Wavelet singular entropy-based feature extraction from a temperature modulated gas sensor

Abstract

This paper demonstrates that a single thermally-modulated semiconductor gas sensor can discriminate and measure concentrations between two different explosive gases (CH4 and H2) and their mixtures. This method uses a novel feature extraction method, which is based on the wavelet singular entropy (WSE). From the time-frequency domain and energy spectrum perspective, wavelet decomposition coefficients and WSE are extracted as the features from the dynamic response of a single SnO2-based sensor in a rectangular temperature mode. Also, distance criterion as the feature evaluation criteria is employed to determine the optimal wavelet function, decomposition level and wavelet coefficients. Experimental results show that, compared with fast Fourier transform (FFT) and discrete wavelet transform (DWT), the WSE technique is more effective in terms of feature extraction and is highly tolerant to the presence of serious additive noise in the sensor response.