A neural approach for improving the measurement capability of an electronicnose

Abstract

Electronic noses, instruments for automatic recognition of odours, aretypically composed of an array of partially selective sensors, a samplingsystem, a data acquisition device and a data processing system. For thepurpose of evaluating the quality of olive oil, an electronic nose based onan array of conducting polymer sensors capable of discriminating oliveoil aromas was developed. The selection of suitable pattern recognitiontechniques for a particular application can enhance the performance ofelectronic noses. Therefore, an advanced neural recognition algorithm forimproving the measurement capability of the device was designed andimplemented. This method combines multivariate statistical analysis and ahierarchical neural-network architecture based on self-organizing maps and errorback-propagation. The complete system was tested using samples composedof characteristic olive oil aromatic components in refined olive oil. Theresults obtained have shown that this approach is effective in groupingaromas into different categories representative of their chemical structure.