earticle

영어

In order to improve the detection accuracy of wavelength, the filtering and curve fitting technologies were applied in the FBG wavelength demodulation system based on tunable F-P filter. These methods could realize the accurate peak-location of output signals of the photo detector. According to the characteristics of noise, the FIR low-pass filter was designed to filter the obtained light power signals so as to provide the input signals with high SNR for the peak-detection algorithms. By analyzing and comparing several typical peak-searching algorithms, the algorithm of Gauss formula nonlinear curve fitting (L-M) was chosen to fit the digitized light power signals. The experimental results show that L-M fitting algorithm reduces the mean square error by 7.5% compared with the Gauss fitting algorithm. For the Gauss signal in the wavelength demodulation system designed in the paper, the L-M algorithm has lower mean square error than other peak-searching algorithms. This algorithm is suitable for FBG wavelength demodulation system based on tunable F-P filter. It can efficiently raise the accuracy of wavelength demodulation system.