A method for demodulating an electrical signal of a single-source single-axis fiber-optic gyroscope

Abstract

The application discloses a kind of single light source single-axis fiber-optic gyroscope electric signal demodulation method, it is related to quantum optics precision measurement technical field, including, the joint probability density function model of quantum fluctuation noise is constructed, through the dynamic correction measured noise parameter of bayesian inference, generate the noise model containing shot noise variance and thermal noise mean;Utilize Y waveguide beam splitting characteristics to generate multiple groups of reference-measurement optical path pairs, calculate the cross-correlation coefficient of interference signal of each group of optical path pairs, and select the two groups of high correlation signal pairs with the highest correlation degree;According to high-precision phase trajectory, based on the phase-velocity conversion relationship of Sagnac effect, the Sagnac phase shift formula is applied to calculate the rotation angular velocity value and output.The application realizes the efficient suppression of shot noise and thermal noise by dynamic noise modeling and parameter adaptive adjustment mechanism, significantly improves the integrity and accuracy of the phase signal after noise reduction.

Description

Technical Field

[0001] This invention relates to the field of quantum optics precision measurement technology, and in particular to a method for demodulating electrical signals in a single-source, single-axis fiber optic gyroscope. Background Technology

[0002] Quantum fluctuation noise, as the background noise in quantum optical systems, presents a key challenge in the field of high-precision phase measurement due to its accurate modeling and suppression. In recent years, research on the statistical properties of noise based on quantum mechanics has deepened, and joint probability density function models have made progress in describing the combined characteristics of shot noise and thermal noise. Meanwhile, multi-path interferometry, through beam splitters to construct reference-measurement optical path pairs, provides a new experimental method for noise correlation analysis. Furthermore, adaptive filtering algorithms, such as the combination of wavelet transform and FIR filters, have shown broad application potential in suppressing electrical signal noise.

[0003] However, existing technologies still have limitations in dynamic noise modeling and multi-path signal utilization. Traditional noise models mostly rely on static parameters or empirical fitting, which makes it difficult to reflect the dynamic changes of quantum fluctuation noise in real time, resulting in insufficient targeting of noise suppression. At the same time, the high correlation characteristics of multi-path interference signals have not been fully explored, and conventional screening methods are prone to introducing correlation errors from fiber perturbation noise, which restricts further improvement in phase extraction accuracy. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope to solve the problems of insufficient dynamic adaptability of existing quantum noise models and insufficient utilization of the high correlation characteristics of multi-optical-path signals, which leads to limited phase extraction accuracy.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope, comprising: constructing a joint probability density function model of quantum fluctuation noise; dynamically correcting the measured noise parameters through Bayesian inference to generate a noise model containing the variance of shot noise and the mean of thermal noise; generating multiple sets of reference-measurement optical path pairs using the beam splitting characteristics of the Y-waveguide; calculating the cross-correlation coefficient of the interference signals of each optical path pair; selecting the two highly correlated signal pairs with the highest correlation; based on the noise model, adjusting the wavelet threshold and FIR filter bandwidth parameters to suppress shot noise and thermal noise in the output electrical signal of the photodetector, and outputting the denoised phase signal; performing a weighted average on the two selected highly correlated signal pairs, with the weight being the cross-correlation coefficient of the interference signals of the corresponding optical path pairs, to extract a high-precision phase trajectory after eliminating fiber disturbance noise; and calculating and outputting the rotational angular velocity value based on the phase-angular velocity conversion relationship of the Sagnac effect using the Sagnac phase shift formula based on the high-precision phase trajectory.

[0008] As a preferred embodiment of the electrical signal demodulation method for the single-source single-axis fiber optic gyroscope described in this invention, the specific steps for constructing the joint probability density function model of quantum fluctuation noise are as follows:

[0009] The sources of quantum fluctuation noise in a single-source, single-axis fiber optic gyroscope were analyzed, and the independent characteristics of shot noise and thermal noise were determined.

[0010] Based on quantum optics theory, shot noise is assumed to follow a Poisson distribution, and thermal noise follows a Gaussian distribution.

[0011] By combining the statistical independence characteristics of shot noise and thermal noise, a joint probability density function model is established;

[0012] Verify the goodness of fit between the joint probability density function model and the actual noise measurement data.

[0013] As a preferred embodiment of the electrical signal demodulation method for the single-source single-axis fiber optic gyroscope described in this invention, the step of generating a noise model containing shot noise variance and thermal noise mean by dynamically correcting the measured noise parameters through Bayesian inference is as follows:

[0014] Acquire time-series data of measured interference signals from fiber optic gyroscopes;

[0015] Based on the joint probability density function model, initial estimates of shot noise variance and thermal noise mean are set, and the posterior probability distribution is updated using Bayesian inference method combined with time series data.

[0016] The corrected variance of shot noise and the mean of thermal noise are extracted from the posterior probability distribution to generate a noise model.

[0017] As a preferred embodiment of the electrical signal demodulation method for the single-source single-axis fiber optic gyroscope described in this invention, the Y-waveguide beam splitting characteristic refers to the characteristic of the Y-waveguide to split the input light into two beams of optical signals with orthogonal polarization directions based on the birefringence effect, which are then transmitted through different paths and then combined for output.

[0018] The specific steps for selecting the two highly correlated signal pairs are as follows:

[0019] Multiple pairs of optical signals between the reference optical path and the measurement optical path are generated by utilizing the beam splitting characteristics of the Y-waveguide, and the cross-correlation coefficient is calculated for the interference signal of each pair of optical paths.

[0020] Sort the optical path pairs by cross-correlation coefficient from highest to lowest, and select the top two groups with the highest cross-correlation coefficient.

[0021] As a preferred embodiment of the electrical signal demodulation method for the single-source single-axis fiber optic gyroscope described in this invention, the wavelet threshold refers to a critical value for selecting signal coefficients that is adaptively determined based on noise intensity and signal sparsity; the FIR filter bandwidth parameter refers to the filter boundary frequency set according to the frequency range of the noise to be suppressed.

[0022] As a preferred embodiment of the electrical signal demodulation method for the single-source single-axis fiber optic gyroscope described in this invention, the specific steps for outputting the noise-reduced phase signal are as follows:

[0023] The electrical signal output by the photodetector is processed using wavelet thresholding for soft thresholding.

[0024] The electrical signal after soft thresholding is input into an FIR filter, low-pass filtered according to the set bandwidth parameters, and continuous phase change information is extracted by Hilbert transform, which is output as the phase signal after denoising by wavelet thresholding and FIR filter.

[0025] As a preferred embodiment of the electrical signal demodulation method for the single-source single-axis fiber optic gyroscope described in this invention, the specific steps for extracting the high-precision phase trajectory after eliminating fiber optic disturbance noise are as follows:

[0026] For the two selected pairs of highly correlated signals, the phase signals after joint noise reduction are extracted respectively, and the two phase signals are weighted and averaged using the cross-correlation coefficient of the two pairs of signals as the weight.

[0027] The phase signal after weighted averaging is filtered by moving average to output a high-precision phase trajectory after eliminating fiber optic disturbance noise.

[0028] As a preferred embodiment of the electrical signal demodulation method for the single-source, single-axis fiber optic gyroscope described in this invention, the specific steps for calculating and outputting the rotational angular velocity value using the Sagnac phase shift formula are as follows.

[0029] The phase change per unit time is extracted from the high-precision phase trajectory, and the rotational angular velocity value is calculated based on the phase-angular velocity conversion relationship of the Sagnac effect.

[0030] Temperature drift compensation and device nonlinearity calibration are performed on the rotational angular velocity value to correct system errors and output the calibrated rotational angular velocity value.

[0031] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the electrical signal demodulation method for a single-source single-axis fiber optic gyroscope as described in the first aspect of the present invention.

[0032] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the electrical signal demodulation method for a single-source single-axis fiber optic gyroscope as described in the first aspect of the present invention.

[0033] The beneficial effects of this invention are as follows: By constructing a joint probability density function model of quantum fluctuation noise and using Bayesian inference to dynamically correct measured noise parameters, the dynamic changes in the variance of shot noise and the mean of thermal noise can be captured in real time, providing accurate statistical basis for adjusting wavelet threshold and FIR filter bandwidth parameters. Compared with traditional fixed-parameter filtering methods, this dynamic adjustment mechanism can significantly improve the targeting of noise suppression. While effectively suppressing shot noise and thermal noise, it preserves the key features of the phase signal to the greatest extent, thereby significantly improving the integrity and accuracy of the phase signal after noise reduction, laying a reliable foundation for subsequent high-precision phase trajectory extraction. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart of the electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope.

[0036] Figure 2 A flowchart for constructing a noise model.

[0037] Figure 3 This is a schematic diagram of the generation and filtering of optical path pairs.

[0038] Figure 4This is a flowchart of signal processing. Detailed Implementation

[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0040] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0041] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0042] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides an electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope, comprising the following steps:

[0043] S1: Construct a joint probability density function model of quantum fluctuation noise, dynamically correct the measured noise parameters through Bayesian inference, and generate a noise model that includes the variance of shot noise and the mean of thermal noise.

[0044] S.1.1: Analyze the sources of quantum fluctuation noise in a single-source, single-axis fiber optic gyroscope and determine the independent characteristics of shot noise and thermal noise.

[0045] Specifically, based on quantum optics theory, the optical link of a single-source, single-axis fiber optic gyroscope (source → coupler → fiber optic loop → detector) is analyzed to identify the main sources of quantum fluctuation noise (shot noise and thermal noise).

[0046] It should be noted that shot noise is caused by the quantum properties of the output light field of a light source (such as a laser), and manifests as random fluctuations in photon counting, belonging to the category of quantum noise; thermal noise is caused by inelastic scattering (such as Rayleigh scattering) and lattice vibration (phonon interaction) induced by the thermal motion of molecules in the optical fiber material, and belongs to classical thermal disturbance noise.

[0047] Furthermore, the photoelectric signal output by the gyroscope is collected by a photodetector, and the noise component is separated by frequency domain filtering.

[0048] Specifically, when shot noise is concentrated in the high-frequency band (on the same order of magnitude as the optical carrier frequency), high-frequency noise samples are extracted using a bandpass filter; when thermal noise is concentrated in the low-frequency band (limited by the optical fiber thermal diffusion time constant), low-frequency noise samples are extracted using a low-pass filter.

[0049] It should be noted that the extracted high-frequency noise samples correspond to shot noise samples, and the low-frequency noise samples correspond to thermal noise samples (because shot noise is strongly correlated with the optical carrier frequency, and thermal noise is dominated by thermal diffusion, the two frequency bands do not overlap and can be completely separated by filtering).

[0050] Calculate the covariance between the shot noise sample and the thermal noise sample. If the covariance approaches zero (for example, the covariance calculated for 1000 samples is 0.002, which is much smaller than the mean of the product of the sample standard deviations of 0.1), then the two are confirmed to be statistically independent.

[0051] S1.2: Based on quantum optics theory, shot noise is assumed to follow a Poisson distribution, and thermal noise follows a Gaussian distribution.

[0052] Specifically, according to the coherent state theory of quantum mechanics, quantum fluctuations in the light field cause the photon count to follow a Poisson distribution. For monochromatic light with an average photon count rate of λ, the probability mass function of shot noise is expressed as:

[0053]

[0054] In the formula, P(n; λ) represents the probability mass function of shot noise, indicating the probability that the photon count is n given an average photon count rate of λ, where n is the discretized photon count value (a non-negative integer, n = 0, 1, 2, ...), λ represents the average photon count rate, and e represents the natural constant. -λ λ represents the exponentially decaying term, reflecting the normalization characteristics of the Poisson distribution. n The term represents the nth power of the average photon count rate, which is proportional to the nth power of the photon count. n! represents the factorial of n, used to normalize the probability distribution.

[0055] Based on the central limit theorem of classical statistical physics, thermal noise (such as phase perturbations) caused by molecular thermal motion in optical fibers is a superposition of numerous independent thermal vibrations, following a Gaussian distribution. The probability density function of the thermal noise is expressed as:

[0056]

[0057] In the formula, f(t; μ,σ) 2 Let σ be the probability density function of the thermal noise, representing the probability density (continuous variable) of the thermal noise taking the value t, where t is the instantaneous value of the thermal noise, μ is the mean of the thermal noise, and σ is the mean of the thermal noise. 2 The variance of thermal noise. π is a mathematical constant, representing the arithmetic square root of 2 multiplied by π. Denotes the normalization coefficients of the Gaussian distribution, ensuring that the probability density integral over the entire domain is 1, 2σ. 2 The scaling factor (t-μ) represents the variance correlation and is used to control the decay rate. 2 This represents the squared deviation of the thermal noise value t from the mean value μ. The core term representing the Gaussian distribution is used to describe the exponential decay of noise values ​​around the mean μ.

[0058] S1.3: Combining the statistical independence characteristics of shot noise and thermal noise, a joint probability density function model is established.

[0059] Specifically, since shot noise (discrete) and thermal noise (continuous) are statistically independent, the joint probability density function model is the product of their probability distributions (the joint distribution of discrete and continuous variables needs to be correlated through integration).

[0060] Assuming the total noise is a linear superposition of the two, the joint probability density function model is expressed as:

[0061]

[0062] N = N s +N t ;

[0063] In the formula, p(N) represents the joint probability density function of the total noise of quantum fluctuations, N is the linear superposition of the two, and represents the instantaneous value of the total noise. s N represents shot noise. t Indicates thermal noise. P(N) represents the infinite summation of the discretized photon count values. s =n) represents the probability that the shot noise takes the value n. This represents the probability density of thermal noise being Nn when the total noise is N.

[0064] S1.4: Verify the goodness of fit between the joint probability density function model and the actual noise measurement data.

[0065] Specifically, under different environmental conditions, the gyroscope output photoelectric signals are collected, noise samples are extracted and standardized (DC component is subtracted, AC noise is retained);

[0066] For each sample group, the empirical mean of shot noise, the empirical mean of thermal noise, and the empirical variance are calculated respectively.

[0067] The Kolmogorov-Smirnov test was used to evaluate the consistency between the predicted distribution and the empirical distribution of the joint probability density function model:

[0068] Specifically, for shot noise, calculate the maximum value D of the cumulative distribution function difference between the theoretical Poisson distribution and the empirical distribution. s For thermal noise, calculate the maximum value D of the cumulative distribution function difference between the theoretical Gaussian distribution and the empirical distribution. t ;

[0069] If D s <Critical value (e.g., when α = 0.05, the critical value is When N = 1000, the critical value is approximately 0.043, and D... t If the threshold is less than the critical value, the joint probability density function model fits the actual data well. Here, α represents the significance level, used to control the probability of misclassification (the smaller α is, the larger the critical value, and the more rigorous the test).

[0070] It should be noted that it is necessary to confirm whether the Kolmogorov-Smirnov test is directly applicable to shot noise (discrete type). If not, the chi-square test or other methods should be used instead. Specifically, for discrete shot noise, the chi-square test is used to evaluate the consistency between the theoretical Poisson distribution and the empirical distribution; for continuous thermal noise, the Kolmogorov-Smirnov test is used.

[0071] S1.5: Obtain time series data of the measured interference signal from the fiber optic gyroscope.

[0072] Specifically, the fiber optic gyroscope is placed in a stable test environment (such as a constant temperature chamber to control temperature fluctuations to less than 0.1℃, and a vibration isolation platform to reduce mechanical vibration interference) and connected to the output terminal and a high-speed data acquisition device.

[0073] Furthermore, the sampling rate and collection duration of the data acquisition equipment can be set;

[0074] For example, setting the sampling rate to 100kHz covers the high-frequency characteristics of shot noise and the low-frequency characteristics of thermal noise; the acquisition duration is 30 seconds to ensure that the acquired interference signal time series contains sufficient noise samples.

[0075] It should be noted that during the acquisition process, the fiber optic gyroscope should be kept in normal working condition (such as stable laser output power and uniform fiber optic ring temperature) to avoid external electromagnetic interference or mechanical shock.

[0076] After acquisition, the DC offset is removed by moving average method, and then a bandpass filter is used to retain the signal components related to shot noise (high frequency band) and thermal noise (low frequency band) to obtain clean time series data of the interference signal.

[0077] S1.6: Based on the joint probability density function model, initial estimates of the variance of shot noise and the mean of thermal noise are set, and the posterior probability distribution is updated using Bayesian inference methods combined with time series data.

[0078] Specifically, based on the established joint probability density function model (shot noise follows a Poisson distribution, and thermal noise follows a Gaussian distribution), initial estimates are set.

[0079] It should be noted that the initial value of the shot noise variance is obtained through theoretical calculation (the shot noise variance is equal to the laser output power divided by the energy of a single photon), and the initial value of the thermal noise mean is determined through the theoretical relationship between fiber temperature, length and thermal noise coefficient (the thermal noise mean is equal to the product of the thermal diffusion related parameters and temperature and length).

[0080] For the variance of shot noise, the inverse gamma distribution is chosen as the prior (conjugate prior, which is convenient for calculation), and the parameters are set as weak information priors (such as shape parameter 2 and scale parameter 1) to reflect the initial assumptions about the variance of shot noise; for the mean of thermal noise, the normal distribution is chosen as the prior, and the mean is the initial estimate.

[0081] Treating time series data as independent and identically distributed samples, the likelihood function is the product of probabilities of the joint probability density function model under the observed data (i.e., the product of the probability density functions of the noise values ​​at each time point that conform to the Poisson or Gaussian distribution).

[0082] It should be noted that the likelihood function is a statistical function used to evaluate the degree of fit between model parameters and observed data. In this context, the likelihood function represents the product of probabilities of the joint probability density function model under given observed time series data; that is, the joint probability of the observed data occurring when the model parameters (such as the variance of shot noise and the mean of thermal noise) are known. Time series data refers to the continuous sequence of sampled values ​​of the measured interferometric signal from a fiber optic gyroscope over a period of time, acquired through high-speed data acquisition equipment (such as an analog-to-digital converter).

[0083] It should also be noted that the continuous sample value sequence contains noise samples under the combined effects of shot noise and thermal noise. Each sampling point corresponds to a noise measurement value at a specific time. The sample values ​​are arranged in chronological order to form a sequence, which serves as the basic observation data for constructing the likelihood function in subsequent Bayesian inference.

[0084] Iterative sampling is performed using the Markov chain Monte Carlo algorithm (specifically, the Metropolis-Hastings sampler): starting from the initial parameter values, candidate parameter values ​​are generated, and the log ratio of the target distribution (the product of prior and likelihood) is calculated.

[0085] It should be noted that the initial parameter values ​​refer to the initial variance of the shot noise and the initial mean of the thermal noise.

[0086] For example, valid samples are retained with an acceptance rate of approximately 23.4%. After 10,000 iterations, the first 2,000 warm-up samples are removed (to ensure distribution convergence), and the remaining 8,000 iterations are used for posterior distribution estimation.

[0087] S1.7: Extract the corrected variance of shot noise and the mean of thermal noise from the posterior probability distribution to generate a noise model.

[0088] Specifically, correction values ​​for shot noise variance and thermal noise mean are extracted from the posterior distribution samples obtained by MCMC sampling.

[0089] Among them, the posterior mean is selected as the correction value (i.e., the arithmetic mean of all retained samples).

[0090] The corrected parameters are substituted into the joint probability density function model to generate a noise model;

[0091] The generated noise model is a joint probability density function model that includes the corrected variance of shot noise and the mean of thermal noise, which can be used for subsequent noise characteristic analysis or gyroscope performance prediction.

[0092] S2: Utilize the beam splitting characteristics of the Y-waveguide to generate multiple sets of reference-measurement optical path pairs, calculate the cross-correlation coefficient of the interference signals of each optical path pair, and select the two sets of highly correlated signal pairs with the highest correlation.

[0093] Among them, the Y-waveguide beam splitting characteristic refers to the characteristic of the Y-waveguide to split the input light into two beams of optical signals with orthogonal polarization directions based on the birefringence effect, which are then transmitted through different paths and then combined for output.

[0094] S2.1: Utilize the beam splitting characteristics of the Y-waveguide to generate multiple pairs of optical signals between the reference optical path and the measurement optical path, and calculate the cross-correlation coefficient for the interference signal of each pair of optical paths.

[0095] Specifically, based on the birefringence effect of the Y-waveguide, the polarization state of the input light is set by a polarization controller, so that the Y-waveguide splits the input light into two orthogonally polarized beams.

[0096] Configure the first optical path: Transmit the first orthogonal beam through the first transmission path of the fiber optic ring, and the second orthogonal beam through the second transmission path of the fiber optic ring. The two beams are combined at the detector end to generate the first set of interference signals. Keeping the Y-waveguide beam splitting structure unchanged, adjust the polarization controller parameters in the fiber optic ring to configure the second optical path: Swap the transmission paths of the two orthogonal beams, so that the first beam passes through the second transmission path and the second beam passes through the first transmission path. After combining, the second set of interference signals is generated.

[0097] Repeat the configuration steps of the first and second optical paths, and generate N optical path pairs by changing the combination of polarization controller parameters. Synchronously acquire the interference signal of each optical path pair to obtain time series data.

[0098] Based on time series data, the cross-correlation coefficient is obtained by calculating the integral value of the product of each set of reference signals and measured signals under different time delays, and then dividing it by the product of the square roots of the energy of the two signals (classical normalized cross-correlation analysis).

[0099] S2.2: Sort the optical path pairs by cross-correlation coefficient from high to low, and select the first two groups with the largest cross-correlation coefficient.

[0100] Specifically, the cross-correlation coefficients of the N groups of optical path pairs are sorted in descending order of numerical value, and the top two optical path pairs are selected as highly correlated signal pairs, denoted as P1 and P2; where N refers to the total number of optical path pairs generated by changing the combination of polarization controller parameters.

[0101] Extract the reference signal and measurement signal corresponding to P1, and the reference signal and measurement signal corresponding to P2;

[0102] Verify whether the cross-correlation coefficients of P1 and P2 satisfy R1 > 0.8 * R max If the condition is not met, the optical path pair is regenerated and the cross-correlation coefficient is calculated.

[0103] Among them, R max R1 represents the highest cross-correlation number among the N groups of optical path pairs, R1 represents the first highly correlated signal pair after sorting (i.e., the group of highly correlated signal pairs with the highest correlation), and 0.8 represents the preset ratio threshold constant (fixed value 0.8). The filtered P1 and P2 are output as highly correlated signal pairs and enter the subsequent weighted average processing flow.

[0104] S3: Based on the noise model, adjust the wavelet threshold and FIR filter bandwidth parameters to suppress shot noise and thermal noise in the output electrical signal of the photodetector, and output the noise-reduced phase signal.

[0105] Among them, the wavelet threshold refers to the critical value used for selecting signal coefficients, which is adaptively determined based on noise intensity and signal sparsity; the FIR filter bandwidth parameter refers to the filter boundary frequency set according to the frequency range of the noise to be suppressed.

[0106] S3.1: The electrical signal output by the photodetector is processed using wavelet thresholding for soft thresholding.

[0107] Specifically, the input electrical signal is decomposed into multi-scale wavelet coefficients to obtain wavelet coefficients in different frequency bands; based on the real-time estimates of the variance of shot noise and the mean of thermal noise in the noise model, the adaptive threshold of the wavelet coefficients in each frequency band is calculated.

[0108] It should be noted that the adaptive threshold calculation process for wavelet coefficients in each frequency band is as follows: Based on the real-time estimates of the shot noise variance and the thermal noise mean in the noise model, the adaptive threshold is determined according to the noise energy distribution ratio for wavelet coefficients in different frequency bands. Specifically, for high-frequency wavelet coefficients (mainly containing shot noise components), the adaptive threshold is proportional to the shot noise variance; for low-frequency wavelet coefficients (mainly containing thermal noise components), the adaptive threshold is proportional to the thermal noise mean. By dynamically adjusting the proportional coefficient, it is ensured that the adaptive threshold for each frequency band can effectively suppress noise while avoiding excessive smoothing of signal details.

[0109] Wavelet coefficients above the adaptive threshold are processed by soft thresholding (the absolute value of the adaptive threshold is subtracted from the wavelet coefficient value and the sign is retained), while wavelet coefficients below the adaptive threshold are directly set to zero. The processed wavelet coefficients are then restored to the denoised time-domain signal using a wavelet reconstruction algorithm.

[0110] S3.2: Input the electrical signal after soft thresholding into the FIR filter, perform low-pass filtering according to the set bandwidth parameters, and extract continuous phase change information through Hilbert transform, which is then used as the phase signal output after joint noise reduction by wavelet thresholding and FIR filter.

[0111] Specifically, the cutoff frequency of the FIR filter is set based on the frequency characteristics of thermal noise (concentrated in the low frequency band) and shot noise (of the same order of magnitude as the optical carrier) in the noise model.

[0112] For example, set the thermal noise cutoff frequency to 0.1 times the optical carrier frequency and the shot noise cutoff frequency to 0.5 times the optical carrier frequency.

[0113] The time-domain signal is input into the FIR filter, and the low-pass filter coefficients are generated using the window function method (such as the Kaiser window) to avoid introducing additional distortion or amplitude distortion of the phase signal during the filtering process, and to ensure that the phase trajectory extracted by the subsequent Hilbert transform is continuous and smooth.

[0114] Perform a Hilbert transform on the filtered time-domain signal to construct an analytic signal;

[0115] The phase trajectory of the analytic signal is extracted by the arctangent function, phase jumps are eliminated, and a continuous phase signal is output.

[0116] S4: Perform a weighted average on the two selected highly correlated signal pairs, with the weight being the cross-correlation coefficient of the interference signals of the corresponding optical path pairs, and extract the high-precision phase trajectory after eliminating fiber disturbance noise.

[0117] S4.1: For the two highly correlated signal pairs selected, extract the phase signals after joint noise reduction, and perform a weighted average of the two phase signals using the cross-correlation coefficient of the two signals as the weight.

[0118] Specifically, the combined denoised phase signals of the highly correlated signal pairs corresponding to P1 and P2 are extracted respectively, and denoted as Phase1 and Phase2.

[0119] Based on the calculated cross-correlation coefficients R1 and R2 (R1 is the cross-correlation coefficient of P1, and R2 is the cross-correlation coefficient of P2), calculate the weighting coefficients: the weighting coefficient of P1 is R1 / (R1+R2); the weighting coefficient of P2 is R2 / (R1+R2).

[0120] Multiply Phade1 and Phase2 by their respective weighting coefficients to obtain the weighted phase signal components. Sum the weighted phase signal components to generate the final weighted average phase signal.

[0121] S4.2: Perform moving average filtering on the weighted average phase signal to output a high-precision phase trajectory after eliminating fiber optic disturbance noise.

[0122] Specifically, set the window size for the moving average filter;

[0123] For example, the number of sampling points covered by the window can be adjusted according to the time scale of fiber optic disturbance noise.

[0124] The weighted average phase signal is sequentially input into the moving average filter, and the arithmetic mean of the phase values ​​of all sampling points within the window is calculated.

[0125] Slide the window along the time axis by one sampling point, repeatedly calculate the average value until the entire phase signal is covered, and output the phase signal after moving average processing as a high-precision phase trajectory.

[0126] It should be noted that the high-precision phase trajectory has eliminated the influence of fiber optic disturbance noise.

[0127] S5: Based on the high-precision phase trajectory and the phase-angular velocity conversion relationship based on the Sagnac effect, the rotational angular velocity value is calculated and output using the Sagnac phase shift formula.

[0128] S5.1: Extract the phase change per unit time from the high-precision phase trajectory, and calculate the rotational angular velocity value based on the phase-angular velocity conversion relationship of the Sagnac effect.

[0129] Specifically, numerical differentiation is performed on the high-precision phase trajectory to calculate the instantaneous phase change rate (i.e., the phase change per unit time) of adjacent phase sampling points. );

[0130] Based on the phase-angular velocity conversion relationship of the Sagnac effect, a mathematical relationship between the phase change rate and the rotational angular velocity is established, expressed as:

[0131]

[0132] In the formula, 8π represents the total phase change caused by rotation when light propagates in an optical fiber ring. 8π represents a constant term derived from the physical derivation of the Sagnac effect. U represents the number of turns in the optical fiber ring. A represents the effective area of ​​a single turn of the optical fiber ring, i.e., the area of ​​the planar region enclosed by a single loop of optical fiber. λ' represents the wavelength of the optical carrier, i.e., the wavelength of the light wave emitted by the light source in free space. c is the speed of light in vacuum, a universal constant. Ω represents the angular velocity of rotation, i.e., the angular rate at which an object rotates about its axis of rotation. Δt represents the time interval, which is the length of the time window used to calculate the phase change.

[0133] The amount of phase change per unit time The measured value is substituted into the mathematical relationship between the phase change rate and the rotational angular velocity to calculate the real-time rotational angular velocity value.

[0134] S5.2: Perform temperature drift compensation and device nonlinearity calibration on the rotational angular velocity value, correct system errors, and output the calibrated rotational angular velocity value.

[0135] Specifically, the internal temperature of the fiber optic gyroscope is monitored in real time by integrating a temperature sensor, and the compensation rules for temperature-phase error are determined based on the correspondence between temperature and phase error.

[0136] For example, polynomial fitting or piecewise linear compensation can be used to determine the correspondence.

[0137] Based on the nonlinear characteristic data calibrated by the device at the factory, calibration coefficients are generated using a lookup table method or a least squares fitting method. Temperature compensation and nonlinear correction calculations are then performed on the real-time rotational angular velocity values ​​in sequence to generate calibrated rotational angular velocity values.

[0138] Nonlinear characteristic data refers to the input-output correspondence data obtained during the device's factory calibration process, which reflects the deviation between the actual output value of the rotational angular velocity and the ideal linear output value.

[0139] It should be noted that temperature compensation and nonlinear correction calculations are performed sequentially on the real-time rotational angular velocity value. The process is as follows: Based on the real-time internal temperature data of the fiber optic gyroscope obtained by the integrated temperature sensor, the phase error compensation value at the current temperature is calculated according to the temperature-phase error correspondence determined in advance by polynomial fitting or piecewise linear compensation method. This phase error compensation value is then subtracted from the measured rotational angular velocity value. Based on the input-output characteristic data calibrated by the device at the factory, the calibration coefficient corresponding to the current rotational angular velocity value is found in the nonlinear calibration coefficient table generated by least squares fitting. The value after deducting the temperature error is then multiplied to eliminate the influence of the inherent nonlinear error of the device.

[0140] The calibrated rotational angular velocity values ​​are converted to units (e.g., from rad / s to ° / h) to output the final rotational angular velocity values ​​that meet the application requirements.

[0141] It should be noted that rad / s represents radians per second, which is a unit of angular velocity in the International System of Units (SI) used to measure angular velocity. It describes the number of radians an object rotates around its axis of rotation per unit time (per second). ° / h represents degrees per hour, which is a unit of angular velocity commonly used in engineering practice and specific application scenarios. It describes the number of angles an object rotates around its axis of rotation per unit time (per hour). The two can be converted to each other through angular unit conversion (1 radian is approximately equal to 57.3 degrees) and time unit conversion (1 hour is equal to 3600 seconds).

[0142] This embodiment also provides a computer device applicable to the electrical signal demodulation method of a single-source single-axis fiber optic gyroscope, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the electrical signal demodulation method of the single-source single-axis fiber optic gyroscope as proposed in the above embodiment.

[0143] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0144] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0145] In summary, this invention constructs a joint probability density function model of quantum fluctuation noise and utilizes Bayesian inference to dynamically correct measured noise parameters. This allows for real-time capture of the dynamic changes in the variance of shot noise and the mean of thermal noise, providing precise statistical data for adjusting wavelet thresholds and FIR filter bandwidth parameters. Compared to traditional fixed-parameter filtering methods, this dynamic adjustment mechanism significantly improves the targeting of noise suppression. While effectively suppressing shot noise and thermal noise, it maximizes the preservation of key phase signal features, thereby significantly improving the integrity and accuracy of the denoised phase signal and laying a reliable foundation for subsequent high-precision phase trajectory extraction.

[0146] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for demodulating electrical signals in a single-source, single-axis fiber optic gyroscope, characterized in that: include, A joint probability density function model of quantum fluctuation noise is constructed, and the measured noise parameters are dynamically corrected through Bayesian inference to generate a noise model that includes the variance of shot noise and the mean of thermal noise. Multiple sets of reference-measurement optical path pairs are generated by utilizing the beam splitting characteristics of the Y-waveguide. The cross-correlation coefficient of the interference signals of each optical path pair is calculated, and the two sets of highly correlated signal pairs with the highest correlation are selected. Based on the noise model, the wavelet threshold and FIR filter bandwidth parameters are adjusted to suppress shot noise and thermal noise in the output electrical signal of the photodetector, and the noise-reduced phase signal is output. The two sets of highly correlated signal pairs selected are weighted and averaged, with the weight being the cross-correlation coefficient of the interference signals of the corresponding optical path pairs, and the high-precision phase trajectory after eliminating fiber disturbance noise is extracted. Based on the high-precision phase trajectory and the phase-angular velocity conversion relationship based on the Sagnac effect, the rotational angular velocity value is calculated and output using the Sagnac phase shift formula. The specific steps for constructing the joint probability density function model of quantum fluctuation noise are as follows: The sources of quantum fluctuation noise in a single-source, single-axis fiber optic gyroscope were analyzed, and the independent characteristics of shot noise and thermal noise were determined. Based on quantum optics theory, shot noise is assumed to follow a Poisson distribution, and thermal noise follows a Gaussian distribution. By combining the statistical independence characteristics of shot noise and thermal noise, a joint probability density function model is established; Verify the goodness of fit between the joint probability density function model and the actual noise measurement data. The process of dynamically correcting measured noise parameters through Bayesian inference to generate a noise model that includes the variance of shot noise and the mean of thermal noise is as follows: Acquire time-series data of measured interference signals from fiber optic gyroscopes; Based on the joint probability density function model, initial estimates of shot noise variance and thermal noise mean are set, and the posterior probability distribution is updated using Bayesian inference method combined with time series data. The corrected variance of shot noise and the mean of thermal noise are extracted from the posterior probability distribution to generate a noise model. The Y-waveguide beam splitting characteristic refers to the characteristic of the Y-waveguide to split the input light into two beams of optical signals with orthogonal polarization directions based on the birefringence effect, which are then transmitted through different paths and then combined for output. The specific steps for selecting the two highly correlated signal pairs are as follows: Multiple pairs of optical signals between the reference optical path and the measurement optical path are generated by utilizing the beam splitting characteristics of the Y-waveguide, and the cross-correlation coefficient is calculated for the interference signal of each pair of optical paths. Sort the optical path pairs by cross-correlation coefficient from highest to lowest, and select the top two groups with the highest cross-correlation coefficient.

2. The electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope as described in claim 1, characterized in that: The wavelet threshold is a critical value used for selecting signal coefficients, adaptively determined based on noise intensity and signal sparsity; the FIR filter bandwidth parameter is the filter boundary frequency set according to the frequency range of the noise to be suppressed.

3. The electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope as described in claim 2, characterized in that: The specific steps for producing the noise-reduced phase signal are as follows. The electrical signal output by the photodetector is processed using wavelet thresholding for soft thresholding. The electrical signal after soft thresholding is input into an FIR filter, low-pass filtered according to the set bandwidth parameters, and continuous phase change information is extracted by Hilbert transform, which is output as the phase signal after denoising by wavelet thresholding and FIR filter.

4. The electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope as described in claim 3, characterized in that: The specific steps for extracting the high-precision phase trajectory after eliminating fiber optic disturbance noise are as follows. For the two selected pairs of highly correlated signals, the phase signals after joint noise reduction are extracted respectively, and the two phase signals are weighted and averaged using the cross-correlation coefficient of the two pairs of signals as the weight. The phase signal after weighted averaging is filtered by moving average to output a high-precision phase trajectory after eliminating fiber optic disturbance noise.

5. The electrical signal demodulation method for a single-source, single-axis fiber optic gyroscope as described in claim 4, characterized in that: The specific steps for calculating and outputting the rotational angular velocity using the Sagnac phase shift formula are as follows. The phase change per unit time is extracted from the high-precision phase trajectory, and the rotational angular velocity value is calculated based on the phase-angular velocity conversion relationship of the Sagnac effect. Temperature drift compensation and device nonlinearity calibration are performed on the rotational angular velocity value to correct system errors and output the calibrated rotational angular velocity value.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the electrical signal demodulation method for a single-source single-axis fiber optic gyroscope as described in any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the electrical signal demodulation method for the single-source single-axis fiber optic gyroscope according to any one of claims 1 to 5.

Images