Machine learning models on demodulation of FBG sensors

Abstract

Pressure, acceleration, vibration, strain, and temperature are all often measured with Fiber Bragg Grating (FBG) sensors. Spectral overlapping in the wavelength domain can occur in a multiplexed FBG network. The main difficulty that will be explored in this project is demodulating a single FBG wave from a made by mixing spectrum. To determine the center Bragg wavelengths of each detector in the overlapping condition, the lowest technique, linear regression used in this project. With a root mean square error of 0.21 pm and an average testing duration of 0.8 milliseconds, a 30 layers residual