Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Signal evaluation of gas sensors with neural nets

: Endres, H.-E.

Foulloy, L. ; International Federation of Automatic Control -IFAC-, Technical Committee on Components and Instruments:
SICICA '97. 3rd IFAC Symposium on Intelligent Components and Instruments for Control Applications. Preprints
Annecy, 1997
Symposium on Intelligent Components and Instruments for Control Applications (SICICA) <1997, Annecy>
Conference Paper
Fraunhofer IFT; 2000 dem IZM eingegliedert
air quality; gas sensor; gas warning; neural network; process control olfactometry; semiconductor sensor; sensor system; signal evaluation system; signal processing algorithm

Due to the lack in selectivity and separability of most common gas sensors, the use of sensor arrays together self adapting systems as artificial neural nets (ANN) is necessary. These systems can evaluate the gas concentration or deliver a binary signal (gas present or not, ventilation on or off, etc.), which gives a chance to solve measurement problems in gas mixtures. Also the estimation of indoor air quality, olfactometric measurements, etc. may be possible using ANN methods. A promising new approach is the use of non-stationary (transient) signals of a single sensor after a certain stimulus (i.e. a temperature change, etc.). This additional information enhances the redundancy of such a system and a single sensor delivers further information (virtual sensor array). Signal evaluation methods for chemical sensors are powerful tools, they are adaptable to specific applications and art compatible to customary micro controllers.