Machine Learning Approach to Data Processing of TFBG-Assisted SPR Sensors

Abstract

Fiber optic sensors are applied in industry, remote sensing, environmental monitoring and healthcare. A special place is occupied by tilted fiber Bragg gratings, which can significantly expand the capabilities provided by standard Bragg sensors. But these gratings have complex spectral responses, therefore, data processing becomes a critical task for achieving maximum performance. In this paper, machine learning methods for processing spectral data of a plasmonic fiber sensor based on a tilted fiber Bragg grating were applied for the first time for the measurement of small refractive index changes. The responses of two similar but not identical sensors were measured in two independent experiments. The model trained on the data of the first sensor was used to analyze data obtained with another sensor. The best resolution achieved in our experiments was <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$9 \times {10^{ - 6}}$</tex-math></inline-formula> RIU.