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

•The new curvature demodulation method adopts the primary function decomposition method to decompose the spectral curve.•The primary function decomposition method converts a large amount of information in the spectral curve into a coefficient vector.•Support vector machine regression is a critical step in extracting curvature information from multidimensional data (coefficient vectors).•The new curvature demodulation method is suitable for the demodulation of other physical quantities with the sensor spectral data as the input.