Electronic Nose

An overview is given of near infrared (NIR) spectroscopy for use in measuring quality attributes of horticultural produce. Different spectrophotometer designs and measurement principles are compared, and novel techniques, such as time and spatially resolved spectroscopy for the estimation of light absorption and scattering properties of vegetable tissue, as well as NIR multi- and hyperspectral imaging techniques are reviewed. Special

The human nose is still the primary `instrument' used to assess the smell or flavour of various industrial products today, despite considerable and sustained attempts to develop new electronic instrumentation capable of mimicking its remarkable ability . In this paper we review the research effort that has been carried out over the past 25 years or so to create an electronic nose . In doing so, we first provide a definition for the term electronic nose, and then discuss some of the technologies that have been explored in what is essentially an intelligent chemical array sensor system. Finally, we summarize the applications of electronic noses to date and suggest where future applications may lie.

Several polymers have been used to detect nitroaromatic explosives by a variety of transduction schemes. Detection relies on both electronic and structural interactions between the sensing material and the analyte. Quenching of luminescent polymers by electron deficient nitroaromatic explosives, such as trinitrotoluene, may be monitored to detect explosives. Resistive sensing using carbon black particles that have been coated with different organic polymers and deposited across metallic leads can also be used to detect nitroaromatic vapors in an electronic nose approach. Frequency changes in surface acoustic wave devices may be monitored to detect nitroaromatics after their adsorption into polymer coatings. Luminescent polymetalloles have recently been investigated for sensing explosives in aqueous-based solutions and for improved visual detection of trace particulates on surfaces.

The development of a robust method for integrating high-performance semiconductors on flexible plastics could enable exciting avenues in fundamental research and novel applications. One area of vital relevance is chemical and biological sensing, which if implemented on biocompatible substrates, could yield breakthroughs in implantable or wearable monitoring systems. Semiconducting nanowires (and nanotubes) are particularly sensitive chemical sensors because of their high surface-to-volume ratios. Here, we present a scalable and parallel process for transferring hundreds of pre-aligned silicon nanowires onto plastic to yield highly ordered films for low-power sensor chips. The nanowires are excellent field-effect transistors, and, as sensors, exhibit parts-per-billion sensitivity to NO 2 , a hazardous pollutant. We also use SiO 2 surface chemistries to construct a 'nano-electronic nose' library, which can distinguish acetone and hexane vapours via distributed responses. The excellent sensing performance coupled with bendable plastic could open up opportunities in portable, wearable or even implantable sensors.

Electronic-nose devices have received considerable attention in the field of sensor technology during the past twenty years, largely due to the discovery of numerous applications derived from research in diverse fields of applied sciences. Recent applications of electronic nose technologies have come through advances in sensor design, material improvements, software innovations and progress in microcircuitry design and systems integration. The invention of many new e-nose sensor types and arrays, based on different detection principles and mechanisms, is closely correlated with the expansion of new applications. Electronic noses have provided a plethora of benefits to a variety of commercial industries, including the agricultural, biomedical, cosmetics, environmental, food, manufacturing, military, pharmaceutical, regulatory, and various scientific research fields. Advances have improved product attributes, uniformity, and consistency as a result of increases in quality control capabilities afforded by electronic-nose monitoring of all phases of industrial manufacturing processes. This paper is a review of the major electronicnose technologies, developed since this specialized field was born and became prominent in the mid 1980s, and a summarization of some of the more important and useful applications that have been of greatest benefit to man.

This paper reviews the range of sensors used in electronic nose (e-nose) systems to date. It outlines the operating principles and fabrication methods of each sensor type as well as the applications in which the different sensors have been utilised. It also outlines the advantages and disadvantages of each sensor for application in a cost-effective low-power handheld e-nose system.

Previous finding shown that the composition of the breath of patients with lung cancer contains information that could be used to detect the disease. These volatiles are mainly alkanes and aromatic compounds. Sensor arrays technology (electronic nose) proved to be useful to screen samples characterised by different headspace composition. Here we investigated the possibility of using an electronic nose to check whether volatile compounds present in expired air may diagnose lung cancer. Breath samples were collected and immediately analysed by an electronic nose. A total of 60 individuals were involved in the experiment. 35 of them were affected by lung cancer, 18 individuals were measured as reference and nine were measured after the surgical therapy. Two individuals were measured twice, before and after the surgical therapy, for a total of 62 measurements. An electronic nose, composed by eight quartz microbalance (QMB) gas sensors, coated with different metalloporphyrins, was used. These sensors show a good sensitivity towards those compounds previously indicated as possible lung cancer markers in breath. The application of a 'partial least squaresdiscriminant analysis' (PLS-DA) found out a 100% of classification of lung cancer affected patients, 94% of reference was correctly classified. The class of post-surgery patients were correctly individuated in 44% of the cases, while the other samples were classified as healthy references. The alteration of breath composition induced by the presence of lung cancer was enough to allow a complete identification of the sample of diseased individuals. Extended studies are necessary to evaluate the resolution of the method, namely the stage at which the disease may be identified in order to use this instrument for early diagnosis. #

Rationale: Electronic noses are successfully used in commercial applications, including detection and analysis of volatile organic compounds in the food industry. Objectives: We hypothesized that the electronic nose could identify and discriminate between lung diseases, especially bronchogenic carcinoma. Methods: In a discovery and training phase, exhaled breath of 14 individuals with bronchogenic carcinoma and 45 healthy control subjects or control subjects without cancer was analyzed. Principal components and canonic discriminant analysis of the sensor data was used to determine whether exhaled gases could discriminate between cancer and noncancer. Discrimination between classes was performed using Mahalanobis distance. Support vector machine analysis was used to create and apply a cancer prediction model prospectively in a separate group of 76 individuals, 14 with and 62 without cancer. Main Results: Principal components and canonic discriminant analysis demonstrated discrimination between samples from patients with lung cancer and those from other groups. In the validation study, the electronic nose had 71.4% sensitivity and 91.9% specificity for detecting lung cancer; positive and negative predictive values were 66.6 and 93.4%, respectively. In this population with a lung cancer prevalence of 18%, positive and negative predictive values were 66.6 and 94.5%, respectively. Conclusion: The exhaled breath of patients with lung cancer has distinct characteristics that can be identified with an electronic nose. The results provide feasibility to the concept of using the electronic nose for managing and detecting lung cancer.

WITNESSING THE SWIFT ADVANCES in the electronic means of seeing and hearing, scientists and engineers scent a market for systems mimicking the human nose. Already commercial systems from several companies are targeting applications, present and potential, that range from quality assurance of food and drugs to medical diagnosis, environmental monitoring, safety and security, and military use.

We demonstrate a versatile class of nanoscale chemical sensors based on single-stranded DNA (ssDNA) for chemical recognition and single-walled carbon nanotube field effect transistors (SWNT FETs) for electronic read-out. SWNT FETs with a nanoscale coating of ssDNA respond to vapors that cause no detectable conductivity change in bare devices. Sensor responses differ in sign and magnitude for different gases and can be tuned by choice of the ssDNA base sequence. Sensors respond and recover rapidly (seconds), and the sensor surface is self-regenerating. Preliminary results of all-atom molecular dynamics simulations agree with experiment. (© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

Exhaled breath contains thousands of gaseous volatile organic compounds (VOCs) that may be used as non-invasive markers of lung disease. The electronic nose analyzes VOCs by composite nano-sensor arrays with learning algorithms. It has been shown that an electronic nose can distinguish the VOCs pattern in exhaled breath of lung cancer patients from healthy controls. We hypothesized that an electronic nose can discriminate patients with lung cancer from COPD patients and healthy controls by analyzing the VOC-profile in exhaled breath. 30 subjects participated in a cross-sectional study: 10 patients with non-small cell lung cancer (NSCLC, [age 66.4+/-9.0, FEV(1) 86.3+/-20.7]), 10 patients with COPD (age 61.4+/-5.5, FEV(1) 70.0+/-14.8) and 10 healthy controls (age 58.3+/-8.1, FEV(1) 108.9+/-14.6). After 5 min tidal breathing through a non-rebreathing valve with inspiratory VOC-filter, subjects performed a single vital capacity maneuver to collect dried exhaled air into a Tedlar bag. Th...

In this article, we review gas sensor application of one-dimensional (1D) metaloxide nanostructures with major emphases on the types of device structure and issues for realizing practical sensors. One of the most important steps in fabricating 1D-nanostructure devices is manipulation and making electrical contacts of the nanostructures. Gas sensors based on individual 1D nanostructure, which were usually fabricated using electron-beam lithography, have been a platform technology for fundamental research. Recently, gas sensors with practical applicability were proposed, which were fabricated with an array of 1D nanostructures using scalable micro-fabrication tools. In the second part of the paper, some critical issues are pointed out including long-term stability, gas selectivity, and roomtemperature operation of 1D-nanostructure-based metal-oxide gas sensors.

Low cost gas multisensor devices can represent an efficient solution for densifying the sparse urban air pollution monitoring mesh. In a previous work, we proposed and evaluated the calibration of such a device using short term on-field recorded data for the Benzene pollution quantification. In this work, we present and discuss the results obtained for CO, NO 2 and total NOx pollutants concentration estimation with the same set up. Conventional air pollution monitoring station is used to provide reference data.

The physico-chemical properties of oil from Moringa oleifera seed were determined following extraction either with petroleum ether or 2% Neutrase 0.8L (a neutral bacterial protease from Bacillus amyloiquefaciens, Novozyme Bagsvaerd Denmark). The enzyme was chosen following a preliminary study conducted on the enzymatic extraction of M. oleifera seed oil using four commercial enzymes that showed Neutrase to be the best enzyme with the highest oil recovery value. The enzymes used were Termamyl 120L, Type L (a-Amylase), Neutrase Ò 0.8L (Neutral protease), Celluclast Ò 1.5 L FG (Cellulase) and Pectinex Ò Ultra SP-L (Pectinase), all from Novozyme, Denmark. The fatty acid compositions of solvent and enzyme-extracted oil from M. oleifera seed were determined. Results showed that the solvent-extracted oil has 67.9% oleic acid compared to 70.0% in enzyme-extracted oil. Results obtained following analysis of extracted oil showed that the oil is highly unsaturated because of the high percentage of oleic acid. Apart from oleic acid, other prominent fatty acids were palmitic (7.8% and 6.8%), stearic (7.6% and 6.5%), and behenic (6.2% and 5.8%) acids for solvent and enzyme-extracted oils, respectively. It was liquid at room temperature and pale-yellow in colour (0.7R + 5.9Y and 0.7R + 3.0Y for solvent and enzyme-extracted oils, respectively). Electronic nose analysis showed that it had flavor similar to that of peanut oil. The melting points estimated by differential scanning calorimetry were found to be 19.0 and 18.9°C for the solvent-and enzyme-extracted oils, respectively. The oil contains 36.7% triolein as the main triacylglycerol. The extraction methods were found to slightly influence the relative amounts of the fatty acids in the oil. The oils extracted using these two methods were found to differ in the percentage composition of their fatty acids. Quality attributes such as relative percent of oleic acid, total percentage of unsaturated fatty acids, iodine value, free fatty acid and unsaponifiable matter contents and the colour of the enzymeextracted oil were better than those of the solvent-extracted oil.

We introduce the Cohesive Energy Density (CED) method, a multiple sampling Molecular Dynamics computer simulation procedure that may offer higher consistency in the estimation of Hildebrand and Hansen solubility parameters. The use of a multiple sampling technique, combined with a simple but consistent molecular force field and quantum mechanically determined atomic charges, allows for the precise determination of solubility parameters in a systematic way ( ϭ 0.4 hildebrands). The CED method yields first-principles Hildebrand parameter predictions in good agreement with experiment [root-mean-square (rms) ϭ 1.1 hildebrands]. We apply the CED method to model the Caltech electronic nose, an array of 20 polymer sensors. Sensors are built with conducting leads connected through thin-film polymers loaded with carbon black. Odorant detection relies on a change in electric resistivity of the polymer film as function of the amount of swelling caused by the odorant compound. The amount of swelling depends upon the chemical composition of the polymer and the odorant molecule. The pattern is unique, and unambiguously identifies the compound. Experimentally determined changes in relative resistivity of seven polymer sensors upon exposure to 24 solvent vapors were modeled with the CED estimated Hansen solubility components. Predictions of polymer sensor responses result in Pearson R 2 coefficients between 0.82 and 0.99.

Fabrication, characterization, and tests of the practical gradient microarray electronic nose with SnO 2 nanowire gas-sensing elements are reported. This novel device has demonstrated an excellent performance as a gas sensor and e-nose system capable of promptly detecting and reliably discriminating between several reducing gases in air at a ppb level of concentration. It has been found that, in addition to the temperature gradient across the nanowire layer, the density and morphological inhomogeneities of nanowire mats define the discriminating power of the electronic nose.

The European fish industry is still reluctant to implement methods other than sensory to monitor freshness and quality of fish products, although general concensus exists about the importance of various quality attributes and the need for methods to monitor quality. The objective of the project FAIR CT98-4076 (MUSTEC) was to evaluate several physico-chemical techniques and to integrate their outputs into a more robust estimate of the freshness quality of fish. The techniques used for this multisensor approach were based on visible light spectroscopy, electrical properties, image analysis, colour, electronic noses and texture. Combining the outputs of the instrumental techniques and calibrating them with sensory scores of Quality Index Method (QIM) for attributes like appearance, smell and texture, gives an Artificial Quality Index (AQI) that can be as accurate and precise as the QIM sensory score. The outcome provides a basis for the construction and industrial exploitation of multi-sensor-devices for defining the quality of fish. #

. Recently, medical diagnostics has emerged to be a promising application area for electronic noses e-nose . In this paper, we review work carried out at Warwick University on the use of an e-nose to diagnose illness. Specifically, we have applied an e-nose to the identification of pathogens from cultures and diagnosing illness from breath samples. These initial results suggest that an e-nose will be able to assist in the diagnosis of diseases in the near future. q

It is well known that insects and other animals use olfactory senses in a wide variety of behavioural processes, namely to recognize and locate food sources, detect predators, and find mates. This article discusses the gathering of olfactive information and its utilization by a mobile robot to find a specific odour source in a room with turbulent phenomena's and multiple sources of odour. Three navigation algorithms are compared with a simple gas sensor and with an electronic nose. Their performance in finding an ethanol source in a room with obstacles is evaluated. The first navigation strategy is based on bacteria chemotaxis. The second strategy is based on the male silkworm moth algorithm that is used to search and track a female moth pheromone plume. The last strategy is based on the estimation of odour geometry and gradient tracking. The electronic nose utilized is composed by an array of different and weakly selective metal oxide gas sensors. The odours are identified and quantified by a pattern recognition algorithm based on an artificial neural network. The test bed for the navigation algorithms was a Nomad Super Scout II mobile robot.

The field of electronic noses, electronic instruments capable of mimicking the human olfactory system, has developed rapidly in the past ten years. There are now at least 25 research groups working in this area and more than ten companies have developed commercial instruments, which are mainly employed in the food and cosmetics industries. Most of the work published to date, and commercial applications, relate to the use of well established static pattern analysis techniques, such as principal components analysis, discriminant function analysis, cluster analysis and multilayer perceptron based neural networks. The authors first review static techniques that have been applied to the steady-state response of different odour sensors, e.g. resistive, acoustic and FET-based. Then they review the emerging field of the dynamic analysis of the sensor array response. Dynamic signal processing techniques reported so far include traditional parametric and nonparametlic ones borrowed from the traditional field of system identification as well as linear filters, time series neural networks and others. Finally the authors emphasise the need for a systems approach to solve specific electronic nose applications, with associated problems of sensor diift and interfesence.

The occurrence of illnesses related with poor ventilation has driven an increasing attention towards indoor air quality monitoring. In buildings equipped with climate control systems, the diseases related to the air quality can be significantly reduced if smart intervening procedures, aiming to control the concentration of pollutants in the indoor air, can be implemented in the heating, ventilation air conditioning unit. When reliable information about both the indoor and outdoor air quality is made available, the climate control system can provide the most appropriate amount of ventilation, ensuring safe and comfortable living conditions.

We have developed a miniature polymer-based chemical gas sensor array on silicon using micromachining technology. The sensors are polymer±carbon black composite ®lms, which swell reversibly and cause a resistance change upon exposure to a wide variety of gases. We have fabricated two types of devices which can measure this resistance change using a``well'' design. These``wells'' contain the polymer± carbon black±solvent liquid volume present during deposition and allow the sensor ®lm to be placed reproducibly in a speci®c and wellconstrained area. After deposition, the solvents evaporate and leave behind a polymer±carbon black residue crust between metal leads on each side of the``well''. Two types of devices, a bulk micromachined sensor and a patterned thick-®lm sensor, have been fabricated, ranging in size from 500 mm  600 mm to 100 mm  100 mm. Since the composite ®lm sensors are not speci®c to any one gas, an array of these sensors, each with a different sensing ®lm, is used to identify gases and gas mixtures through the pattern response of the array. Six polymer±carbon black composite ®lms were deposited into the sensor array and exposed to three chemical gases at ®ve different concentration levels. The sensors were able to uniquely detect these gas vapors and demonstrated a linear response to concentration levels between 2000 and 10,000 ppm. Results also indicate that a reduction in sensor area by an order of magnitude (from 4.32 to 0.30 mm 2 ) does not reduce sensor response. This design allows the integration of circuits to process the changes in resistance which will permit the realization of a completely integrated miniature gas sensor. # 2001 Published by Elsevier Science B.V. (J.W. Judy). 0925-4005/01/$ ± see front matter # 2001 Published by Elsevier Science B.V. PII: S 0 9 2 5 -4 0 0 5 ( 0 0 ) 0 0 6 3 8 -9

Anopheles funestus, were screened for their midgut bacterial contents. The midgut from each blood-fed mosquito was screened with two different detection pathways, one culture independent and one culture dependent. Bacterial species determination was achieved by sequence analysis of 16S rRNA genes. Altogether, 16 species from 14 genera were identified, 8 by each method. Interestingly, several of the bacteria identified are related to bacteria known to be symbionts in other insects. One isolate, Nocardia corynebacterioides, is a relative of the symbiont found in the vector for Chagas' disease that has been proven useful as a paratransgenic bacterium. Another isolate is a novel species within the ␥-proteobacteria that could not be phylogenetically placed within any of the known orders in the class but is close to a group of insect symbionts. Bacteria representing three intracellular genera were identified, among them the first identifications of Anaplasma species from mosquitoes and a new mosquito-Spiroplasma association. The isolates will be further investigated for their suitability for a paratransgenic Anopheles mosquito.

An investigation into the use of an electronic nose to predict the class and growth phase of two potentially pathogenic micro-organisms, Eschericha coli (E. coli ) and Staphylococcus aureus (S. aureus), has been performed. In order to do this we have developed an automated system to sample, with a high degree of reproducibility, the head space of bacterial cultures grown in a standard nutrient medium. Head spaces have been examined by using an array of six different metal oxide semiconducting gas sensors and classified by a multi-layer perceptron (MLP) with a back-propagation (BP) learning algorithm. The performance of 36 different pre-processing algorithms has been studied on the basis of nine different sensor parameters and four different normalization techniques. The best MLP was found to classify successfully 100% of the unknown S. aureus samples and 92% of the unknown E. coli samples, on the basis of a set of 360 training vectors and 360 test vectors taken from the lag, log and stationary growth phases. The real growth phase of the bacteria was determined from optical cell counts and was predicted from the head space samples with an accuracy of 81%. We conclude that these results show considerable promise in that the correct prediction of the type and growth phase of pathogenic bacteria may help both in the more rapid treatment of bacterial infections and in the more efficient testing of new anti-biotic drugs.

Some predictions concerning the technological expansion for the multisensor systems are made based on the analysis of the state-of-the-art Electronic Nose technology and correlation between their principles of functioning and the main mechanisms for olfactory reactions in living organisms. Emphasis is placed on the dynamic aspect of such systems performance, in particular, the additional value of the kinetic behavior when forming chemical images for composite multicomponent media using the multisensor arrays. ᮊ

The appreciation of food is based on the perception of many senses; in fact for a total estimation the human senses are involved. Biomimetics as the 'abstraction of good design from nature' is to analyze and tap nature's huge reservoir of potential innovative solutions. These biological solutions are cost-efficient, multi-functional and environmentally friendly. Human sense inspired sensor technologies such as multi arrays of sensors are of interest to food industry for food control and sensory evaluation. These systems mimic the human sense to provide a decision on final food quality and safety. Nanotechnology techniques and biosensors are the recent advances in these systems. Odor and taste sensors are interestingly under consideration to food engineers and scientist for application in food processing tasks such as drying, packaging, sorting and so on. Soft computing together with computer development is the promising outlook to enhance the performance of these sensing systems to food quality and safety characterization. This paper deals with some aspects of such systems (odor sensor and taste sensor) and some of more recent applications and advances for food control in food research and technology are discussed as well.

Drift is an important issue that impairs the reliability of gas sensing systems. Sensor aging, memory effects and environmental disturbances produce shifts in sensor responses that make initial statistical models for gas or odor recognition useless after a relatively short period (typically few weeks). Frequent recalibrations are needed to preserve system accuracy. However, when recalibrations involve numerous samples they become expensive and laborious. An interesting and lower cost alternative is drift counteraction by signal processing techniques. Orthogonal Signal Correction (OSC) is proposed for drift compensation in chemical sensor arrays. The performance of OSC is also compared with Component Correction (CC). A simple classification algorithm has been employed for assessing the performance of the algorithms on a dataset composed by measurements of three analytes using an array of seventeen conductive polymer gas sensors over a ten month period.

The performance of a pattern recognition system is dependent on, among other things, an appropriate data-preprocessing technique. In this paper, we describe a method to evaluate the performance of a variety of these techniques for the problem of odor classification using an array of gas sensors, also referred to as an electronic nose. Four experimental odor databases with different complexities are used to score the data-preprocessing techniques. The performance measure used is the cross-validation estimate of the classification rate of a K nearest neighbor voting rule operating on Fisher's linear discriminant projection subspace.

In this paper, we have (analyzed using a metal oxide sensor (MOS)-based electronic nose (EN)) five tea samples with different qualities, namely, drier month, drier month again over-fired, well-fermented normal fired in oven, well-fermented over-fired in oven, and under-fermented normal fired in oven. The flavour of tea is determined mainly by its taste and smell, which are determined by hundreds of volatile organic compounds (VOC) and non-volatile organic compounds present in tea. Tea flavour is traditionally measured through the use of a combination of conventional analytical instrumentation and human organoleptic profiling panels. These methods are expensive in terms of for example time and labour. The methods are also inaccurate because of a lack of either sensitivity or quantitative information. In this paper an investigation has been made to determine the flavours of different tea samples using an EN and thus to explore the possibility of replacing existing analytical and profiling panel methods. The technique uses an array of four MOSs, each of, which has an electrical resistance that has partial sensitivity to the headspace of tea. The signals from the sensor array are then conditioned by suitable interface circuitry resulting in our tea data-set. The data were processed using principal component analysis (PCA), fuzzy C means (FCM) algorithm. The data were then analyzed following the neural network paradigms, following the self-organizing map (SOM) method along with radial basis function (RBF) network and probabilistic neural network (PNN) classifier. Using FCM and SOM feature extraction techniques along with RBF neural network, we achieved 100% correct classification for the five different tea samples, each of which have different qualities. These results prove that our EN is capable of discriminating between the flavours of teas manufactured under different processing conditions, viz. over-fermented, over-fired, under-fermented, etc.

Ziehl-Neelsen (ZN) staining for the diagnosis of tuberculosis (TB) is time-consuming and operator dependent and lacks sensitivity. A new method is urgently needed. We investigated the potential of an electronic nose (EN) (gas sensor array) comprising 14 conducting polymers to detect different Mycobacterium spp. and Pseudomonas aeruginosa in the headspaces of cultures, spiked sputa, and sputum samples from 330 cultureproven and human immunodeficiency virus-tested TB and non-TB patients. The data were analyzed using principal-component analysis, discriminant function analysis, and artificial neural networks. The EN differentiated between different Mycobacterium spp. and between mycobacteria and other lung pathogens both in culture and in spiked sputum samples. The detection limit in culture and spiked sputa was found to be 1 ؋ 10 4 mycobacteria ml ؊1. After training of the neural network with 196 sputum samples, 134 samples (55 M. tuberculosis culture-positive samples and 79 culture-negative samples) were used to challenge the model. The EN correctly predicted 89% of culture-positive patients; the six false negatives were the four ZN-negative and two ZN-positive patients. The specificity and sensitivity of the described method were 91% and 89%, respectively, compared to culture. At present, the reasons for the false negatives and false positives are unknown, but they could well be due to the nonoptimized system used here. This study has shown the ability of an electronic nose to detect M. tuberculosis in clinical specimens and opens the way to making this method a rapid and automated system for the early diagnosis of respiratory infections.

Smell is probably the least understood and exploited of the principal human senses, yet it is clearly important to both product and process control in many industries, such as foodstuffs, beverages, tobacco and perfumery. Advances in the field of integrated microelectronic devices have led to new instruments, robots, capable of vision and complex touch or taction, but not yet of smell. This paper reviews the research effort that has been carried out at Warwick University over recent years into the development of an electronic instrument that can mimic the human sense of smell. The approach that we have adopted is to construct a microprocessor-controlled system comprising an array of solid-state chemical gas sensors (with overlapping partial sensitivities to odorants) and associated signal processing and pattern recognition. This electronic system' is based upon our present knowledge of the biological system. Our earliest electronic nose consisted of an array of only three to twelve tin dioxide thick-film sensors, yet it can discriminate betwe_en alcohols, beverages, tobacco blends and coffees. Current efforts are reported towards the fabrication of an integrated microsensor metal oxide array, the development of other electronic devices using polymeric materials, and the implementation of various patternrecognition techniques, including correlation, principalcomponent analysis, cluster analysis and artificial neural networks. Finally, the application areas most likely to arouse widespread interest in the next decade are discussed.

Since the first developments of electronic noses, food analysis has been considered as one of its most useful applications. In this paper an electronic nose based on quartz microbalances coated with metallo-porphyrins and related compounds is presented and illustrated. Extensive tests on various substances playing key roles in food analysis show that sensing properties of the sensing materials (in terms of sensitivity and selectivity) can be exploited for electronic nose applications devoted to the analysis of various kinds of foods. The versatility of this system has been successfully tested on different kinds of foods, such as fish, meat, vegetable and wine for which results are shown. © 1997 Elsevier Science S.A. 0925-4005/97/$17.00 © 1997 Elsevier Science S.A. All rights reserved. PII S 0 9 2 5 -4 0 0 5 ( 9 7 ) 0 0 1 7 5 -5

The quality of virgin olive oils is assessed through the determination of several analytical parameters, whose values must be within the ranges established by the different institutions involved. In addition to official methods, there is a strong need for simple, rapid, and environmentally friendly techniques for the quality control of virgin olive oils and for addressing the challenging task of determining geographical origin and detecting adulterants. Toward this purpose, some of the most interesting applications based on optical spectroscopic techniques, on the measurement of electrical characteristics and on the use of instruments equipped with electronic chemical sensors, including also other promising techniques are herein discussed. These techniques, adequately coupled with an appropriate statistical approach, appear to be promising for assessment of several quality-related parameters. The prediction of sensory attributes and of the oxidative status of virgin olive oils have also been reviewed by adopting these selected techniques, which are also considered to be potentially appropriate solutions for identification of the geographical origin of virgin olive oils and to assess their adulteration with cheaper oils. Overall, the techniques discussed are promising and cutting-edge approaches for the establishment of useful portable instruments for in situ monitoring of the quality of virgin olive oils. Practical applications: The simple, rapid, and environmentally friendly analytical approaches discussed herein represent promising analytical tools for assuring the authenticity and monitoring the quality of virgin olive oils. Such innovative techniques and tools need to be ring-tested and validated. Some innovative reviewed approaches will permit to develop useful portable instruments able to perform in situ appropriate controls also by small laboratories or olive oil mills with limited technical facilities. These equipments will be potentially usable also by trained " non-professional analytical skilled " people. Some other approaches, rapid but more expensive, will be applicable mainly by quality control labs and will increase the number of samples analyzed per day, thus fostering laboratory proficiency and an effective fighting against olive oil fraud.

Techniques for measuring odours in the field are reviewed. The possibility of relating results of field odour measurements and model outputs is investigated. Chemical analysis, though reliable and consolidated, is mostly unsuitable for odour assessment. Human panels (trained or untrained) are necessary for direct assessment of odour in the field. Electronic noses or sensors represent a promising technology for environmental odour monitoring.

The detection and simultaneous identification of a range of microorganisms by measuring the volatile compounds produced from plate cultures has been carried out using an electronic nose and a neural network classifier. Headspace samples were taken from ...

Low cost gas multisensor devices can represent an efficient solution for densifying the sparse urban air pollution monitoring mesh. In a previous work, we proposed and evaluated the calibration of such a device using short term on-field recorded data for the Benzene pollution quantification. In this work, we present and discuss the results obtained for CO, NO 2 and total NOx pollutants concentration estimation with the same set up. Conventional air pollution monitoring station is used to provide reference data.

In this study, we analyzed the transient of microsensors based on tin oxide sol-gel thin film. A novel method to this research field for data analysis and discrimination among different volatile organic compounds is presented. Moreover; several feature extraction methods have been considered, both steady-state (fractional change, relative, difference and log) and transient (Fourier and wavelet descriptors, integral and derivatives) information. Feature extraction methods have been validated qualitatively (by using principal component analysis) and quantitatively on the classification rate (by using a radial basis function neural network). #

An electronic nose and an electronic tongue, in combination with multivariate analysis, have been used to verify the geographical origin and the uniqueness of specific extra virgin olive oils. The olive oil samples belong to a small production, located in the lake of Garda (north of Italy) and distinguished with a European Protected Denomination of Origin trademark since 1998. In order to obtain a complete description of oil samples, free acidity, peroxide value, ultraviolet indices, and phenol content have been also determined. The dataset includes 36 Garda oils and 17 oils from other regions. Two classification models have been built by means of Counterpropagation Artificial Neural Networks in order to separate Garda and not-Garda oils, as follows: first, by using all the chemical variables and sensor signals; second, by using electronic tongue sensors; finally, by using four selected electronic nose sensors. All the models have been also tested with 19 commercial olive oil samples. Neural networks have provided very satisfactory results and have indicated the electronic nose as the most appropriate tool for the characterization of the analyzed oils. These results have suggested how electronic nose, in combination with neural networks, could represent a fast, cheap and functional method to classify and describe extra virgin olive oils from a circumscribed geographical area.

An electronic nose based system, which employs an array of inexpensive commercial tin-oxide odour sensors, has been used to analyse the state of ripeness of bananas. Readings were taken from the headspace of three sets of bananas during ripening over a period of 8-14 days. A principal-components analysis and investigatory techniques were used to define seven distinct regions in multisensor space according to the state of ripeness of the bananas, predicted from a classification of banana-skin colours. Then three supervised classifiers, namely Fuzzy ARTMAP, LVQ and MLP, were used to classify the samples into the observed seven states of ripeness. It was found that the Fuzzy ARTMAP and LVQ classifiers outperformed the MLP classifier, with accuracies of 90.3% and 92%, respectively, compared with 83.4%. Furthermore, these methods were able to predict accurately the state of ripeness of unknown sets of bananas with almost the same accuracy, i.e. 90%. Finally, it is shown that the Fuzzy ARTMAP classifier, unlike LVQ and MLP, is able to perform efficient on-line learning in this application without forgetting previously learnt knowledge. All of these characteristics make the Fuzzy-ARTMAP-based electronic nose a very attractive instrument with which to determine non-destructively the state of ripeness of fruit.

An electronic nose (e-nose), the Cyrano Sciences' Cyranose 320, comprising an array of thirty-two polymer carbon black composite sensors has been used to identify six species of bacteria responsible for eye infections when present at a range of concentrations in saline solutions. Readings were taken from the headspace of the samples by manually introducing the portable e-nose system into a sterile glass containing a fixed volume of bacteria in suspension. Gathered data were a very complex mixture of different chemical compounds. Linear Principal Component Analysis (PCA) method was able to classify four classes of bacteria out of six classes though in reality other two classes were not better evident from PCA analysis and we got 74% classification accuracy from PCA. An innovative data clustering approach was investigated for these bacteria data by combining the 3-dimensional scatter plot, Fuzzy C Means (FCM) and Self Organizing Map (SOM) network. Using these three data clustering...

A combined approach based on a multisensor system, to get chemical information from liquid samples through the analysis of the Ž . Ž solution and its head space is illustrated and commented. Sensors operating in liquid electronic tongue and in the head space electronic . nose are based on the same sensitive materials: the metalloporphyrins. This gives the opportunity of a certain uniformity of interaction putting in evidence the differences between the solution and its related volatile part. The combined system has been tested in two different experiments in the fields of clinical and food analysis, respectively. Results have shown a net increase of information when the combined systems are used. The sensor fusion procedure has been optimised in order to achieve the best results. q

A new drift compensation method based on Common Principal Component Analysis (CPCA) is proposed. The drift variance in data is found as the principal components computed by CPCA. This method finds components that are common for all gasses in feature space. The method is compared in classification task with respect to the other approaches published where the drift direction is estimated through a Principal Component Analysis (PCA) of a reference gas. The proposed new method -employing no specific reference gas, but information from all gases -has shown the same performance as the traditional approach with the best-fitted reference gas. Results are shown with data lasting 7-months including three gases at different concentrations for an array of 17 polymeric sensors.

The use of volatile production patterns produced by bacterial contaminants in urine samples were examined using electronic nose technology. In two experiments 25 and 45 samples from patients were analysed for specific bacterial contaminants using agar culture techniques and the major UTI bacterial species identified. These samples were also analysed by incubation in a volatile generation test tube system for 4 Á/5 h. The volatile production patterns were then analysed using an electronic nose system with 14 conducting polymer sensors. In the first experiment analysis of the data using a neural network (NN) enabled identification of all but one of the samples correctly when compared to the culture information. Four groups could be distinguished, i.e. normal urine, Escherichia coli infected, Proteus spp. and Staphylococcus spp. In the second experiment it was again possible to use NN systems to examine the volatile production patterns and identify 18 of 19 unknown UTI cases. Only one normal patient sample was mis-identified as an E. coli infected sample. Discriminant function analysis also differentiated between normal urine samples, that infected with E. coli and with Staphylococcus spp. This study has shown the potential for early detection of microbial contaminants in urine samples using electronic nose technology for the first time. These findings will have implications for the development of rapid systems for use in clinical practice. #

Different supervised pattern recognition treatments were applied to the signals generated by an electronic nose for the classification of vegetable oils. The system, comprising six metal oxide semiconductor sensors, was used to generate a pattern of the volatile compounds present in the samples. Feature selection techniques were employed to choose a set of optimally discriminant variables. The K-nearest neighbours (KNN), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), soft independent modelling of class analogy (SIMCA) and artificial neural networks (ANN) were applied to model the different classes. The results obtained indicated good classification and prediction capabilities, the neural networks being those that afforded the best results.

The present study investigates the potential of Radial Basis Function (RBF) neural networks for the prediction of reference evapotraspiration (ETo). The study utilizes daily climatic data of temperature, relative humidity, sunshine hours, wind speed, and rainfall for five years collected from Mosul meteorological station, north of Iraq. Thirteen RBF networks each using varied input combination of climatic variables have been trained and tested. The network output is compared with estimated daily Penman-Monteith ETo values. To evaluate the performance of RBF networks, the same networks in the studied cases were re-trained using the wellknown feedforward-backpropagation (FF-BP) networks. In addition, the effect of including a time index within the inputs of considered networks is investigated.