Dynamic multi-sensor operation and read-out for highly selective gas sensor systems

We describe hardware and algorithms which enable highly selective and sensitive operation of the two gas sensor types used in the SENSIndoor project. The resistance of a metal-oxide semiconductor (MOS) type can rise above 1 G Omega in temperature cycled operation (TCO), which is measured using a logarithmic amplifier. A silicon-carbide based, gas-sensitive field-effect transistor (SiC-FET) driven with a combination of TCO and gate-bias cycled operation (GBCO) is used as second, complimentary sensor. The cyclic sensor signals exhibit distinct shape changes depending on the gas present which is captured by pattern recognition. In this study we use Linear Discriminant Analysis (LDA) for discrimination and Partial Least Squares Regression (PLSR) for quantification of ppb concentrations of target VOCs in changing ppm concentrations of interfering gases for indoor air quality assessment. (C) 2016 The Authors. Published by Elsevier Ltd