Abstract

In this paper, an adaptive filtering scheme using the least mean square (LMS) algorithm is proposed to reduce the noise floor of fiber optic sensing system based on phase generated carrier. The application of ellipse fitting algorithm can eliminate the effects of phase modulation depth fluctuation and carrier phase delay to ensure the accurate demodulation of the signal to be measured. The demodulation results are processed by an adaptive filtering procedure, which can effectively reduce harmonic distortion and noise. The experimental data show that the error signal converges to minimum value within 0.25 s, the filtered signal improves the signal-to-noise ratio at 1 kHz by about 40 dB compared with the conventional algorithm, and the system noise floor is reduced by 20 dB to 1.77 μrad/√Hz on average. This technique provides a strong support for the implementation of fiber optic sensing system in complex noise environment.