An alternative global feature extraction of temperature modulated micro-hotplate gas sensors array using an energy vector approach

In this paper the modulation of the operating temperature of metal oxide gas sensors combined with a novel global feature extraction method based on the energy signal of the gas sensor array response is introduced. The method is demonstrated solving a practical application in identification and prediction of pollutant species using an array of four metal oxide semiconductor gas sensors. Features extracted by means of the global feature extraction approach have been compared with the standard temperature modulation feature extraction method. Different validation techniques have been implemented, which show that the method performs well and is of interest for simple, small size and real time gas analyzers. Excellent results have been obtained, not only accurately identifying and predicting the species studied, but also dramatically reducing down to 1 s the portion of the response transient necessary to achieve a sufficiently accurate prediction.