A novel method of curvature demodulation for Mach-Zehnder interferometer optical fiber curvature sensors

Abstract

This paper constructs a novel curvature demodulation method for Mach-Zehnder interferometer (MZI) optical fiber sensors. The method is composed of two parts: spectral curve decomposition and support vector machine regression (SVR) of curvature. The authors fabricated the MZI fiber optic curvature sensor and carried out curvature loading experiments. The curvature demodulation method proposed in this paper includes the following steps: spectral data acquisition, data normalization, coefficient vectors calculation, and curvature calculation. The minimum root mean square errors (RMSE) of the curvature demodulation method is 0.02133 m−1. This method has two characteristics: wide curvature demodulation range and unitive curvature regression expression. The curvature demodulation method is not only limited to the MZI sensor curvature demodulation but also applicable to the demodulation of other physical quantities with the spectral data as input. This paper provides a new idea for solving similar demodulation problems.