Rapid detection of target gases with an early prediction approach

Abstract

Certain gas sensing applications such as breath analyzers and food spoilage detection sensors require rapid step response to the target gases. However, the response time of semiconductor gas sensors is in a few to several minutes. In order to reduce the response time, we implemented an approach of early prediction in which the initial response data (first 30 seconds data) was compared with pre-defined output at different concentrations. Finally, the concentration corresponding to the pre-defined output which has less deviation with initial response data was determined as the concentration of the exposed gas. Though this method cannot predict the concentration of the target gas exactly, it can be used to identify the range of concentrations with much accuracy. By measuring the response of the sensor to exposure of combination of different gases, this technique can be utilized in identifying the multiple gases in the environment simultaneously with reasonable accuracy that is sufficient for many practical applications.