An on-line polarization-aligned optical vector analysis device and method

Abstract

The application provides an online polarization alignment optical vector analysis device and method, belonging to the technical field of optical fiber device testing, comprising a tunable laser source, a quadrature polarization generation module, a quadrature polarization interference fringe feedback module, a test interferometer module, a polarization diversity detection module, a test interference information feedback module and a to-be-tested device information acquisition module, characterized in that: after connecting the device, the quadrature polarization alignment is realized by minimizing the interference signal of the quadrature polarization generation module, the quadrature polarization interrogation light is constructed, and then the polarization diversity system maximum polarization alignment is realized by maximizing the amplitude of the first idle peak of the detector. The feedback control module for realizing polarization alignment is based on the Bayesian optimization algorithm, the polarization diversity detection module is used for coherent detection after polarization alignment, then the transmission matrix of the device is obtained, and the optical parameters of the device are obtained by demodulation. The test device can still effectively suppress the measurement error caused by polarization misalignment after accessing the device, and improve the reliability of the system test.

Description

Technical Field

[0001] This invention belongs to the field of fiber optic device testing technology, specifically relating to an online polarization alignment light vector analyzer based on Bayesian optimization. Background Technology

[0002] Spectral response characteristics are crucial parameters for the development and manufacturing of optical devices, optical systems, and photonic integrated circuits. Testing the spectral response is essential during the production, fabrication, and application of these devices. Furthermore, optical devices exhibit polarization-dependent characteristics, and obtaining these characteristics is also significant for device evaluation. In recent years, next-generation photonic systems, such as airborne and shipborne optical propulsion control systems, optically controlled phased array radar systems, high-speed fiber optic communication systems, and integrated photonic chip systems, have placed higher demands on the optical devices used in these applications, requiring precise testing of multiple dimensions of information from these devices. Moreover, the continuous development, fabrication, and application of high-precision optical devices and core photonic integrated chips necessitate that spectral response measurement techniques simultaneously meet the requirements of high spectral resolution, large measurement bandwidth, and high measurement speed.

[0003] Existing spectral response measurement techniques are mainly divided into three categories: optical vector analysis techniques based on broadband mapping, optical vector analysis techniques based on electro-optic modulation, and optical vector analysis techniques based on swept light sources and optical interferometers. Optical vector analysis techniques based on broadband mapping measure the spectral response of the optical device under test by using a broadband light source and a spectrometer. This method is simple to operate, but it can only measure the amplitude response of the optical device under test and cannot meet the measurement requirements of some optical delay line devices. Optical vector analysis techniques based on electro-optic modulation modulate a broadband, finely controllable electrical signal into the optical domain and use a high-precision microwave measuring instrument to obtain the spectral response of the optical device under test. It can achieve Hz-level testing [T. Qing, S. Li, Z. Tang, B. Gao, S. Pan, "Optical vector analysis with attometer resolution, 90-dB dynamicrange and THz bandwidth" Nature Communication, 10(1), 5135(2019)], but its measurement range is only 8nm, and its sensitivity is low due to direct detection. Optical vector analysis (OVA) based on swept-frequency light sources and optical interferometers is widely used to test the loss, delay, and polarization-related characteristics of optical devices. It boasts high measurement accuracy, large measurement bandwidth, and high measurement speed [Apparatus and method for the complete characterization of optical devices including loss, birefringence, and dispersion effects, US7042573B2]. It can also test the reflection characteristics of devices [Calculation of distributed birefringence and polarization mode dispersion from optical frequency domain reflectionometry, WO2020214637A1]. However, polarization misalignment is a major source of error in OVA. Traditional commercial instruments like the LUNA OVA5100 use offline polarization alignment, which significantly reduces the reliability and efficiency of testing. A single calibration cannot guarantee the accuracy of every measurement unless calibration is performed before each measurement, which greatly impacts measurement efficiency. Therefore, it is necessary to develop an online polarization alignment method to improve the accuracy of OVA.

[0004] Online polarization state adjustment commonly utilizes polarization analysis devices combined with intelligent algorithms to rapidly iterate to the target polarization state. In 2020, Zhang Xinliang et al. from Huazhong University of Science and Technology proposed a polarization stabilization system [A Control Method and System for Achieving Polarization Stabilization, CN112485930B]. This system uses a fast positioning algorithm to quickly adjust the output polarization state of any input light after passing through a polarization controller to the vicinity of any specified target polarization state on a Poincaré sphere. Then, a stochastic gradient descent algorithm further stabilizes the output light polarization state to the target polarization state, thus achieving stabilization of any polarization state to any set target polarization state. However, external polarization analysis devices are expensive and complicate the optical path. Another approach is to directly utilize the interference fringes in the optical path to maximize feedback polarization parallel alignment, using the Proportion Integral Differential (PID) algorithm for iteration [J. Bogdanski, J. Ahrens, and M. Bourennane. "Single mode fiber birefringence compensation in Sagnac and "plug & play" interferometric setups", Optics Express 17(6)4485-4494(2009).]. However, PID is prone to getting trapped in local extrema.

[0005] This invention provides an online polarization alignment method based on Bayesian optimization. Bayesian optimization is an emerging global optimization algorithm characterized by fast convergence of approximate calculations based on substitution models, making it highly promising for polarization alignment in OVA. When connecting the device under test (DUT), two auxiliary interference signals are used to perform minimization and maximization judgments to determine whether the polarization is aligned. The optimization objective function for polarization alignment contains a large number of local extrema, which significantly reduces the convergence speed of the Bayesian optimization algorithm. A data preprocessing method based on trigonometric function transformation is introduced to effectively solve this problem. OVA based on Bayesian optimization for online polarization alignment can effectively improve the reliability of the test. Summary of the Invention

[0006] The purpose of this invention is to provide an online polarization alignment light vector analysis device with short testing time and simple structure. This method can effectively suppress measurement errors caused by polarization misalignment.

[0007] An online polarization-aligned optical vector analyzer (OVA) device includes a tunable laser source 101, an orthogonal polarization generation module 2, an orthogonal polarization interference fringe feedback module 3, a test interferometer module 4, a polarization diversity detection module 5, a test interference information feedback module 6, and a device-under-test (DUT) information acquisition module 7. The tunable laser source 101 generates linearly swept continuous light with a wavelength varying over time, which is injected into the orthogonal polarization generation module 2. The orthogonal polarization interference fringe feedback module 3 acquires data from the interference fringes output by the orthogonal polarization generation module 2. Through rapid iteration via a first feedback control module 303, the first polarization controller 204 located on the upper arm 201a of the interferometer is adjusted until the amplitude of the output interference fringes is minimized, achieving orthogonal polarization alignment. Then, a pair of orthogonal polarization interrogation beams 21 are output to the test interferometer module 4. The amplitude and phase of the orthogonal polarization interrogation beams 21 change according to the amplitude and phase of the DUT 402 located on the upper arm 401a of the test interferometer, and then an optical signal carrying the amplitude and phase information of the DUT 402 is output. The polarization diversity detection module 5 decomposes the optical signal output by the test interferometer module 4 into two perpendicular polarization directions.

[0008] The second interference fringe acquisition module 601 acquires the fringes of the p detector 502 and the s detector 503, and then performs a fast Fourier transform 602. The signal corresponding to the p detector 502 or the s detector 503 is input to the first interference peak information acquisition module 603. The first interference peak 603a is selected according to the time delay difference τ between the two arms of the orthogonal polarization generation module 2. Then the amplitude value 603b of the first interference peak is read. The amplitude value 603b of the first interference peak is then processed by the second amplitude trigonometric function transformation module 604 and then input to the second feedback control module 605. The second feedback control module 605 iterates rapidly to determine the next adjustment angle of the second polarization controller 404 located on the lower arm 401b of the test interferometer. The feedback adjustment is made to maximize the amplitude value 603b of the first interference peak, thereby maximizing polarization alignment. Finally, the p detector signal 701a and the s detector signal 701b after the fast Fourier transform 602 are input to the device under test information acquisition module (7) to obtain the transmission characteristics of the device.

[0009] The orthogonal polarization interference fringe feedback module 3 is characterized in that: the first interference fringe acquisition module 301 acquires data of the interference fringes of the orthogonal polarization generation module 2 and calculates the amplitude of the orthogonal polarization interference fringe 301a. Then, the amplitude of the orthogonal polarization interference fringe 301a is processed by the first amplitude trigonometric function transformation module 302 and then input into the first feedback control module 303 to determine the next adjustment angle of the first polarization controller 204.

[0010] The orthogonal polarization generation module 2 is characterized in that: a linearly swept continuous beam is injected into the first coupler 201 and split into two beams. One beam, as the first test beam m1, passes through the lower arm 201b of the interferometer with the first extended fiber 203 and is injected into the second input end 202b of the second coupler. The other beam, as the input reference beam m2, is injected into the upper arm 201a of the interferometer with the first polarization controller 204. The first polarization controller 204 is adjusted to rotate the polarization direction of the input reference beam m2 and inject it into the first input end 202a of the second coupler. Then, the two beams are combined in the second coupler 202 and input into the first detector 203 through the first output end 202c of the second coupler to cause interference. The interference fringes are input into the orthogonal polarization interference fringe feedback module 3 and fed back to minimize the amplitude 301a of the orthogonal polarization interference fringes. Finally, the first test beam m1 and the second test beam n1, which are orthogonal to the polarization direction of the first test beam m1, are output through the second output end 202d of the second coupler to the test interferometer module 4.

[0011] The first amplitude trigonometric transformation module (302) is characterized in that: the trigonometric transformation is cot(y1 / V) max ×π / 2), where y1 is the amplitude of the orthogonal polarization interference fringes (301a), V max This is the maximum voltage value of the data acquisition card;

[0012] The second amplitude trigonometric transformation module (604) is characterized in that: the trigonometric transformation is tan(y2 / V) max ×π / 2), where y2 is the amplitude of the first interference peak (603b);

[0013] The test interferometer module 4 is characterized in that: the orthogonally polarized light 21 is split into two beams by the third coupler 401. One beam is output from the upper arm 401a of the test interferometer through the device under test 402 to the first input terminal 405a of the fourth coupler. The other beam is injected through the lower arm 401b of the test interferometer, and then sequentially passes through the second extension fiber 403 and the second polarization controller 404 to the second input terminal 405b of the fourth coupler. The two beams are combined at the fourth coupler and then input to the polarization diversity detection module 5.

[0014] The first feedback control module 303 and the second feedback control module 605 are characterized in that: the feedback control algorithm is a Bayesian optimization algorithm;

[0015] The device under test information acquisition module 7 is characterized in that: after polarization alignment is achieved, the p detector information 701a and s detector information 701b after the fast Fourier transform 602 are input into the second and fourth interference peak filtering modules 701, and the p detector τ is selected. D -τ time delay difference correlation peak a and p detector τ D +τ delay difference correlation peak b, and s detector τD -τ time delay difference correlation peak c and s detector τ D +τ Delay difference correlation peak d, τ D The test interferometer module 4 measures the time delay difference between the two arms. The peaks are filtered to obtain the relevant first real signal a1, second real signal b1, third real signal c1, and fourth real signal d1. Then, a Hilbert transform 702 is performed to obtain the first complex signal a2, second complex signal b2, third complex signal c2, and fourth complex signal d2. A Hilbert transform 702 is then performed, followed by Jones matrix demodulation 703 to obtain the device Jones matrix 703a. The optical parameters of the device under test (DUT) 402 are then calculated. These optical parameters include insertion loss, group delay, dispersion, polarization mode dispersion, and polarization-dependent loss.

[0016] Orthogonal polarization interferometry (OPI) measures optical vector characteristics using a pair of orthogonally polarized linearly swept continuous lasers with fixed time delays as interrogation beams. Coherent detection is performed using a polarization analysis detection system to obtain the Jones matrix of the device. This method can simultaneously demodulate the amplitude, phase, and polarization vector transmission characteristics of the optical device under test as a function of frequency. Because optical devices possess polarization characteristics, injecting orthogonally polarized interrogation beams into the device yields information about the device in two different polarization directions, thus obtaining the device's polarization characteristics. Since there is a certain time delay between the two polarization states of the interrogation beams, they are temporally separated in the detection signal and can be distinguished. This is why a certain time delay needs to be introduced between the two orthogonally polarized interrogation beams in this method. Polarization diversity detection requires aligning the input polarization state to maximize its alignment with the detection system to ensure the test signal is maximized. The interference signals in the p and s polarization directions are respectively:

[0017]

[0018]

[0019] The transmission matrix of the device is H = [h ij ] 2×2 β and γ are the polarization alignment angles of the orthogonal polarization generation module 2 and the test interferometer module 4, respectively. We found that the first peak has the largest amplitude when γ = 45°, indicating that calibration can also be performed when the device under test 402 is connected. The only additional step is filtering to extract the signal in the frequency domain, which corresponds to the first peak in the spatial domain. Therefore, online polarization alignment can be achieved by feeding back the fringe amplitude. In addition, rapid iteration using intelligent algorithms is required to achieve OVA with online polarization alignment.

[0020] Bayesian optimization fits the objective function using a probabilistic surrogate model and selects the most likely point for evaluation based on the fitting result. By considering the points from previous iterations and continuously experimenting, it can determine the optimal parameters of the objective more quickly. The Bayesian algorithm is an excellent global optimization algorithm that, compared to other optimization algorithms, focuses more on reducing the cost of evaluation, requiring fewer iterations, and is faster. The most important core idea of ​​Bayesian optimization is "Bayes' theorem":

[0021]

[0022] Where f represents the unknown objective function, and D 1：t Let p(f) represent the observed set, and p(f[D] represent the prior probability distribution of f. 1：t The expression denoted by represents the posterior probability distribution of the objective function. Combining this with Bayesian optimization enables rapid iterative polarization alignment. However, the objective function for polarization alignment contains numerous local extrema, significantly slowing the convergence speed of the Bayesian optimization algorithm. The two fringe amplitudes used for alignment are transformed by the cotangent and tangent functions, respectively. Using the transformed data as the objective function improves the algorithm's convergence speed.

[0023] The Bayesian optimization algorithm process is as follows:

[0024] Step 1 801: Preprocess the fringe amplitude data as the objective function f; according to the requirements of orthogonal polarization alignment and maximizing polarization alignment, perform cotangent function and tangent function transformations on the objective function respectively;

[0025] Step 2 802: Randomly generate n initialization points within the polarization angle range; after adjusting the polarization angles of the first polarization controller 204 and the second polarization controller 404, obtain the amplitude of the orthogonal polarization interference fringes 301a and the amplitude of the first interference peak 603b under different polarization angles;

[0026] Step 3 803: Establish the surrogate model using the probabilistic surrogate model function; the probabilistic surrogate function is a Gaussian process (GP) regression, which estimates the mean μ(x) and variance σ(x) of other points based on the searched points;

[0027] Step 4 804: The sampling function explores other point values; based on the model's predicted distribution, the sampling function explores a large number of point values ​​to select the next sampling point; this sampling function is the upper confidence interval function (UCB), UCB = μ(x) + kσ(x), where k is the upper confidence interval parameter;

[0028] Step 5.805: Determine the standard deviation (std(f)) of the function values ​​at ten consecutive sampling points. i ~f i+10 Is it less than the current maximum value f? i(max)1%; if yes, proceed to step seven 806; if no, return to step three 803 using sampling point set iteration g05a, and then repeat step three 803 to step five 805.

[0029] Step 6.806: Output the target fringe amplitude and the corresponding polarization adjustment angle;

[0030] After achieving polarization alignment and demodulating to obtain the Jones matrix of the device, the insertion loss (IL) is calculated using the following formula:

[0031]

[0032] The formula for calculating group delay (GD) is:

[0033]

[0034] The formula for calculating polarization-dependent loss (PDL) is:

[0035]

[0036] IL max and IL min These are the eigenvalues ​​of the matrix obtained by multiplying the Jones matrix of the device by its conjugate transpose.

[0037] The formula for calculating polarization dispersion (PMD) is as follows:

[0038]

[0039] Among them, ρ1 and ρ2 are H(ω+Δω)H(ω) -1 eigenvalues.

[0040] The advantages of this invention are:

[0041] 1) This invention is a Bayesian-optimized online polarization alignment light vector characteristic testing device, which can realize high-precision testing of the vector characteristics of devices in a large wavelength measurement range. It has a simple structure and uses online polarization alignment to suppress polarization misalignment during the measurement process, effectively suppressing the measurement error caused by polarization misalignment and improving the reliability of measurement.

[0042] 2) The Bayesian optimization algorithm is used to achieve polarization alignment through rapid iteration. The cotangent and tangent functions are transformed on the objective function to highlight the target extremum points, which can effectively avoid getting trapped in local extrema and improve the convergence speed of the optimization algorithm. Attached Figure Description

[0043] Figure 1 This is a diagram of an online polarization alignment light vector analyzer based on Bayesian optimization.

[0044] Figure 2These are the interference peak images after fast Fourier transform of the P and S detectors.

[0045] Figure 3 This is a flowchart of the process for acquiring information about the device under test.

[0046] Figure 4 This is a flowchart of Bayesian optimized polarization alignment.

[0047] Figure 5 This is a comparison of the insertion loss results of the test cell before and after calibration using Bayesian-optimized online polarization alignment light vector analysis and the LUNA OVA5100 instrument.

[0048] Figure 6 This is a magnified comparison of the insertion loss results of the test cell after Bayesian-optimized online polarization alignment light vector analysis and calibration with the LUNA OVA5100 instrument.

[0049] Figure 7 Comparison of Bayesian-optimized online polarization alignment light vector analysis and the polarization correlation loss results of the test cell before and after calibration with the LUNA OVA5100 instrument. Detailed Implementation

[0050] To clearly illustrate the online polarization alignment light vector analysis device of the present invention, the present invention will be further described in conjunction with the embodiments and accompanying drawings, but this should not be construed as limiting the scope of protection of the present invention.

[0051] Using a gas absorption chamber (primarily containing acetylene) as the device under test, the main optoelectronic components and their parameters are as follows:

[0052] The tunable laser source 101 is a Keysight 81606A laser. When polarization alignment is used to acquire fringes, the wavelength tuning range is 0.1 nm and the sweep rate is 200 nm / s. When testing devices, the wavelength tuning range is 1525 nm to 1545 nm and the sweep rate is 40 nm / s.

[0053] The data acquisition card has a sampling rate of 2MHz / s, is triggered via LabVIEW software, and has a maximum acquisition voltage of 2.5V.

[0054] The first coupler 201, the second coupler 202, the third coupler 401, and the fourth coupler 405 all have a splitting ratio of 50:50, an insertion loss of less than 0.5dB, and an operating wavelength of 1550nm.

[0055] The first polarization controller 204 and the second polarization controller 404LUNA MPC202 have an adjustment speed of 65μs each time.

[0056] The arm length difference of orthogonal polarization generation module 2 is 2m, and the arm length difference of test interferometer module 4 is 10m;

[0057] The peak value of the stripes within the range of approximately 1550 nm ± 0.02 nm is used as an evaluation index for stripe amplitude.

[0058] like Figure 1 An online polarization-aligned optical vector analyzer (OVA) device includes a tunable laser source 101, an orthogonal polarization generation module 2, an orthogonal polarization interference fringe feedback module 3, a test interferometer module 4, a polarization diversity detection module 5, a test interference information feedback module 6, and a device-under-test (DUT) information acquisition module 7. The tunable laser source 101 generates linearly swept continuous light with a wavelength varying over time, which is injected into the orthogonal polarization generation module 2. The orthogonal polarization interference fringe feedback module 3 acquires data from the interference fringes output by the orthogonal polarization generation module 2. Through rapid iteration via a first feedback control module 303, the first polarization controller 204 located on the upper arm 201a of the interferometer is adjusted until the amplitude of the output interference fringes is minimized, achieving orthogonal polarization alignment. Then, a pair of orthogonal polarization interrogation beams 21 are output to the test interferometer module 4. The amplitude and phase of the orthogonal polarization interrogation beams 21 change according to the amplitude and phase of the DUT 402 located on the upper arm 401a of the test interferometer, and then an optical signal carrying the amplitude and phase information of the DUT 402 is output. The polarization diversity detection module 5 decomposes the optical signal output by the test interferometer module 4 into two perpendicular polarization directions.

[0059] The second interference fringe acquisition module 601 acquires the fringes of the p detector 502 and the s detector 503, and then performs a fast Fourier transform 602. The signal corresponding to the p detector 502 or the s detector 503 is input to the first interference peak information acquisition module 603. The first interference peak 603a is selected according to the time delay difference τ between the two arms of the orthogonal polarization generation module 2. Then the amplitude value 603b of the first interference peak is read. The amplitude value 603b of the first interference peak is then processed by the second amplitude trigonometric function transformation module 604 and then input to the second feedback control module 605. The second feedback control module 605 iterates rapidly to determine the next adjustment angle of the second polarization controller 404 located on the lower arm 401b of the test interferometer. The feedback adjustment is made to maximize the amplitude value 603b of the first interference peak, thereby maximizing polarization alignment. Finally, the p detector signal 701a and the s detector signal 701b after the fast Fourier transform 602 are input to the device under test information acquisition module (7) to obtain the transmission characteristics of the device.

[0060] The orthogonal polarization interference fringe feedback module 3 is characterized in that: the first interference fringe acquisition module 301 acquires data of the interference fringes of the orthogonal polarization generation module 2 and calculates the amplitude of the orthogonal polarization interference fringe 301a. Then, the amplitude of the orthogonal polarization interference fringe 301a is processed by the first amplitude trigonometric function transformation module 302 and then input into the first feedback control module 303 to determine the next adjustment angle of the first polarization controller 204.

[0061] The orthogonal polarization generation module 2 is characterized in that: a linearly swept continuous beam is injected into the first coupler 201 and split into two beams. One beam, as the first test beam m1, passes through the lower arm 201b of the interferometer with the first extended fiber 203 and is injected into the second input end 202b of the second coupler. The other beam, as the input reference beam m2, is injected into the upper arm 201a of the interferometer with the first polarization controller 204. The first polarization controller 204 is adjusted to rotate the polarization direction of the input reference beam m2 and inject it into the first input end 202a of the second coupler. Then, the two beams are combined in the second coupler 202 and input into the first detector 203 through the first output end 202c of the second coupler to cause interference. The interference fringes are input into the orthogonal polarization interference fringe feedback module 3 and fed back to minimize the amplitude 301a of the orthogonal polarization interference fringes. Finally, the first test beam m1 and the second test beam n1, which are orthogonal to the polarization direction of the first test beam m1, are output through the second output end 202d of the second coupler to the test interferometer module 4.

[0062] The first amplitude trigonometric transformation module (302) is characterized in that: the trigonometric transformation is cot(y1 / V) max ×π / 2), where y1 is the amplitude of the orthogonal polarization interference fringes (301a), V max This is the maximum voltage value of the data acquisition card;

[0063] The second amplitude trigonometric transformation module (604) is characterized in that: the trigonometric transformation is tan(y2 / V) max ×π / 2), where y2 is the amplitude of the first interference peak (603b);

[0064] The test interferometer module 4 is characterized in that: the orthogonally polarized light 21 is split into two beams by the third coupler 401. One beam is output from the upper arm 401a of the test interferometer through the device under test 402 to the first input terminal 405a of the fourth coupler. The other beam is injected through the lower arm 401b of the test interferometer, and then sequentially passes through the second extension fiber 403 and the second polarization controller 404 to the second input terminal 405b of the fourth coupler. The two beams are combined at the fourth coupler and then input to the polarization diversity detection module 5.

[0065] The first feedback control module 303 and the second feedback control module 605 are characterized in that: the feedback control algorithm is a Bayesian optimization algorithm;

[0066] The device under test information acquisition module 7 is characterized in that: after polarization alignment is achieved, the p detector information 701a and s detector information 701b after the fast Fourier transform 602 are input into the second and fourth interference peak filtering modules 701, and the p detector τ is selected. D -τ time delay difference correlation peak a and p detector τ D +τ delay difference correlation peak b, and s detector τ D -τ time delay difference correlation peak c and s detector τ D +τ Delay difference correlation peak d, τ D The test interferometer module 4 measures the time delay difference between the two arms. The peaks are filtered to obtain the relevant first real signal a1, second real signal b1, third real signal c1, and fourth real signal d1. Then, a Hilbert transform 702 is performed to obtain the first complex signal a2, second complex signal b2, third complex signal c2, and fourth complex signal d2. A Hilbert transform 702 is then performed, followed by Jones matrix demodulation 703 to obtain the device Jones matrix 703a. The optical parameters of the device under test (DUT) 402 are then calculated. These optical parameters include insertion loss, group delay, dispersion, polarization mode dispersion, and polarization-dependent loss.

[0067] Figure 5 This is a comparison of the cell insertion loss results before and after calibration using the Bayesian-optimized online polarization alignment OVA and the commercial instrument LUNA OVA5100. The insertion loss test result 901 after online polarization alignment is consistent with the insertion loss result 902 after calibration using LUNA OVA5100. At the same time, the comparison shows that due to polarization misalignment, the insertion loss result 903 before calibration by LUNA OVA5100 is deviated by nearly 2dB. Figure 6 This is a partial magnified comparison of the cell insertion loss results of the online polarization alignment OVA based on Bayesian optimization and the corrected results of the commercial instrument LUNA OVA5100. The magnified results of the online polarization alignment insertion loss test (905) and the corrected insertion loss magnified results (904) show basically the same trend, and the positions of the absorption peaks also match.

[0068] Figure 7 This is a comparison of the cell polarization-related loss results before and after calibration using the Bayesian-optimized online polarization alignment OVA and the commercial instrument LUNA OVA5100. The overall trend of the polarization-related loss test result 907 after online polarization alignment and the polarization-related loss result 908 after calibration by LUNA OVA5100 is consistent. At the same time, the comparison shows that due to polarization misalignment, the position of the polarization-related loss result 906 before calibration by LUNA OVA5100 deviates by nearly 2dB.

[0069] Experimental results show that this method can effectively suppress errors caused by polarization misalignment, greatly improve the accuracy of system measurements, and enhance the reliability of testing.

Claims

1. A light vector analyzer device for online polarization alignment, comprising a tunable laser source (101), an orthogonal polarization generation module (2), an orthogonal polarization interference fringe feedback module (3), a test interferometer module (4), a polarization diversity detection module (5), a test interference information feedback module (6), and a device-under-test information acquisition module (7), characterized in that: A tunable laser source (101) generates a linearly swept continuous light whose wavelength varies with time, which is injected into the orthogonal polarization generation module (2). The orthogonal polarization interference fringe feedback module (3) collects data on the interference fringes output by the orthogonal polarization generation module (2). The first feedback control module (303) iterates rapidly and adjusts the first polarization controller (204) located on the upper arm of the interferometer until the amplitude of the output interference fringes is minimized, thus achieving orthogonal polarization alignment. Then, a pair of orthogonal polarization interrogation beams are output to the test interferometer module (4). The amplitude and phase of the orthogonal polarization interrogation beams change with the amplitude and phase of the device under test located on the upper arm of the test interferometer, and then an optical signal carrying the amplitude and phase information of the device under test is output. The polarization diversity detection module (5) decomposes the optical signal output by the test interferometer module (4) into two perpendicular polarization directions. The second interference fringe acquisition module (601) acquires the fringes of the p detector (502) and the s detector (503), and then performs a fast Fourier transform. The signal corresponding to the p detector (502) or the s detector (503) is input into the first interference peak information acquisition module (603). The first interference peak is selected according to the time delay difference τ between the two arms of the orthogonal polarization generation module (2). The amplitude of the first interference peak is then read and processed by the second amplitude trigonometric function transformation module (604). The first interference peak amplitude is then input into the second feedback control module (605) for rapid iteration to determine the next adjustment angle of the second polarization controller (404) located in the lower arm of the test interferometer. The feedback adjustment is made to maximize the amplitude of the first interference peak to achieve maximum polarization alignment. Finally, the p detector signal and the s detector signal after the fast Fourier transform are input into the device under test information acquisition module (7) to obtain the transmission characteristics of the device. The first interference fringe acquisition module (301) acquires data of the interference fringes of the orthogonal polarization generation module (2) and calculates the amplitude of the orthogonal polarization interference fringes. The amplitude of the orthogonal polarization interference fringes is processed by the first amplitude trigonometric function transformation module (302) and input into the first feedback control module (303) to determine the next adjustment angle of the first polarization controller (204). The orthogonal polarization generation module (2) specifically includes: after the linearly swept continuous light is injected into the first coupler (201), it is split into two beams. One beam is used as the first test light and is injected into the second input end of the second coupler through the lower arm of the interferometer with the first extended fiber (). The other beam is used as the input reference light and is injected into the upper arm of the interferometer with the first polarization controller (204). The first polarization controller (204) is adjusted to rotate the polarization direction of the input reference light and is injected into the first input end of the second coupler. Then the two beams are combined in the second coupler and are input into the first detector through the first output end of the second coupler to cause interference. The interference fringes are input into the orthogonal polarization interference fringe feedback module (3) and fed back to minimize the amplitude of the orthogonal polarization interference fringes. Finally, the first test light and the second test light orthogonal to the polarization direction of the first test light are output through the second output end of the second coupler to the test interferometer module (4).

2. The optical vector analyzer device according to claim 1, characterized in that: The amplitude of the orthogonal polarization interference fringes is processed by the first amplitude trigonometric function transformation module (302), wherein the trigonometric transformation in the first amplitude trigonometric function transformation module is cot(y1 / V) max ×π / 2), where y1 is the amplitude of the orthogonal polarization interference fringes, V max This is the maximum voltage value of the data acquisition card.

3. The optical vector analyzer device according to claim 2, characterized in that: The amplitude of the first interference peak is then processed by the second amplitude trigonometric function transformation module (604), wherein the trigonometric transformation in the second amplitude trigonometric function transformation module is tan(y2 / V). max ×π / 2), where y2 is the amplitude of the first interference peak.

4. The optical vector analyzer device according to claim 3, characterized in that: The test interferometer module (4) specifically includes: the orthogonal polarization interrogation light is split into two beams by the third coupler, one beam is output from the upper arm of the test interferometer through the device under test to the first input end of the fourth coupler, and the other beam is injected through the lower arm of the test interferometer, and then sequentially through the second extended optical fiber and the second polarization controller to the second input end of the fourth coupler. The two beams are combined at the fourth coupler and then input into the polarization diversity detection module (5).

5. The optical vector analyzer device according to claim 1, characterized in that: After polarization alignment is achieved, the device under test information acquisition module (7) inputs the p detector information and s detector information after fast Fourier transform into the second interference peak filtering module and the fourth interference peak filtering module, and selects the p detector τ D -τ time delay difference correlation peak and p detector τ D +τ time delay difference correlation peak, and s detector τ D -τ time delay difference correlation peak and s detector τ D +τ time delay difference related peak, τ D The time delay difference between the two arms of the interferometer module (4) is tested. The correlation peak is filtered to obtain the first real signal, the second real signal, the third real signal and the fourth real signal. Then, Hilbert transform is performed to obtain the first complex signal, the second complex signal, the third complex signal and the fourth complex signal. Then, Jones matrix demodulation is performed to obtain the device Jones matrix. Finally, the optical parameters of the device under test are calculated based on the device Jones matrix. The optical parameters of the device under test include the insertion loss, group delay, dispersion, polarization mode dispersion and polarization correlation loss of the device under test.

Images