Fluctuation-Enhanced Sensing With Zero-Crossing Analysis for High-Speed and Low-Power Applications

A new method to generate fingerprints of chemical agents has been introduced in this paper. The method is based on the use of the zero-crossing statistics at fluctuation-enhanced sensing. It is a new version of Ben Kedem's original method based on low-pass filters. To improve computation time and energy efficiency, high-pass filtering is used, and in doing this in the simplest possible way, local zero levels for short-time subwindows are defined and a zero-crossing counting by the use of such windows is carried out. The method turns out to be an effective tool to identify noise processes with different spectra or amplitude distribution, with at least 1000 times less calculation and correspondingly lower energy need than that of the Kedem or the fast Fourier transform methods. We demonstrate the usability of the method by the analysis and recognition of different stochastic processes with similar and different spectra.