Reference Techniques for Electronic Noses for Indoor Air Monitoring Applications

摘要

An electronic nose (EN) with its high gas discrimination power can be a suitable instrument for indoor air monitoring tasks. Referencing is one of the important signal pre-processing steps to adapt the EN system to the special task for indoor air monitoring considering the background air conditions with usually not constant behavior. Several methods to achieve an indoor air reference have been developed which take into account that the background composition can differ from room to room and is subject to changes. Techniques based on hard- or software solutions have been tested. While the hardware solutions suffer from its higher effort, slow response and service dependence, the advantages of the software solutions convince especially by its flexible adaptability to the specificities of the monitoring tasks and to the distinctions of the premises to be kept under surveillance. A self-adaptive reference algorithm based on a recursive digital low-pass filter has been tested successfully for online refreshment of an air reference by continuous analysis of the EN data obtained in selected monitoring periods of the past. Fast event detection with rise times of seconds can be handled as well as tracking of much slower changes such as the usual diurnal cycles of indoor air quality.