Pattern recognition electronic nose systems

Fig. 2. The principle configuration of an electronic nose system where the analyte mixture is contacted with a chemical sensor array that produces raw data which subsequently are treated with a pattern recognition algorithm that delivers the predicted result...

An electronic nose (Fig. 1) uses an array of non-specific broadly tuned sensors to discriminate odours [5-11]. The odours are analyzed by sensor array data with pattern recognition methods [4, 13]. A customized electronic nose set-up has been developed such that the same can be used in quality estimation of cardamom for monitoring of volatile emission pattern. The electronic nose consists of (a) sensor array, (b) micro-pump with programmable sequence control, (c) PC-based data acquisition and (d) olfaction software as illustrated in Fig. 2. The same system is used for study of cardamom. 

Keywords electronic nose principal component analysis pattern recognition chemical sensors sensor arrays olfaction system multivariate data analysis. 

J.W. Gardner and P.N. Bartlett, Pattern recognition in odour sensing. In Sensors and Sensory Systems for an Electronic Nose, Vol. 212, Kluwer Academic Publishers, Dordrecht, 1992, pp. 161-179. 

An electronic nose is an instrument which comprises an array of electronic chemical sensors with partial sensitivity and an appropriate pattern recognition system capable of recognising simple or complex odours ... 

Figure 1. Schematic architecture of an electronic nose showing an array of chemical sensors, pre-processing, array processing and finally a supervised pattern recognition system. A crude analogue to the biological system is shown as well. See chapter 2 of reference [2] for details of the biology of olfaction.

Electronic noses and tongues are arrays of sensors used to characterize complex samples, with the former being arrays of gas sensors while the latter are composed of liquid sensors [1]. These devices are composed of a chemical sensing system and a pattern recognition (PR) system [usually an artificial neural network (ANN)]. The array sensing system allows different properties to be measured simultaneously [2], Each chemical, which reaches the sensor array, will produce a characteristic pattern and therefore a database of patterns will be built up for a series of chemicals [2],... 

Nabarun Bhattacharyya is Associate Director in Centre for Development of Advanced Computing (C-DAC), Kolkata, India, which is a premier R D Institute under Department of Information Technology, Government of India. He is a Ph.D. from Jadavpur University, Kolkata, India. He has authored more than 100 papers in peer-reviewed journals and conferences and two book chapters on the topics electronic nose, electronic vision and electronic tongue. His research areas focus on machine olfaction, soft computing, pattern recognition embedded systems for agricultural and environmental applications. He is a member of IEEE. 

SAW devices are also considered as electronic noses. An electronic nose is a device that is composed of a chemical-sensing system and a pattern-recognition system. Electronic... 

It can be concluded that gas sensors can be made very small and with small consumption of the molecules to be detected. They can, therefore, be used to analyze gas evalution from small samples and with small gas flows with both industrial and medical applications. Furthermore, the use of sensor arrays gives the possibility to construct electronic noses for more complicated situations related to identification, classification and quantification of gas mixtures or odours. The final microanalysis system for gases may perhaps consist of microfabricated gas handling, an integrated sensor array and a dedicated chip for signal treatment and pattern recognition. 

An electronic nose functions by analyzing a sensor array response to a complete aroma, i.e., there is no separation of aroma components. The sensor array response to any given aroma is correlated (pattern recognition software) to sensory panel data. Using nenral network software and many training samples, the system determines a sensor response pattern that is representative of a fresh milk vs. a spoiled milk, for example. This technique is relatively young but has generated a substantial number of publications [5,85]. It is particularly attractive for quality control applications where an acceptable/unacceptable decision is often needed. 

At first glance the technique appears to be ideal in that there is no need for separation of volatiles. This can resnlt in very rapid analysis. Also, it seems to be based on a process similar to the human olfactory system in that both the electronic nose and human olfactory systems consist of a host of receptors (sensors) and yield a pattern of response to any given aroma. The brain, in the case of the human, and the computer, in the case of the electronic nose, make judgments based on a pattern recognition process as to the aroma and its quality. Thus, the speed is attractive and theoretical foundation appears to be rational. 

---