Sensor array signal processing for cross-sensitivity compensation in non-specific organic semiconductor gas sensors
A fundamental limitation of many chemically sensitive organic semiconductor materials is their high susceptibility to cross-interference resulting from interactions with background species other than those actively being detected. Such cross-sensitivities often preclude their use in 'real' sensor applications, particularly where discrete and selective gas sensing systems are required. It has been hypothesised, however, that this lack of specificity can largely be overcome with the adoption of a multi-element sensor array, thereby allowing the compensation of unwanted sensitivities through suitable signal processing. This thesis describes how such a multi-element sensor array of different gas sensitive metallophthalocyanine films, constructed on a single substrate, was used as the sensing element in an 'intelligent' chemical sensor. Since the individual sensors show varying degrees of gas sensitivity, the individual responses of each to any particular analyte will give rise to a characteristic change in the output 'pattern' comprised of each of the sensor resistances. By monitoring the change in this pattern of responses on exposure to specific gases of pre-determined concentration and employing a suitable feature extraction algorithm, the characteristic responses to particular analytes was learnt, and a knowledge base, from which future inferences may be drawn, was constructed. The success of suitable signal processing techniques to accommodate the inherent cross-sensitivities exhibited by these materials is demonstrated. The results demonstrate the viability of pattern recognition methods to analyse gas mixtures by comparing particular features of the combined array response with those previously learnt during a gas recognition training phase.