The Study Of Hilbert-Huang Transform Apply In Fiber-optic Extrinsic Fabry-Perot Interferometer’s Spectrum Detection

For the demodulation of reflectance spectrum of fiber optic EFPI sensor, how achieve the EFPI cavity length from the reflectance spectrum is a main job for demodulate the system of fiber optic EFPI sensor. In generally the demodulation algorithm have unimodal demodulation algorithm, bimodal demodulation algorithm, fast Fourier transform algorithm, and cursor demodulation algorithm, etc. In this algorithms, just fast Fourier transform algorithm don’t need filtering, the other algorithms must filter the signal before demodulate. But the precision of fast Fourier transform algorithm is affect by the sampling precision of the spectrometer, it can’t approach a good precision. In order to get high-accuracy cavity length of the fiber optic EFPI sensor, how to achieve the EFPI spectrum from the noisy EFPI spectrum is a important part of the research of EFPI system demodulation. Because of the unsteadiness of light source of EFPI system and the wastage in the fiber transmission, the EFPI spectrum come to be a nonstationary and nonlinear signal. The signal process methods based on Fourier transform have its shortages in the field to process the nonstationary and nonlinear signal, but the Hilbert-Huang transform (HHT) has its advantages than other algorithms in process the nonstationary and nonlinear signal. So we proposed to use the HHT algorithm to deal with the EFPI spectrum signal, get the cavity length information of the EFPI sensor.The algorithm of this article is modulate in the matlab. Using the HHT algorithm to achieve the information of EFPI form the simulative EFPI spectrum added noise, and get the cavity length by bimodal demodulation algorithm. It ensure that it is available to use the HHT algorithm to demodulate the EFPI spectrum. Analyzed the demodulation accuracy used this algorithm, and proposed the solutions to deal with the problems like end effects and mode mixing, and restrict the effect cause by those problems.The resolutions can’t solve this problems, so we put forward to use the HHT algorithm to filter the EFPI spectrum. Simulate the EFPI spectrum added the noise, use the HHT algorithm to filter the signal, then demodulate the signal with bimodal demodulation algorithm. This way it can get an EFPI cavity length which its accuracy at more than10nm. In order to get the performance of HHT filter algorithm, in the experiment it used FIR filter and wavelet transform to filter the signal, and use the bimodal demodulation algorithm to demodulate the EFPI cavity length. Compare the accuracy of the EFPI cavity length, we conclude that the HHT filter performance is better than the other two algorithms.In the end of this article, it used the HHT algorithm to filter the true EFPI spectrum and demodulate with bimodal demodulation algorithm. The demodulation accuracy have been affect by the EFPI cavity length, the deviation is very large. So we used the unimodal demodulation algorithm combined with bimodal demodulation algorithm to demodulate the EFPI spectrum and get the EFPI cavity length is differ than the result demodulate by the cursor demodulation algorithm only in2nm. The envelops of the true EFPI spectrum can affect the demodulate accuracy, so we normalized the signal by delete the EFPI envelops, then using the unimodal combined bimodal demodulation algorithm to demodulate the EFPI spectrum, and get the EFPI cavity length is differ than the result demodulate by the cursor demodulation algorithm only in0.1nm. This article demonstrate that the HHT algorithm is available using in the EFPI spectrum, and have a considerable prospect.