A wide-area time transfer method based on adaptive random noise constraint

Abstract

The application discloses a wide-area time transfer method based on adaptive random noise constraint, and the method comprises the following steps: obtaining historical observation data of a receiver, and calculating a receiver clock error sequence based on a precise point positioning model; performing variance analysis on the receiver clock error sequence, identifying a dominant noise type under different smoothing times, and determining a time interval for segmented modeling; selecting a corresponding variance estimator or constructing a mixed variance model according to the dominant noise type, and calculating an adaptive random noise constraint value; introducing the adaptive random noise constraint value into a process noise matrix of the precise point positioning model, and performing dynamic constraint estimation on the receiver clock error to realize time transfer. The method of the embodiment of the application can overcome systematic bias introduced by a traditional fixed constraint strategy, adapt to differences in physical characteristics of different atomic clocks, and significantly improve the receiver clock error estimation accuracy and the frequency stability of wide-area time transfer.

Description

Technical Field

[0001] This invention relates to the field of satellite navigation precision positioning and high-precision timing technology, and in particular to a wide-area time transfer method based on adaptive random noise constraints. Background Technology

[0002] Precise Point Positioning (PPP) technology of Global Navigation Satellite System (GNSS) has become the core means of international time transfer. Real-time time transfer technology has been applied to national economic construction, social life, and defense science and technology. With the rapid development of science and technology, people have an increasing demand for highly reliable and accurate time, and meeting the needs of real-time users is the primary issue in achieving high-precision time transfer. Precise point positioning has certain limitations in achieving high-precision wide-area time transfer. This is because when directly using the traditional white noise model to estimate the receiver clock bias, it is affected by the instability of orbit and clock bias products, unmodeled errors, and multipath effects, making it difficult to fully reflect the actual variation characteristics of the receiver clock bias. It may absorb some unknown noise, thereby reducing frequency stability and seriously interfering with the estimation of receiver clock bias parameters. To mitigate the impact of the above errors on clock bias estimation, the inter-epoch constraint method is applied to the estimation of receiver clock bias. This method usually uses the prior noise of the receiver clock as constraint information, and constrains the clock bias parameters by introducing constraint variance between adjacent epochs. The advantage of the inter-epoch constraint method is that it can combine the physical characteristics of atomic clocks to a certain extent, making the clock bias estimation results smoother and more stable. However, traditional inter-epoch constraint methods often use a single prior noise, and the physical characteristics and random uncertainties of atomic clocks from different brands can easily introduce systematic biases into fixed constraint strategies in practical applications, thus affecting the effectiveness of random noise constraint estimation. Currently, the only improved method is to constrain clock errors at different stations based on a single Allan variance. However, this method theoretically suffers from limitations: the calculation of constraint values ​​often utilizes variance data and does not consider the differences in the impact of dominant noise types on clock error estimation under different smoothing periods; it may also fail to converge when dealing with specific noise levels, resulting in highly misleading evaluations in complex noise environments.

[0003] In summary, this invention proposes a receiver clock bias estimation method with adaptive random noise constraints. This method is theoretically more complete and can estimate the receiver clock bias more reasonably and accurately, thereby improving the stability of wide-area high-precision time transfer. Summary of the Invention

[0004] The main objective of this invention is to provide a wide-area time transfer method based on adaptive random noise constraints.

[0005] Another objective of this invention is to propose a wide-area time transfer device based on adaptive random noise constraints.

[0006] The third objective of this invention is to provide a computer device.

[0007] A fourth objective of this invention is to provide a non-transitory computer-readable storage medium.

[0008] To achieve the above objectives, a first aspect of the present invention proposes a wide-area time transfer method based on adaptive random noise constraints, comprising:

[0009] Historical observation data from the receiver is acquired, and the receiver clock error sequence is calculated based on a precise single-point positioning model. A variance analysis was performed on the receiver clock error sequence to identify the dominant noise type under different smoothing times and to determine the time interval for segmented modeling. Select the corresponding variance estimator or construct a mixed variance model based on the dominant noise type, and calculate the adaptive random noise constraint value for each time interval. The adaptive random noise constraint value is introduced into the process noise matrix of the precise single-point positioning model to dynamically constrain and estimate the receiver clock error in order to achieve time transfer.

[0010] In one embodiment of the present invention, the step of acquiring historical observation data of the receiver and calculating the receiver clock bias sequence based on a precise single-point positioning model includes: Simultaneously, pseudorange and carrier phase observations are used to perform ionospherically free combined precise single-point positioning to calculate receiver clock error parameters and construct observation equations; The historical observation data is processed based on the observed value equation to obtain the receiver clock error sequence.

[0011] In one embodiment of the present invention, the step of performing variance analysis on the receiver clock bias sequence to identify the dominant noise type under different smoothing times and determine the time interval for segmented modeling includes: Calculate the Allan and Hadamard variances of time series data using historical clock difference data; Based on the Allan variance and the Hadamard variance analysis, the dominant noise of clock error is identified in different smoothing time periods. The dominant noise type is identified in different smoothing time periods according to the characteristic slope corresponding to each type of dominant noise. The time interval for segmented modeling is determined by combining the characteristic slope results.

[0012] In one embodiment of the present invention, the step of selecting the corresponding variance estimator or constructing a mixed variance model according to the dominant noise type and calculating the adaptive random noise constraint value corresponding to each time interval includes: Based on phase-modulated white noise, frequency-modulated white noise, frequency-modulated random walk noise and frequency-modulated random run noise, the corresponding noise figures can be obtained by using least squares fitting of the Hadamard correspondence with noise and the Allan correspondence with noise, and the random walk noise can be determined. Calculate the atomic clock noise variance calculated by Hadamard variance and Allan variance; Select the corresponding variance estimator or construct a mixed variance model based on the dominant noise type, and calculate the adaptive random noise constraint value for each time interval.

[0013] In one embodiment of the present invention, the step of selecting the corresponding variance estimator or constructing a mixed variance model according to the dominant noise type and calculating the adaptive random noise constraint value corresponding to each time interval includes: When analyzing the dominant noise type, if the identification results show that the dominant noise is phase-modulated white noise or frequency-modulated white noise, the stability is calculated using Allan variance. When the dominant noise is frequency-modulated random running noise, the Hadamard variance is selected. When the dominant noise is frequency-modulated random walk noise, the sensitivity difference between Allan variance and Hadamard variance to frequency-modulated random walk noise is identified through theoretical derivation. Based on this sensitivity ratio, a hybrid constraint model is constructed, and hybrid variance calculation is introduced. A fusion algorithm with inversely proportional weighting based on sensitivity is established. After normalization, the optimal weighting coefficient is obtained. The optimal weighting coefficient is used to calculate the hybrid variance to adaptively determine the random noise constraint value.

[0014] In one embodiment of the present invention, the step of introducing the adaptive random noise constraint value into the process noise matrix of the precise single-point positioning model to dynamically constrain the receiver clock bias for time transfer includes: The calculated process noise matrix is ​​updated by adding random noise information and incorporating the adaptive random noise constraint value. By combining the constraint variance obtained from the solution with the clock error estimate of the previous epoch, a one-step prediction is made to obtain the predicted value and variance of the clock error in the current epoch. Based on the updated process noise matrix and the predicted value and variance of the clock bias, the receiver clock bias is dynamically constrained and estimated to achieve time transfer.

[0015] To achieve the above objectives, a second aspect of the present invention provides a wide-area time transfer device based on adaptive random noise constraints, comprising: The data processing module is used to acquire historical observation data from the receiver and calculate the receiver clock error sequence based on the precise single-point positioning model. The noise characteristic analysis module is used to perform variance analysis on the receiver clock error sequence, identify the dominant noise type under different smoothing times, and determine the time interval for segmented modeling. The adaptive constraint construction module is used to select the corresponding variance estimator or construct a mixed variance model according to the dominant noise type, and calculate the adaptive random noise constraint value corresponding to each time interval. The dynamic estimation and transfer module is used to introduce the adaptive random noise constraint value into the process noise matrix of the precise single-point positioning model, and to perform dynamic constraint estimation of the receiver clock error to achieve time transfer.

[0016] To achieve the above objectives, a third aspect of this application provides a computer device comprising a processor and a memory; wherein the processor runs a program corresponding to the executable program code stored in the memory, for implementing a wide-area time transfer method based on adaptive random noise constraints as described in the first aspect embodiment.

[0017] To achieve the above objectives, a fourth aspect of this application provides a non-transitory computer-readable storage medium storing a computer program that, when executed by a processor, implements a wide-area time transfer method based on adaptive random noise constraints as described in the first aspect embodiment.

[0018] The embodiments of the present invention have the following beneficial effects: This invention leverages the complementary strengths of Allan and Hadamard variances to form a dual variance fusion strategy, achieving precise noise suppression, effective suppression of frequency drift and unmodeled errors, and adaptation analysis for the physical characteristics of atomic clocks from different brands. Simultaneously, it combines an adaptive segmented constraint algorithm to dynamically optimize the noise constraint strategy and achieve dynamic and accurate estimation of receiver clock bias. This not only improves the frequency stability of receiver clock bias by an order of magnitude, solving the systematic bias problem of traditional methods, but also significantly enhances the anti-interference capability of the time transfer link and reduces link fluctuations, ultimately significantly improving the overall accuracy, stability, and algorithm adaptability of BeiDou / GNSS wide-area time transfer. Attached Figure Description

[0019] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart illustrating a wide-area time transfer method based on adaptive random noise constraints, provided for an embodiment of the present invention; Figure 2 A comparison chart of clock bias stability of a single-station receiver provided in an embodiment of the present invention; Figure 3A comparison chart of time-transmission link stability provided for embodiments of the present invention; Figure 4 This is a structural diagram of a wide-area time transfer device based on adaptive random noise constraints, provided in an embodiment of the present invention. Detailed Implementation

[0020] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] The following description, with reference to the accompanying drawings, describes a wide-area time transfer method based on adaptive random noise constraints proposed according to an embodiment of the present invention.

[0023] Example 1 This embodiment provides a wide-area time transfer method based on adaptive random noise constraints. For example... Figure 1 As shown, the method includes the following steps: S1: Obtain historical observation data from the receiver and calculate the receiver clock error sequence based on the precise single-point positioning model.

[0024] Specifically, this step uses both pseudorange and carrier phase observations to perform ionospherically-free combined PPP calculation of receiver clock bias parameters, and the constructed observation equation is as follows:

[0025] in, , Each is a receiver, measured in units of distance. To satellite The pseudorange and phase observations; For the corresponding geometric distance, For satellite clock bias, For receiver clock bias, For the corresponding zenith tropospheric delay, For the corresponding projection function, For floating-point ambiguity parameters, For the corresponding wavelength, and These are pseudorange and phase measurement noise, respectively, which include multipath errors.

[0026] The historical observation data is processed based on the observed value equation to obtain the receiver clock error sequence.

[0027] S2, perform variance analysis on the receiver clock error sequence to identify the dominant noise type under different smoothing times and determine the time interval for segmented modeling.

[0028] Specifically, this step uses historical clock difference data to calculate the Allan variance and Hadamard variance of the time series. First, the Allan variance of the time series is calculated using the following formula:

[0029] in, To smooth out time, The number of samples, As a smoothing factor, it is generally taken as... .

[0030] Next, the Hadamard variance is calculated using historical clock error data. The formula is as follows:

[0031] in, To smooth out time, The number of samples, As a smoothing factor, it is generally taken as... .

[0032] Furthermore, based on the correspondence between the characteristic slopes of the Log-Log plots of Allan and Hadamard variances and the dominant noise type, the dominant noise type of clock error in different smoothing time periods was analyzed: when the dominant noise is phase-modulated white noise (QN), the characteristic slope is -1; when the dominant noise is frequency-modulated white noise (RWN), the characteristic slope is -0.5; when the dominant noise is frequency-modulated random walk noise (FRW), the characteristic slope is 0; and when the dominant noise is frequency-modulated random run noise (RRW), the characteristic slope is 0.5. Combining the above characteristic slope identification results, the time interval for segmented clock error modeling was further determined. The statistical characteristics of random noise were modeled using a power-law spectrum, the expression of which is:

[0033] in, The spectral density represents the relative frequency deviation, and f represents the sideband frequency. The noise figure is The energy spectrum noise intensity coefficient, This indicates the upper limit of the crystal oscillator's output frequency.

[0034] Next, based on the interconversion relationship between phase data and frequency data, the time-domain power-law spectral model can be obtained, and its expression is:

[0035] in, Spectral density representing time deviation.

[0036] Frequency stability is mainly measured using time-domain variance. Spectral density of relative frequency deviation in the frequency domain Analysis reveals the following relationship between the two:

[0037] in, The variance transfer function determines the sensitivity to variance frequencies. Let Allan be the variance transfer function; Let be the Hadamard variance transfer function. From this, we can obtain the relationship between the slope of the time-domain variance Log-Log plot and various types of noise.

[0038] S3. Select the corresponding variance estimator or construct a mixed variance model according to the dominant noise type, and calculate the adaptive random noise constraint value corresponding to each time interval.

[0039] Specifically, considering the non-convergence of traditional Allan variance under FM random run noise and frequency drift conditions, and the fact that frequency drift is easily confused with random walk noise, Allan variance exhibits high sensitivity and stability in handling short-term clock error estimation, while Hadamard variance is superior due to its excellent ability to suppress frequency drift. Therefore, this method selects an appropriate variance estimator for different dominant noise types. The specific process is as follows: First, the strong correlation and stationarity between epochs imply that prior constraints can be introduced into the receiver clock bias to enhance the constraint strength of the stochastic model. Therefore, before implementing adaptive random noise constraints, the feasibility of this method must be verified by performing autocorrelation analysis on the receiver clock bias, as shown in the formula:

[0040] in, It is the autocorrelation function. As the expected value, As a stationary random process, most signals and noise encountered in communication systems are considered to be stationary random processes; and These are the mean and variance, respectively, and t is time.

[0041] Secondly, according to phase-modulated white noise Frequency-modulated white noise FM random walk noise and frequency-modulated random running noise By using least squares fitting of the correspondence between Hadamard noise and noise and Allan noise, the corresponding noise figures can be obtained, thus determining the characteristics of the frequency-modulated random walk noise. The fitting correlation formula is as follows:

[0042]

[0043]

[0044]

[0045] in, For phase-modulated white noise, For frequency-modulated white noise, For frequency-modulated random walk noise, For random running noise, The atomic clock noise variance calculated for the Hadamard variance. The variance of atomic clock noise calculated for Allan variance. At the speed of light, The sampling interval is denoted as .

[0046] Next, based on the identified dominant noise type, the corresponding variance estimator is selected or a mixed variance model is constructed. When frequency-modulated random walk noise is the dominant noise, the Allan variance and Hadamard variance satisfy the following formula:

[0047]

[0048] As the formula shows, the Allan variance is twice as sensitive to frequency-modulated random walk noise as the Hadamard variance. Based on this physical fact, according to the sensitivity ratio... A hybrid constraint model is constructed, and hybrid variance calculation is introduced to establish a fusion algorithm with inversely proportional sensitivity weighting. The weighting coefficients are then normalized to obtain the optimal weighting coefficients:

[0049] in, For Hadamard variance weights, Using Allan variance weights, this hybrid constraint strategy effectively combines the short-term sensitivity of Allan variance with the long-term stability of Hadamard variance, forming a variance estimation system with complementary advantages.

[0050] Finally, through in-depth analysis of the atomic clock performance, the dominant noise types within different smoothing time windows were identified, and the noise model was piecewise estimated and dynamically adjusted accordingly. The selection rules for the variance estimator were as follows: when the dominant noise was phase-modulated white noise or frequency-modulated white noise, Allan variance was used for stability calculation; when the dominant noise was frequency-modulated random run noise, Hadamard variance was used; and when the dominant noise was frequency-modulated random walk noise, the above-mentioned mixed variance model was used. Finally, combining the results of the corresponding variance estimator or the mixed variance model, the adaptive random noise constraint value was calculated, and its formula is:

[0051] S4, the adaptive random noise constraint value is introduced into the process noise matrix of the precise single-point positioning model to dynamically constrain the receiver clock error to achieve time transfer.

[0052] Specifically, the most commonly used PPP model in the time-frequency domain is the ionosphere-free combined model. This model uses both pseudorange and carrier phase observations to estimate unknown parameters. The ionosphere-free combined model can be expressed as:

[0053] in, , Each is a receiver, measured in units of distance. To satellite The pseudorange and phase observations; For the corresponding geometric distance, For satellite clock bias, For receiver clock bias, For the corresponding zenith tropospheric delay, For the corresponding projection function, For floating-point ambiguity parameters, For the corresponding wavelength, and These are pseudorange and phase measurement noise, respectively, which include multipath error. The receiver hardware delay is absorbed by the receiver clock bias during receiver clock bias calculation; the actual calculated receiver clock bias is the sum of the receiver hardware delay value and the actual receiver clock bias value.

[0054] Next, based on this model, random noise information is added to the process noise matrix, and the calculated adaptive random noise constraint values ​​are incorporated to complete the update of the process noise matrix. The formula is as follows:

[0055] in, The process noise matrix for adding random noise information.

[0056] Furthermore, by combining the calculated constraint variance with the clock error estimate from the previous epoch, the predicted clock error and variance for the current epoch are obtained through a one-step prediction, as shown in the formula:

[0057]

[0058] in, and They are respectively and The predicted state value at time; for arrive The state transition matrix at time t is an identity matrix. and They are respectively and The variance-covariance matrix at time points; for The process noise matrix at time step 1.

[0059] Finally, based on the updated process noise matrix and the predicted values ​​and variances of the clock bias, dynamic constraint estimation is performed on the receiver clock bias. After hardware delay correction is applied to the calculated receiver clock bias, the time difference between station A and station B is calculated using the following formula to achieve wide-area time transfer:

[0060] in, and These are the receiver clock differences for site A and site B, respectively. and These are the local times of site A and site B, respectively. This is a reference time.

[0061] In summary, the technical advantages of this invention are also reflected in its ability to accurately obtain various noise coefficients through least-squares fitting, and to systematically suppress the influence of various types of noise by combining mixed variance calculation and optimal weighting strategies. Compared with traditional white noise models, this invention achieves a significant breakthrough in ensuring the long-term stability and reliability of time transfer, providing a more complete technical solution for the field of high-precision time transfer.

[0062] To fully verify the effectiveness of the present invention, 12 different regional stations equipped with high-precision hydrogen atomic clocks were selected, with DRAO, WTZR, and BOR1 as the central nodes respectively. Nine time links were designed according to the different types and brands of atomic clocks connected to the receivers of the stations. The station information is shown in Table 1.

[0063] Table 1. Information on the measuring station and external atomic clock

[0064] It should be noted that this experiment selected October 23 to November 19, 2022 as the observation period. First, the clock bias was calculated using the traditional PPP method, and the clock bias parameters were treated as white noise. Then, autocorrelation analysis, dominant noise type determination, and process noise identification were performed on the clock bias sequence. Based on the ionospheric PPP function and the method proposed in this invention for estimating receiver clock bias using adaptive random noise constraints, as shown in Table 2, two PPP processing schemes were designed and experimentally verified using observation data from 7 days (324 to 330 days) in 2022.

[0065] Table 2 Data Processing Strategies

[0066] Specifically, the adaptive random noise model scheme significantly improves receiver clock bias estimation results in both short-term and long-term stability compared to the traditional white noise model scheme, achieving a stability of 100s compared to the traditional white noise model scheme. -13 The magnitude dropped to 10 -14 The adaptive random noise model scheme improves the stability of the model by 30% over 100 seconds compared to the traditional white noise model scheme; and improves the stability of the model by 10% over 10,000 seconds. -14 The magnitude dropped to 10 -15 The adaptive random noise model scheme achieves approximately 20% improvement in stability over traditional white noise models in kilosecond and 10,000-second intervals; for different brands, the initial ADEV of some devices is concentrated around 10. -13 Nearby, some devices had initial values ​​close to 10. -12 ~10 -13 It can be seen that the short-term noise of some hydrogen atomic clock devices is slightly high. Compared with the traditional model, the PPP model with adaptive random noise constraints significantly improves the frequency stability of the receiver clock bias at stations with external hydrogen atomic clocks, with the best improvement reaching about one order of magnitude. In particular, the frequency stability improvement is most significant for hydrogen atomic clock devices with slightly high short-term noise. Figure 2 The figure shown is a comparison of the clock bias frequency stability of a single-station receiver in this embodiment. The left side shows the results of the traditional white noise model scheme, and the right side shows the results of the adaptive random noise model scheme.

[0067] Furthermore, time-transfer link analysis results show that the adaptive random noise model scheme significantly outperforms the traditional white noise model scheme in terms of clock error stability during time transfer. All three types of hydrogen atomic clocks exhibit significant improvements across various sampling intervals, with the type showing the best performance. Under the adaptive random noise model scheme, the time transfer stability improvement percentage can reach up to approximately 50% within sampling intervals from 480s to 1920s. Figure 3 The figure shown is a comparison of the stability of the time-transfer link in this embodiment. The left side shows the results of the traditional white noise model scheme, and the right side shows the results of the adaptive random noise model scheme.

[0068] In summary, the experimental data fully validates the superior performance of this invention. The frequency stability of hydrogen atomic clock stations with slightly higher short-term noise is improved by an order of magnitude. Other types of hydrogen atomic clock stations also show significant improvements, with a marked reduction in clock difference sequence fluctuations and a qualitative improvement in stability. Addressing the limitations of existing real-time PPP time transfer methods, such as the instability of track and clock difference data transmission, susceptibility to re-convergence issues, and insufficient compatibility with different brands of atomic clocks, this invention determines the piecewise function interval through in-depth analysis of historical clock difference data and adaptively calculates the variance constraint value, demonstrating excellent compatibility. It is not only compatible with mainstream hydrogen clock brands but also achieves effective performance improvements for different types of atomic clocks, such as rubidium and cesium clocks.

[0069] Example 2 This invention also provides a wide-area time transfer device based on adaptive random noise constraints, such as... Figure 4 As shown, the device 10 includes: The data processing module 100 is used to acquire historical observation data from the receiver and calculate the receiver clock error sequence based on the precise single-point positioning model. The noise characteristic analysis module 200 is used to perform variance analysis on the receiver clock error sequence, identify the dominant noise type under different smoothing times, and determine the time interval for segmented modeling. The adaptive constraint construction module 300 is used to select the corresponding variance estimator or construct a mixed variance model according to the dominant noise type, and calculate the adaptive random noise constraint value corresponding to each time interval. The dynamic estimation and transmission module 400 is used to introduce the adaptive random noise constraint value into the process noise matrix of the precise single-point positioning model, and to perform dynamic constraint estimation of the receiver clock error to achieve time transmission.

[0070] Example 3 To implement the methods of the above embodiments, the present invention also provides a computer device, which includes a memory and a processor; wherein the processor runs a program corresponding to the executable program code by reading executable program code stored in the memory, so as to implement the various steps of the methods described above.

[0071] Example 4 To implement the above embodiments, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in the foregoing embodiments.

[0072] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0073] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0074] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A wide-area time transfer method based on adaptive random noise constraints, characterized in that, Includes the following steps: Historical observation data from the receiver is acquired, and the receiver clock error sequence is calculated based on a precise single-point positioning model. A variance analysis was performed on the receiver clock error sequence to identify the dominant noise type under different smoothing times and to determine the time interval for segmented modeling. Based on the dominant noise type, select the corresponding variance estimator or construct a mixed variance model, and calculate the adaptive random noise constraint values ​​for each time interval, including: Based on phase-modulated white noise, frequency-modulated white noise, frequency-modulated random walk noise and frequency-modulated random run noise, the corresponding noise figures are obtained by least-squares fitting of the Hadamard-noise correspondence and the Allan-noise correspondence, and the random walk noise is determined. Calculate the atomic clock noise variance based on Hadamard variance and the atomic clock noise variance based on Allan variance, respectively; When analyzing the dominant noise type, if the identification results show that the dominant noise is phase-modulated white noise or frequency-modulated white noise, the stability is calculated using Allan variance. When the dominant noise is frequency-modulated random running noise, the Hadamard variance is used for stability calculation. When the dominant noise is frequency-modulated random walk noise, determine the sensitivity of the Allan variance and Hadamard variance to frequency-modulated random walk noise, and determine the weighting relationship based on the relative magnitude of their sensitivity, so that the weights are inversely proportional to the sensitivity. The weights are normalized to obtain the optimal weighting coefficients for the Allan variance and the Hadamard variance. The Allan variance and the Hadamard variance are weighted and fused using the optimal weighting coefficients to obtain the mixed variance. The random noise constraint value is then adaptively determined based on the mixed variance. Calculate the adaptive random noise constraint value for each time interval; The adaptive random noise constraint value is introduced into the process noise matrix of the precise single-point positioning model to dynamically constrain and estimate the receiver clock error in order to achieve time transfer.

2. The method according to claim 1, characterized in that, The process of acquiring historical observation data from the receiver and calculating the receiver clock bias sequence based on a precise single-point positioning model includes: Simultaneously, pseudorange and carrier phase observations are used to perform ionospherically free combined precise single-point positioning to calculate receiver clock error parameters and construct observation equations; The historical observation data is processed based on the observed value equation to obtain the receiver clock error sequence.

3. The method according to claim 1, characterized in that, The step of performing variance analysis on the receiver clock bias sequence to identify the dominant noise type under different smoothing times and determine the time interval for segmented modeling includes: Calculate the Allan and Hadamard variances of time series data using historical clock difference data; Based on the Allan variance and the Hadamard variance analysis, the dominant noise of clock error is identified in different smoothing time periods. The dominant noise type is identified in different smoothing time periods according to the characteristic slope corresponding to each type of dominant noise. The time interval for segmented modeling is determined by combining the characteristic slope results.

4. The method according to claim 1, characterized in that, The process noise matrix of the precise single-point positioning model, which incorporates the adaptive random noise constraint value into the model, and dynamically estimates the receiver clock bias to achieve time transfer, includes: The calculated process noise matrix is ​​updated by adding random noise information and incorporating the adaptive random noise constraint value. By combining the constraint variance obtained from the solution with the clock error estimate of the previous epoch, a one-step prediction is made to obtain the predicted value and variance of the clock error in the current epoch. Based on the updated process noise matrix and the predicted value and variance of the clock bias, the receiver clock bias is dynamically constrained and estimated to achieve time transfer.

5. A wide-area time transfer device based on adaptive random noise constraints, characterized in that, The device includes: The data processing module is used to acquire historical observation data from the receiver and calculate the receiver clock error sequence based on the precise single-point positioning model. The noise characteristic analysis module is used to perform variance analysis on the receiver clock error sequence, identify the dominant noise type under different smoothing times, and determine the time interval for segmented modeling. The adaptive constraint construction module is used to select the corresponding variance estimator or construct a mixed variance model according to the dominant noise type, and calculate the adaptive random noise constraint values ​​corresponding to each time interval, including: Based on phase-modulated white noise, frequency-modulated white noise, frequency-modulated random walk noise and frequency-modulated random run noise, the corresponding noise figures are obtained by least-squares fitting of the Hadamard-noise correspondence and the Allan-noise correspondence, and the random walk noise is determined. Calculate the atomic clock noise variance based on Hadamard variance and the atomic clock noise variance based on Allan variance, respectively; When analyzing the dominant noise type, if the identification results show that the dominant noise is phase-modulated white noise or frequency-modulated white noise, the stability is calculated using Allan variance. When the dominant noise is frequency-modulated random running noise, the Hadamard variance is used for stability calculation. When the dominant noise is frequency-modulated random walk noise, determine the sensitivity of the Allan variance and Hadamard variance to frequency-modulated random walk noise, and determine the weighting relationship based on the relative magnitude of their sensitivity, so that the weights are inversely proportional to the sensitivity. The weights are normalized to obtain the optimal weighting coefficients for the Allan variance and the Hadamard variance. The Allan variance and the Hadamard variance are weighted and fused using the optimal weighting coefficients to obtain the mixed variance. The random noise constraint value is then adaptively determined based on the mixed variance. Calculate the adaptive random noise constraint value for each time interval; The dynamic estimation and transfer module is used to introduce the adaptive random noise constraint value into the process noise matrix of the precise single-point positioning model, and to perform dynamic constraint estimation of the receiver clock error to achieve time transfer.

6. A computer device, characterized in that, Including processor and memory; The processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement a wide-area time transfer method based on adaptive random noise constraints as described in any one of claims 1-4.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements a wide-area time transfer method based on adaptive random noise constraints as described in any one of claims 1-4.

Images