Neue Entwicklungen in der Testmethodik für Gassensorsysteme

: Endres, H.-E.; Göttler, W.; Jander, H.D.; Drost, S.; Sberveglieri, G.

Abstract
The task for a gas sensing system is to measure one or more specified gases and to suppress any disturbance due to other gases and humidity. Nearly all gas sensors have a lack in their selectivity and show interferences towards several other gases. A common method to suppress interferences is the use of multisensor systems together with an intelligent signal evaluation system. These systems come more and more in the focus of applied research. Fore most semiconducting sensors were used for such applications, but other sensors are also in the field of research. Self learning systems like Artificial Neural Networks (ANN) are one of the best methods for signal evaluation systems of sensor signals. Common aspect of all these systems is the immense necessity of training data for the calibration of such systems. This implies a sophisticated measurement technology as well as improved methods to handle with calibration data and training such systems.