Fourier analysis for the quantitative detection of carbon monoxide and hydrogen gases using metal oxide gas sensor array

Abstract

The project aims to propose a Fourier approach and spectral analysis to predict the pure and mixture gas concentration quantitatively using cheap sensors which do not have the stability and reproducibility. In this analysis, various practical limitations on manual gas injection, glitches in the data due to the hardware, inconsistent recovery of sensor and drifting problems were taken into account. Several signal processing techniques of Fast Fourier Transform (FFT) were applied such as: interpolation, truncation, periodization and normalization on the finite time response of resistance/conductance variation. In this project, five methods were evaluated and then a FFT model based on conductance time-response characterization was selected for extensive analysis.on better prediction.