Application of an artificial neural network for simultaneous measurement of bending curvature and temperature with long period fiber gratings

Abstract An artificial neural network approach is proposed for simultaneous measurement of bending curvature and temperature for the embedded long period grating bending sensors. The sensing system is composed of two different types of long period gratings: one is H 2 -loaded and the other is Bo/Ge co-doped. By measuring the variations of the amplitudes of the transmission dips, the bending and temperature can be determined simultaneously. The root-mean-square (RMS) errors of the measurements of the bending curvature and the temperature are 0.0072 m −1 and 0.1898 °C, respectively.