Multimodal demodulation algorithm for fiber optic MEMS fabry perot sensors

This paper addresses the issue of low demodulation accuracy in interferometric signals caused by significant errors in direct peak finding and positioning during multi-peak demodulation of fiber-optic MEMS Fabry Perot Sensors. To tackle this problem, we propose a novel approach that involves decomposing the fiber-optic Fabry Perot sensing signals through preliminary peak finding. By separating the multi-peak interferometric signals into individual single-peak signals, we then perform segmented Gaussian fitting on the extracted data. To improve the accuracy of the signal peak finding algorithm, the spectra are subjected to spectral smoothing and filtering using the Savitzky-Golay method. The feasibility and superiority of the algorithm were verified through MATLAB simulations and conducted demodulation experiments using fiber optic MEMS Fabry Perot sensors. The results demonstrate a linearity of 99.98% in the fitted curve between temperature and cavity length, showcasing the efficacy of this algorithm. Furthermore, compared with the direct peak-finding algorithm, the demodulation accuracy is significantly improved by more than ten-fold.