Method for correcting phase jump caused by polarization-induced fading in optical fiber phase demodulation

Patent metadata

Field	Value
Publication number	US-11223426-B2
Application number	US-202117180735-A
Country	US
Kind code	B2
Filing date	Feb 20, 2021
Priority date	Jun 9, 2020
Publication date	Jan 11, 2022
Grant date	Jan 11, 2022

Abstract

A method for correcting a phase jump caused by polarization-induced fading in optical fiber phase demodulation, including the steps of: 1, selecting a demodulated phase in the case of non-depolarization as historical sample data; 2, determining an autoregressive coefficient and a moving average coefficient of the autoregressive moving average model for the demodulated phase; 3, establishing a Kalman prediction model for the demodulated phase, and deriving recursive equations of the Kalman prediction model for the demodulated phase; and 4, judging whether a jump point exists in the actual demodulated phase, determining polarization states of lights if the jump point exists, and correcting the jump point when the polarization states of the lights are in polarization orthogonality by replacing the actual demodulated phase with a predicted phase value. The disclosure ensures the correctness of subsequent vibration-based signal processing.

First claim

What is claimed is:

1. A method for correcting a phase jump caused by polarization-induced fading in optical fiber phase demodulation, comprising steps of:
step 1: analyzing optical signals output by three terminals of a 3×3 optical fiber coupler in a Mach Zender interferometer, and selecting a demodulated phase in a case of non-depolarization as historical sample data;
step 2: using a historical sample sequence selected in step 1, building an autoregressive moving average model for the demodulated phase through time sequence analyses, determining orders of the autoregressive moving average model for the demodulated phase through Akaike Information Criterion or Bayesian Information Criterion, and determining an autoregressive coefficient and a moving average coefficient of the autoregressive moving average model for the demodulated phase through a least-square estimation;
step 3: representing a Kalman state vector with the demodulated phase, initializing parameters such as a Kalman state transition matrix, a system noise vector, and a prediction output matrix with the autoregressive moving average model for the demodulated phase obtained in step 2, establishing a Kalman prediction model for the demodulated phase, and deriving recursive equations for the Kalman prediction model for the demodulated phase;
step 4: performing real-time prediction on the demodulated phase with the Kalman prediction model for the demodulated phase, subtracting a predicted phase from an actual phase demodulated with an arctangent algorithm and a phase deconvolution algorithm to judge whether a jump point exists in an actual demodulated phase, determining a polarization state of light if the jump point exists, and correcting the jump point when the polarization state of the light is determined to be depolarized by replacing an actual demodulated phase value with a predicted phase value.

2. The method for correcting a phase jump caused by polarization-induced fading in optical fiber phase demodulation according to claim 1 , wherein: in step 1, in a case of non-depolarization, a phase difference of 120° exists in the optical signals output by three terminals of a 3×3 optical fiber coupler, a photoelectric detector detects that the optical signals output by three terminals of the 3×3 optical fiber coupler do not tend to be equal, and the demodulated phase in the case of non-depolarization is selected as the historical sample data {φ(k)} through a judgment of an amplitude of the optical signals output by three terminals of the 3×3 optical fiber coupler.

3. The method for correcting a phase jump caused by polarization-induced fading in optical fiber phase demodulation according to claim 1 , wherein: in step 2, the autoregressive moving average model for the demodulated phase is built through the historical sample data {φ(k)}, and the demodulated phase at a current moment of the autoregressive moving average model for the demodulated phase and the demodulated phase at a historical moment satisfy a following relational equation:
φ( k+ 1)= a 1 φ( k )+ a 2 φ( k− 1)+ . . . + a p φ( k−p+ 1)

e ( k+ 1)+θ 1 e ( k )+ . . . +θ q e ( k−q+ 1)  (1)

where φ(k+1), φ(k), φ(k−1), . . . , φ(k−q+1) represent demodulated phase values at moments k+1, k, k−1 . . . , k−p+1, respectively; e(k+1), e(k), . . . , e(k−q+1) is a white noise residual error sequence, representing a residual error between the current moment and the historical moment, a n (n=1, 2, . . . , p) is an autoregressive coefficient, p is a regression order, θ m (m=1, 2, . . . , q) is a moving average coefficient, and q is a moving average order;
the orders p and q of the autoregressive moving average model for the demodulated phase are determined through Akaike Information Criterion or Bayes Information Criterion, and the autoregressive coefficient a n and the moving average coefficient θ m of the autoregressive moving average model for the demodulated phase are determined through a selected historical sample sequence {φ(k)} of the demodulated phase in conjunction with a least-square estimation, so that a specific form of the autoregressive moving average model for the demodulated phase is determined.

4. The method for correcting a phase jump caused by polarization-induced fading in optical fiber phase demodulation according to claim 3 , wherein: in step 2, determining the orders of the autoregressive moving average model for the demodulated phase through Akaike Information Criterion or Bayesian Information Criterion comprises specifically:
determining the orders of the autoregressive moving average model for the demodulated phase to be ARMA (p, q) through Akaike Information Criterion, where p and q satisfy a following relational equation;
min AIC= n ln {circumflex over (σ)} ε 2 +2( p+q+ 1)

determining the orders of the autoregressive moving average model for the demodulated phase to be ARMA (p, q) through Bayesian information criterion, where p and q satisfy a following relational equation;
min BIC=ln( n )·( p+q+ 1)−2 ln {circumflex over (σ)} ε 2

where min AIC is a most appropriate order of the autoregressive moving average model for the demodulated phase through Akaike Information Criterion, min BIC is a most appropriate order of the autoregressive moving average model for the demodulated phase through Bayesian Information Criterion, n is a number of the demodulated phase φ(k) selected as a sample, and {circumflex over (σ)} ε 2 is an estimation of white noise variance of the autoregressive moving average model.

5. The method for correcting a phase jump caused by polarization-induced fading in optical fiber phase demodulation according to claim 4 , wherein: parameters a n (n=1, 2, . . . , p), θ m (m=1, 2, . . . , q) of the autoregressive moving average model for the demodulated phase are determined through the historical demodulated phase sample sequence {φ(k)} in conjunction with the least-square estimation, so that the specific form of the autoregressive moving average model for the demodulated phase is determined.

6. The method for correcting a phase jump caused by polarization-induced fading in optical fiber phase demodulation according to claim 1 , wherein: the Kalman prediction model for the demodulated phase established in step 3 is expressed as:

{

Φ
⁡

(

k
+
1

)

=

A
⁡

(

k
+
1

,
k

)

⁢

Assignees

• Univ Wuhan Tech

Inventors

• Li Zhengying
• Fan Minlang
• Wang Honghai
• Wu Jun
• Wang Jiaqi

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N20/00

  Machine learning · CPC title

• H04B10/697

  Arrangements for reducing noise and distortion · CPC title

• H04B10/2507 Primary

  for the reduction or elimination of distortion or dispersion · CPC title

• G01H9/004

  using fibre optic sensors (light guides per se G02B6/00, acousto-optical devices specially adapted for gating or modulating in optical wave guides G02F1/125) · CPC title