONTENTS
| Corrado Di Natale, Eugenio Martinelli, Arnaldo D'Amico - Pre-processing and pattern recognition methods for artificial olfaction systems: a review | 3 |
| Martin Kollár - A transducer interface for resistive sensor elements based on the use of a flip-flop | 27 |
| Bożenna Kalus-Jęcek, Zygmunt Kusmierek - New instruments to measure parameters of infra-low frequency voltage signals | 47 |
| Dariusz Borkowski - On-line instantaneous frequency estimation and voltage/current coherent resampling method | 59 |
| Zbigniew Kłos - Calibration of extremely sensitive ranges of DC picoammeters with analog active I/U conversion | 75 |
| Tomasz Bojko - Smart sensor solutions for mechanical measurements and diagnostics | 95 |
| Contents of the Volume XI/2004 M&MS | 107 |
ABSTRACTS & FULL PAPERS
|
CORRADO DI NATALE1,2, EUGENIO MARTINELLI1, ARNALDO D'AMICO1,2
1University of Rome "Tor Vergata"
Department of Electronic Engineering, Roma, Italy
2CNR-IMM, Sezione di Roma, Roma, Italy
PRE-PROCESSING AND PATTERN RECOGNITION METHODS FOR ARTIFICIAL OLFACTION SYSTEMS: A REVIEW
Data analysis is a fundamental part of artificial olfaction systems. Nonetheless, while electrical and mechanical components of electronic noses are deeply investigated and optimised for the specific applications, data analysis is sometimes performed without any critical understanding about hypotheses on which it is funded.
Several methodologies were imported into this field and utilised to correlate electronic nose data to the information required for classification purposes. Chemometrics and neural networks are the sources of the most popular methods currently in use.
In this paper some of the most used methodologies of data analysis are reviewed and their use for electronic nose data is discussed. Although the discussion is given considering chemical sensor arrays, the conclusions are generic and valid for a broader range of sensor systems.
Full paper (285kB): M&MS_2005_003.pdf
|
|
MARTIN KOLLÁR
University of Technology in Košice
Department of Theory of Electrical Engineering and Measurement
A TRANSDUCER INTERFACE FOR RESISTIVE SENSOR ELEMENTS BASED ON THE USE OF A FLIP-FLOP
This paper presents a new transducer interface. This interface serves the resistive bridges. The A/D conversion is based on the use of an auto-compensatory system with two digital-to-analog converters (DACs). A new measurement technique based on the use of a flip-flop circuit is used to obtain high accuracy. The main advantages of the depicted architecture are: a) Calibration accuracy depends only on LSB and maximal differential non-linearity (max(DNL1)) of the first digital-to-analog converter (DAC). b) Resultant accuracy of measurement depends on calibration accuracy and on accuracy of the second DAC. c) Calculation of a correction formula (or look up table) and input amplifier are not required.
The experimental measurement circuit with flip-flop was constructed and simulated to verify operation of the measurement.
Keywords: Wheatstone bridge, transducer interface, self-calibration, flip-flop circuit, measurement
Full paper (296kB): M&MS_2005_027.pdf
|
|
BOŻENNA KALUS-JĘCEK, ZYGMUNT KU¦MIEREK
Technical University of ŁódĽ
Institute of Theoretical Electrotechnics, Metrology and Material Science
NEW INSTRUMENTS TO MEASURE PARAMETERS OF INFRA-LOW FREQUENCY VOLTAGE SIGNALS
The paper presents the design, the principles of operation and the technical specification of two types of infra-low frequency voltage instruments: the integrating converter instrument and the microprocessor instrument. Both instruments are designed for the following frequency range: 0.01 - 10 Hz. The integrating converter instrument measures the mean modulus value and the rms value of voltage. The microprocessor instrument measures the mean modulus value, the rms value, the shape coefficient and the frequency of the measured signal at the same time. The measurement takes from one to four periods of the measured signal.
Keywords: infra-low frequency, rms value, mean modulus value
Full paper (194kB): M&MS_2005_047.pdf
|
|
DARIUSZ BORKOWSKI
AGH University of Science and Technology
ON-LINE INSTANTANEOUS FREQUENCY ESTIMATION AND VOLTAGE/CURRENT COHERENT RESAMPLING METHOD
An innovatory on-line method of power system frequency estimation and voltage/current signals resampling is presented. The method is designed to increase the accuracy of power system equivalent harmonic impedance estimation using simple DAQ systems. Estimation of instantaneous frequency uses 5-th degree polynomial interpolation of filtered signal zero crossings time positions. Final resampling uses cubic splines interpolation for the calculation of output signal values. The details of resampling algorithm are described.
The results of simulation tests demonstrating the typical behavior of a power system were shown as well as results of real voltage signal resampling. The method shows good capabilities of tracking varying harmonic phase angles in reference to the fundamental one while phase/frequency of fundamental harmonic varies too. The computational complexity allows to implement the method as a realtime version on DSP. Comparison of the amount of math operations used in this and another method [1] is shown too.
Keywords: frequency estimation, resampling, interpolation, coherent sampling, power system
Full paper (513kB): M&MS_2005_059.pdf
|
|
ZBIGNIEW KŁOS
Wrocław University of Technology
Institute of Electrical Machines, Drives and Metrology
CALIBRATION OF EXTREMELY SENSITIVE RANGES OF DC PICOAMMETERS WITH ANALOG ACTIVE I/U CONVERSION
The paper deals with the problem of calibration of extremely sensitive picoammeters with measuring ranges close to limits of the measuring possibilities. The presented method of calibration bases on a resistive source of very low DC standard current, namely a current calibrator. Procedure for the design of a proper calibrator, criteria of optimal selection of high value resistance standards and a detailed analysis of the accuracy of obtained calibrating current have been described. The principle of evaluation of measurement uncertainty and method of assessment of combined uncertainty for calibrated picoammeters with analog and digital display have been introduced. It has been shown that in an averagely equipped measuring and research laboratory and using relatively simply accessible resistance standards with values of 1010 - 1011 Ω, it is possible to calibrate the most sensitive ranges of picoammeters with active i/u conversion.
Keywords: calibration, electrometric picoammeter, active i/u conversion, uncertainty of measurement
Full paper (298kB): M&MS_2005_075.pdf
|
|
TOMASZ BOJKO
AGH University of Science and Technology Cracow
Department of Robotics and Machine Dynamics
SMART SENSOR SOLUTIONS FOR MECHANICAL MEASUREMENTS AND DIAGNOSTICS
Application of many sensors located on a mechanical structure causes serious technical problems. One of the main problems is wiring of sensors which is time-consuming and may change the system output due to high structure loading and damping. To overcome this problem new smart sensor designs have been proposed. One of the most promising technologies in this field is wireless data transfer. In the paper requirements for the developed smart wireless sensor are specified.
The important part of a smart sensor is the transducer which transforms certain information from the physical world to user defined values - mostly the voltage. In the paper the application of industrial accelerometers was specified and descriptions of basic types of accelerometers are given. For smart sensor design two modern embedded accelerometers were specified and laboratory-tested.
Keywords : transducers, smart sensors, wireless networks
Full paper (183kB): M&MS_2005_095.pdf
|
|
