Kazimierz Brudzewski

Milk classification by means of an electronic tongue and Support Vector Machine neural network

The aim of this work was the classification of milk samples with the use of Support Vector Machine networks. An electronic tongue, based on a sensor array of miniaturized solid-state potentiometric electrodes, was used for measurements of milk originating from various dairies (i.e. various brands) and with different fat content. The sensors were mounted into the measurement flow-cell developed at Warsaw University of Technology. Their signals were input to the Support Vector Machine neural network without a pre-processing stage. The results of the classification of milk by trademark and by fat content proved the proposed system to be very efficient.

Recognition of Coffee Using Differential Electronic Nose

This paper studies the application of the differential electronic nose for the recognition of coffee, particularly the forgery of it, made by mixing two different quality coffee brands (the mediocre product and the high-quality coffee type), sold as the high-quality coffee brand. Since the beans are practically unrecognizable by the shape and visual inspection, the only solution to this problem is the application of the chemical analysis. The usually applied approach is the liquid chromatography. However, it is a laborious and expensive method, requiring special equipment and an experienced operator. In this paper, we propose the application of the differential electronic nose, relying its decision on the measurement of the coffee smell by the semiconductor gas sensors organized in the form of a matrix. We will show that differential electronic nose applying the special procedure of signal processing is of sufficient sensitivity for the recognition of the forgery of coffee and performs much better than the classical electronic nose (e-nose).

Gas analysis system composed of a solid-state sensor array and hybrid neural network structure

Abstract This paper presents the application of the hybrid neural network to the solution of the calibration problem of the solid state sensor array used for the gas analysis. The applied neural network is composed of two parts: the selforganizing Kohonen layer and multilayer perceptron (MLP). The role of the Kohonen layer is to perform the feature extraction of the data and MLP network fulfils the role of the estimator of the concentration of the gas components. The obtained results have shown that the array of partially selective sensors, cooperating with hybrid neural network, can be used to determine the individual analyte concentrations in a mixtures of gases with good accuracy. The hybrid network is a reasonably small net and as a result, it learns faster and reaches good generalization ability with a reasonably small sized training data set. The system has the two interesting features, i.e. lower calibration cost and good accuracy.

Neuro-fuzzy TSK network for calibration of semiconductor sensor array for gas measurements

The neuro-fuzzy network applying Takagi-Sugeno-Kang (TSK) fuzzy reasoning for the calibration of the semiconductor sensor array is developed in this paper. The structure, as well as the learning algorithm of the neuro-fuzzy network, is presented and tested on the example of estimation of the concentration of gas components in the gaseous mixture (so-called artificial nose problem). The results of numerical experiments are presented and discussed.

Differential electronic nose and support vector machine for fast recognition of tobacco

Highlights? We developed differential electronic nose using 2 sensor arrays for recognition of cigarettes.? We applied support vector machine as the final pattern recognition unit in this artificial nose. ? We performed experiments of recognition of 11 brands of cigarettes using the artificial nose. The paper presents the application of differential electronic nose and support vector machine for recognition of the cigarette brands. Differential nose combines two identical gas sensor arrays mounted into two chambers. One array measures the gas of interest and the second the environmental air. Only differential signals of these two arrays are registered and subjected to further processing. Such organization of measurement eliminates the need for determination of the baseline value of resistance, since the differential signal is free of it. This system combined with support vector machine used as the classifier is well suited for very quick and on-line dynamic mode measurements. We have applied this system to the recognition of 11 cigarette brands on the basis of smell of their leaves. The performed experiments have proved that the developed e-nose based on the differential signals is capable to recognize the cigarette smells very quickly and with high accuracy.

Neuro-fuzzy network for flavour recognition and classification

This paper presents the neuro-fuzzy Takagi-Sugeno-Kang (TSK) network for the recognition and classification of flavor. The important role in this process fulfills the self-organizing process used for the creation of the inference rules. The self-organizing neurons perform the role of clustering data into fuzzy groups with different membership values (the preprocessing stage). Applying the automatic control of clusters, we have the optimal size of the TSK network. The developed measuring system has been applied for the recognition of flavor of different brands of beer. The fuzzy neural network is used for processing signals obtained from the semiconductor sensor array. The results of numerical experiments have confirmed the excellent performance of such solutions.

An attempt to apply Elman's neural-network to the recognition of methane pulses

Abstract This paper describes an attempt to apply Elmans neural-network to the recognition of gas pulses. In the present approach a small array of semiconductor oxide sensors (TGS813 Figaro, MQK-T KenLuck Electr., and W-Capteur) is used to recognize the pulses of mixtures of air and methane.

Fuzzy self-organizing hybrid neural network for gas analysis system

The paper presents the gas analysis system applying the self-organizing fuzzy hybrid neural network. The network is composed of the self-organizing competitive fuzzy layer and the supervised multilayer perceptron (MLP) subnetwork, connected in cascade. The characteristic features of this network structure for gas analysis systems are discussed and the results of experiments compared to standard neural solutions based on MLP or classical hybrid network employing the Kohonen layer.

Optimal selection of sensors based on the catalogue data : a vector-space approach to cross-sensitivity effects

Abstract Many sensors and systems for detecting toxic gases have been developed. Electrochemical and solid-state sensors are affected by various interferences. An effective method for the selection of a set of sensors based on a vector-space approach to the cross-sensitivity problem is presented. We believe that this method of selecting gas sensors will provide a new tool for designers of monitoring systems.

Spectral response of ellipsometric sensors

Abstract Aspects of the spectral response of ellipsometric sensors (ESs) are discussed. An essential feature of the ES is the spectral selectivity, which can be tuned in the UV, Vis or near-IR spectral region. The tuning effect is achieved by a change in thickness of thin films of a reactive material deposited on the test and reference mirrors. The transmittance function of the ES calculated for polyaniline (PANI) layers as a reactive mateial for the cases of chemisorption and bulk chemical reaction is presented. Optimization of the PANI layer thickness (tuning effect of the spectral sensitivity of the ES) enables us to obtain sensors with an optimal response to the molecules to be detected.