Regularization and the inflection point method for sensor signal in gas concentration measurement

Abstract

In this work, we focus on a transient state of sensor as a source of chemical information. This problem deserves attention for many reasons. As an example, utilization of transient states may allow for faster and more accurate information acquisition. We have considered a transient state gas sensor response to test analyte applied for quantitative gas determination. We examined the inflection point which is a promising feature of a signal resulting from transient sensor responses. This particular feature was found to be linearly dependent on the target gas concentration. However, the applicability of inflection point method is dependent on the precision of its determination. It raises a number of issues related to signal regularization. The regularization problem was studied from a theoretical point of view. A number of a-priori estimates for regularization strategies were obtained yielding some practical algorithms for calculating the inflection point in a stable way. Next, they were validated using numerical simulations and real data analysis procedure. This work provides a rigorous mathematical evaluation of the inflection point method and it may contribute to the improvement of gas sensor performance.