Real-time atomic time calculation method

Abstract

The invention belongs to the technical field of time frequency. The invention provides a real-time atomic time calculation method. According to the embodiment of the invention, modeling is carried out on the basis of the weighted average thought assisted by the exponential filter coefficient, the exponential filter coefficient suppresses high-frequency noise, estimation fluctuation caused by short-term fluctuation is avoided, meanwhile, the smooth effect of the whole iterative calculation is enhanced, and errors caused by estimation are reduced. According to the design, real-time control is carried out on an adaptive forgetting factor based on the real-time change of the atomic clock state, the forgetting rate is dynamically adjusted in real time according to a prediction error and a parameter change rate, the defect of lagging response under conventional fixed parameters is avoided, robustness can be guaranteed and quick response can be realized at the same time in a complex dynamic change scene of the system, and the system performance is improved. And the method has the advantages of noise resistance and time-varying characteristics.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the technical field of time and frequency, and in particular, to a method for calculating real-time atomic time. Background Art

[0002] Due to the adjustment of the time scale algorithm of the International Bureau of Weights and Measures (BIPM) (improving the frequency algorithm in 2011 and the weight algorithm in 2014), the proportion of the weight of hydrogen atomic clocks in the reduction of international standard time is relatively high, that is, hydrogen atomic clocks play an increasingly important role in timekeeping work. In addition, hydrogen atomic clocks have good short-term stability and are often used as the master clocks of the timekeeping systems in timekeeping laboratories to participate in the measurement and comparison of atomic clocks.

[0003] Atomic time is an important reference quantity for the stable operation of the timekeeping system and a crucial part of timekeeping work. For the operation of each atomic clock in the timekeeping system, a stable reference is required to obtain a highly reliable output of timekeeping results. If only one atomic clock is used as the reference, when this clock malfunctions or fails, the timekeeping of the entire system will experience a large jump or interruption. Therefore, it is necessary to calculate a continuous, stable, and reliable time reference quantity, that is, atomic time.

[0004] The International Bureau of Weights and Measures (BIPM) generates the international standard time UTC based on more than 400 atomic clocks of different types in more than 80 timekeeping laboratories around the world. BIPM obtains the time difference between the atomic clocks in each timekeeping laboratory around the world and the standard time UTC (PTB) of the international comparison center station through a remote time comparison link, and uses a weighted average algorithm to obtain the free atomic time EAL. Subsequently, it is calibrated by fountain clocks to obtain the international atomic time TAI, and then the international standard time UTC is obtained and released at the beginning of each month. The data released this month is the calculation result of the previous month. Therefore, UTC has a delay of nearly 40 days, and there is a need for real-time high-precision time signals locally, that is, it is necessary to establish a real-time time reference system to generate the time signals, time information, and time codes required for national time service broadcasts. Therefore, it is of great significance to independently generate and maintain the real-time local time UTC(k).

[0005] In related technologies, the real-time local time UTC(k) requires real-time atomic time as a steering reference. Currently, the real-time state changes of atomic clocks and the modeling of the iterative characteristics of real-time atomic time are not comprehensively considered, resulting in low stability and low accuracy in the calculation of real-time atomic time.

[0006] Therefore, it is necessary to improve one or more problems existing in the above-mentioned related technical solutions.

[0007] It should be noted that this section aims to provide background or context for the technical solutions of the present disclosure stated in the claims. The descriptions herein are not admitted to be prior art merely because they are included in this section. Summary of the Invention

[0008] An object of embodiments of the present disclosure is to provide a real-time atomic time calculation method, thereby at least to some extent overcoming one or more problems caused by limitations and defects of related technologies.

[0009] According to an embodiment of the present disclosure, there is provided a real-time atomic time calculation method, the method comprising: Determine the type of atomic clock, set the comparison period, and collect the phase comparison data of the atomic clock within a corresponding time period; Set the length of the phase comparison data according to the stability characteristics of the atomic clock, and determine the time interval for real-time atomic time calculation; wherein, the time interval for real-time atomic time calculation is synchronized with the comparison period; Use the quadratic least squares method to perform outlier detection and preprocessing on the phase comparison data, and calculate the real-time prediction rate and real-time prediction frequency drift of the atomic clock according to the preprocessed phase comparison data; wherein, the quadratic least squares method includes a first least squares fitting and a second least squares fitting; According to the real-time prediction rate, real-time prediction frequency drift and phase comparison data, dynamically calculate the weight coefficients of each atomic clock in combination with an adaptive forgetting factor; Integrate the weight coefficients of all atomic clocks, clock error prediction data and the phase comparison data of each atomic clock, and calculate the real-time atomic time through weighted fusion.

[0010] Furthermore, the length of the phase comparison data is one day, and the iteration interval is synchronized with the comparison period as one hour.

[0011] Furthermore, in the step of using the quadratic least squares method to perform outlier detection and preprocessing on the phase comparison data, it includes: Perform a first least squares fitting on the phase comparison data, and calculate the residuals of the first fitting data; Calculate the standard deviation of the first fitting data according to the residuals of the first fitting data; Use three times the standard deviation as the threshold for outlier discrimination to perform outlier detection and filter out the outliers; Perform a second least squares fitting on the phase comparison data after filtering out the outliers to obtain the preprocessed phase comparison data.

[0012] Furthermore, calculate the real-time prediction rate and real-time prediction frequency drift of the atomic clock according to the preprocessed phase comparison data:

[0013] Among them, is the real-time prediction rate, is the real-time prediction frequency drift, is the initial moment, is the th time period, is the th phase comparison data of the time period, and -1 represents the inverse operation of the matrix.

[0014] Furthermore, in the step of dynamically calculating the weight coefficients of each atomic clock according to the real-time prediction rate, the real-time prediction frequency drift, and the phase comparison data, combined with the adaptive forgetting factor, it includes: Calculating the predicted clock error data according to the real-time prediction rate and the real-time prediction frequency drift; Calculating the absolute phase prediction error according to the phase comparison data and the predicted clock error data; Dynamically adjusting the adaptive forgetting factor according to the absolute phase prediction error; among them, the initial value of the adaptive forgetting factor is 0.99, the step size is 0.001, and the change range of the adaptive forgetting factor is controlled between 0.90 and 0.99 by the step size; Based on the phase comparison data, using the multi-corner hat method to calculate the initial value of the product of the single-clock stability and the square of the time interval for the iterative weight coefficient component parameter; Calculating the weight coefficient component using the exponential filter coefficient and the initial value of the iteration, and normalizing the weight coefficient component to obtain the weight coefficient.

[0015] Furthermore, the expression of the predicted clock error data is:

[0016] Among them, is the initial phase, is the i th real-time prediction rate of the atomic clock, is the i th real-time prediction frequency drift of the atomic clock; The expression of the absolute phase prediction error is:

[0017] Among them, is the phase comparison data, represents the dynamic update of the absolute phase prediction error; The expression of the adaptive forgetting factor is:

[0018] The expression of the weight coefficient component is:

[0019] Among them, is the exponential filtering coefficient, and is the initial value of the weight coefficient component parameter iteration; The expression of the weight coefficient is:

[0020] Among them, N is the number of atomic clocks.

[0021] Furthermore, by integrating the weight coefficients of all atomic clocks, the clock error prediction data, and the phase comparison data of each atomic clock, the real-time atomic time is calculated through weighted fusion:

[0022] Among them, is the phase comparison result between two atomic clocks.

[0023] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects: In the embodiments of the present disclosure, through the above real-time atomic time calculation method, based on the weighted average idea and supplemented by the exponential filtering coefficient for modeling, the exponential filtering coefficient suppresses high-frequency noise, avoids estimation fluctuations caused by short-term fluctuations, and at the same time enhances the smoothing effect of the entire iterative calculation, thereby reducing the error caused by estimation. At the same time, an adaptive forgetting factor is designed for real-time control based on the real-time change of the atomic clock state, and the forgetting rate is dynamically adjusted in real time according to the prediction error and the parameter change rate, avoiding the disadvantages of lag response under conventional fixed parameters. In the scenario of complex dynamic changes in the system, it can ensure both robustness and achieve fast response, and has the advantages of anti-noise and time-variation. In addition, this application can effectively realize dynamic iterative calculation in real time and converge quickly, with good robustness and flexibility; and the calculated atomic time has the advantages of long-term and short-term stability, and can provide an effective reference for the real-time driving control of the master clock of the timekeeping system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings herein are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0025] Figure 1 Shows a step diagram of a real-time atomic time calculation method in an exemplary embodiment of the present disclosure; Figure 2 Shows a specific flowchart of the real-time atomic time calculation method in an exemplary embodiment of the present disclosure; Figure 3Shows the comparison result of the fluctuation range of the overall atomic time generated by the method of the present application and the conventional method in the exemplary embodiment of the present disclosure; Figure 4 Shows the comparison result of the ALLAN deviation between the method of the present application and the conventional method in the exemplary embodiment of the present disclosure. Detailed implementation manners

[0026] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more complete and will fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0027] In addition, the accompanying drawings are only schematic illustrations of the embodiments of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities.

[0028] A real-time atomic time calculation method is provided in the present example embodiment. Referring to Figure 1 as shown in, the real-time atomic time calculation method may include: Step S101: Determine the type of atomic clock, set the comparison period, and collect the phase comparison data of the atomic clock within the corresponding time period; Step S102: Set the length of the phase comparison data according to the stability characteristics of the atomic clock, and determine the time interval for real-time atomic time calculation; wherein, the time interval for real-time atomic time calculation is synchronized with the comparison period; Step S103: Use the quadratic least squares method to detect and preprocess outliers in the phase comparison data, and calculate the real-time prediction rate and real-time prediction frequency drift of the atomic clock according to the preprocessed phase comparison data; wherein, the quadratic least squares method includes the first least squares fitting and the second least squares fitting; Step S104: Dynamically calculate the weight coefficients of each atomic clock according to the real-time prediction rate, real-time prediction frequency drift, and phase comparison data, in combination with an adaptive forgetting factor; Step S105: Integrate the weight coefficients of all atomic clocks, clock error prediction data, and the phase comparison data of each atomic clock, and calculate the real-time atomic time through weighted fusion.

[0029] Through the above real-time atomic time calculation method, based on the weighted average idea and supplemented by an exponential filtering coefficient for modeling, the exponential filtering coefficient suppresses high-frequency noise, avoids estimation fluctuations caused by short-term fluctuations, and at the same time enhances the smoothing effect of the entire iterative calculation, thereby reducing the error caused by estimation. At the same time, an adaptive forgetting factor is designed for real-time control based on the real-time change of the atomic clock state, and the forgetting rate is dynamically adjusted in real time according to the prediction error and the parameter change rate, avoiding the disadvantages of lagging response under conventional fixed parameters. In the scenario of complex dynamic changes in the system, it can ensure both robustness and achieve fast response, with the advantages of anti-noise and time-variability. In addition, this application can effectively perform dynamic iterative calculations in real time and converge quickly, with good robustness and flexibility; and the calculated atomic time has the advantages of long-term and short-term stability, which can provide an effective reference for the real-time driving control of the master clock of the timekeeping system.

[0030] Next, with reference to Figures 1 to 4 each step of the above real-time atomic time calculation method in the exemplary embodiment will be described in more detail.

[0031] In step S101, determine the type of atomic clock, set the comparison period, and collect the phase comparison data of the atomic clock within the corresponding time period.

[0032] Specifically, the clock group system, multi-channel phase comparator, multi-channel counter, and switch of the timekeeping system cooperate with each other to obtain the atomic clock comparison data (i.e., phase comparison data). The comparison period is one hour, and the comparison data file is stored in the time-frequency database. Based on this comparison data file, the phase comparison data of the VCH-1003M hydrogen atomic clock is collected as the research basis.

[0033] In step S102, set the length of the phase comparison data according to the stability characteristics of the atomic clock, and determine the time interval for real-time atomic time calculation; wherein, the time interval for real-time atomic time calculation is synchronized with the comparison period.

[0034] Specifically, the iteration interval is determined based on the length of the original data set (i.e., phase comparison data) and the time interval for real-time atomic time calculation. Since the hydrogen atomic clock has a relative advantage in stability within one day, the length of the original data set is set to one day, and the time interval for real-time calculation is synchronized with the comparison period of the timekeeping system. From step 101, the phase comparison period is obtained as one hour, so it can be set to one hour.

[0035] In step S103, use the quadratic least squares method to detect and preprocess the outliers in the phase comparison data, and calculate the real-time prediction rate and real-time prediction frequency drift of the atomic clock based on the preprocessed phase comparison data; wherein, the quadratic least squares method includes the first least squares fitting and the second least squares fitting.

[0036] Specifically, before dynamically calculating the atoms each time, it is necessary to detect outliers in the data. Here, a quadratic least squares method is proposed for detection. The first time is to detect outliers and filter them out, and the second time is to fit the valid values within the original outlier time stamps. The reason for using the two - time least squares fitting method is that the first least squares fitting will be skewed due to outliers, and the fitting estimate will deviate or even fail. Therefore, the first least squares fitting only performs outlier detection and corresponding filtering. In this way, the second least squares fitting is implemented after the first - time data pre - processing, which is reasonable and effective.

[0037] Perform real - time prediction rate of the atomic clock based on the pre - processed data and real - time prediction frequency drift of the atomic clock The calculations are as follows: (1) Where, is the real - time prediction rate, is the real - time prediction frequency drift, is the initial time, is the th time period, is the th phase comparison data of the time period, and - 1 represents the inverse operation of the matrix.

[0038] In step S104, according to the real - time prediction rate, real - time prediction frequency drift, and phase comparison data, the weight coefficients of each atomic clock are dynamically calculated in combination with the adaptive forgetting factor.

[0039] Specifically, first calculate the absolute error of phase prediction as follows: (2) Where, represents the phase comparison data of the time - keeping system, represents the predicted clock error data, and the predicted clock error data is obtained by combining the frequency prediction data and the frequency drift data with the corresponding time stamps, expressed as: (3) Where, is the initial phase. If the data starts from 0, this item is 0.

[0040] Subsequently, calculate the adaptive forgetting factor: (4) Usually, the timekeeping system is in a stable state. A higher initial value can retain more historical information. Therefore, the initial value of the adaptive forgetting factor is set to 0.99, where 0.001 is the adjustment step size. The small step size can avoid overshoot. When the state of the atomic clock deteriorates or the noise increases, the forgetting factor (i.e., the adaptive forgetting factor) is reduced, so as to reduce the weight of the atomic clock and minimize the negative impact on the current estimate. The forgetting factor range is limited to 0.90 - 0.99 because it is necessary to balance the long-term estimation of the atomic clock and the short-term adaptive state change of the atomic clock.

[0041] Subsequently, the weight coefficient component is calculated: (5) Among them, is the exponential filtering coefficient, and its purpose is to reduce the error caused by the estimation. is the initial value of the iteration, which is obtained by multiplying the stability of a single clock and the square of the calculation interval through the multi-corner hat method using prior information. The weight can be expressed as: (6) Numerical normalization is performed here.

[0042] In step S105, by integrating the weight coefficients of all atomic clocks, the clock difference prediction data, and the phase comparison data of each atomic clock, the real-time atomic time is calculated through weighted fusion.

[0043] Specifically, the real-time atomic time is calculated using the following formula: (7) Among them, represents the number of atomic clocks, represents the phase comparison result between every two atomic clocks, is the weight obtained in step 104, is the clock difference prediction data in step 104. In a specific embodiment, the specific process of the real-time atomic time calculation method of the present application is as Figure 2 shown.

[0044] 1. Collect the phase comparison data of several VCH-1003M hydrogen atomic clocks for one month in April 2025 as the research basis. The data update period is every hour, totaling 720 hours. The time for real-time atomic time calculation is the same as the data comparison period, also in hours, and the fixed interval data is based on days.

[0045] 2. Before each atomic calculation, anomaly detection, filtering, and supplementation of the data of each atomic clock within a fixed interval are performed first. Here, according to the characteristics of the atomic clock, the least-squares fitting order for anomaly value detection is determined. After analysis, the frequency drift of the VCH-1003M type hydrogen atomic clock is generally relatively small, and the main distribution noise is white frequency noise WFM. Therefore, the least-squares fitting order is set to 1. First, the residuals between the measured data and the first-order fitting data are calculated, and then the standard deviation is calculated on this basis to measure the overall dispersion of the data. According to the characteristics of the Gaussian distribution, three times the standard deviation is used as the basis for anomaly value discrimination. After detecting the anomaly values in the atomic clock comparison data, the anomaly values are filtered out, and then least-squares fitting interpolation is carried out based on the processed data to ensure the continuity and reliability of the data.

[0046] 3. Based on the above preprocessed data, combined with formula (1), the prediction rate and prediction frequency drift of each clock at the latest moment are calculated. A quadratic polynomial needs to be constructed to meet the dual-index prediction requirements of the atomic clock. Considering the best linear unbiased estimation, the prediction estimator is determined with the minimum variance of the estimator, and the result is more reliable. In addition, compared with the machine algorithm that needs to be repeatedly optimized, considering the overall real-time performance of the algorithm, the calculation is more concise and efficient. And the absolute error of the phase prediction is also calculated in real-time and dynamically updated.

[0047] 4. The weight distribution of each hydrogen atomic clock at the latest moment is calculated using formulas (2) to (6). First, the absolute error between the predicted clock difference and the measured clock difference of each clock is calculated, where the predicted clock difference is obtained based on formula (3). Here, calculating the absolute error is convenient for making direct decisions on the results. The forgetting factor is adaptively adjusted according to the size of the absolute error. When the absolute error increases, that is, the predictability of the atomic clock decreases; when the state of the atomic clock changes, the size of the forgetting factor is adjusted, ultimately affecting the weight distribution. At the same time, an exponential filtering coefficient is used to assist in control to further reduce the error caused by the estimated value. And calculate the initial value , to calculate this value, first collect the phase comparison data of several VCH-1003M type hydrogen atomic clocks for one month in March 2025. First, perform anomaly value detection on the data, and then use the triangular hat method and the ALLAN variance method to find the stability of a single clock. Finally, multiply it by the square of the calculation interval time to obtain it.

[0048] 5. Substitute the rates, predicted frequency drifts, and weights of each atomic clock into formula (7), and the real-time atomic time can be calculated. The relevant calculation results are as Figure 3 shown, and the real-time atomic time relative to the conventional one can be obtained. It can be seen that the overall fluctuation range of the atomic time generated by this application is more convergent, and the fluctuation range is further reduced.

[0049] For the obtained real-time atomic time, calculate the ALLAN deviation, and the result is as Figure 4As shown, it can be seen that the ALLAN deviation of the real-time atomic time of this application is generally smaller than that of the conventional atomic time, that is, the long-term and short-term stability of this application is better than that of the real-time atomic time obtained by the conventional method.

[0050] Through the above real-time atomic time calculation method, based on the weighted average idea and supplemented by the exponential filtering coefficient for modeling, the exponential filtering coefficient suppresses high-frequency noise, avoids the estimation fluctuation caused by short-term fluctuations, and at the same time enhances the smoothing effect of the entire iterative calculation, thereby reducing the error caused by estimation. At the same time, an adaptive forgetting factor is designed for real-time control based on the real-time change of the atomic clock state, and the forgetting rate is dynamically adjusted in real time according to the prediction error and the parameter change rate, avoiding the drawback of the lag response under the conventional fixed parameters. In the scenario of complex dynamic changes in the system, it can ensure both robustness and achieve fast response, and has the advantages of anti-noise and time-variation. In addition, this application can effectively perform dynamic iterative calculation in real time and converge quickly, with good robustness and flexibility; and the calculated atomic time has the advantages of long-term and short-term stability, which can provide an effective reference for the real-time driving control of the master clock of the timekeeping system.

[0051] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present disclosure, the meaning of "a plurality" is two or more unless otherwise specifically defined.

[0052] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine the different embodiments or examples described in this specification.

[0053] Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include the common general knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the appended claims.

Claims

1. A real-time atomic time calculation method, characterized in that, The method includes: Determining the type of atomic clock, setting the comparison period, and collecting the phase comparison data of the atomic clock within the corresponding time period; Setting the length of the phase comparison data according to the stability characteristics of the atomic clock, and determining the time interval for calculating the real-time atomic time; wherein, the time interval for calculating the real-time atomic time is synchronized with the comparison period; Using the quadratic least squares method to detect and preprocess outliers in the phase comparison data, and calculating the real-time prediction rate and real-time prediction frequency drift of the atomic clock based on the preprocessed phase comparison data; wherein, the quadratic least squares method includes the first least squares fitting and the second least squares fitting; Dynamically calculating the weights of each atomic clock based on the real-time prediction rate, real-time prediction frequency drift, and phase comparison data, in combination with an adaptive forgetting factor; Calculating the real-time atomic time through weighted fusion by integrating the weights of all atomic clocks, clock error prediction data, and the phase comparison data of each atomic clock.

2. The real-time atomic time calculation method according to claim 1, wherein The length of the phase comparison data is one day, and the iteration interval is synchronized with the comparison period as one hour.

3. The real-time atomic time calculation method according to claim 1, characterized in that, In the step of using the quadratic least squares method to detect and preprocess outliers in the phase comparison data, it includes: Performing the first least squares fitting on the phase comparison data and calculating the residuals of the first fitting data; Calculating the standard deviation of the first fitting data based on the residuals of the first fitting data; Using three times the standard deviation as the threshold for outlier discrimination to detect outliers and filter out the outliers; Performing the second least squares fitting on the phase comparison data after filtering out the outliers to obtain the preprocessed phase comparison data.

4. The real-time atomic time calculation method according to claim 3, wherein Calculating the real-time prediction rate and real-time prediction frequency drift of the atomic clock based on the preprocessed phase comparison data: Among them, is the real-time prediction rate, is the real-time prediction frequency drift, is the initial moment, is the th time period, is the phase comparison data of the th time period, and -1 represents the inverse operation of the matrix.

5. The real-time atomic time calculation method according to claim 4, wherein In the step of dynamically calculating the weight coefficients of each atomic clock based on the real-time prediction rate, real-time prediction frequency drift, and phase comparison data, in combination with an adaptive forgetting factor, it includes: Calculating the predicted clock error data based on the real-time prediction rate and real-time prediction frequency drift; Calculating the absolute phase prediction error based on the phase comparison data and the predicted clock error data; Dynamically adjusting the adaptive forgetting factor based on the absolute phase prediction error; wherein, the initial value of the adaptive forgetting factor is 0.99, the step size is 0.001, and the change range of the adaptive forgetting factor is controlled between 0.90 and 0.99 through the step size; Based on the phase comparison data, using the multi-corner hat method to calculate the product of the stability of a single clock and the square of the time interval to obtain the initial value of the weight coefficient component iteration; Calculating the weight coefficient component using the exponential filter coefficient and the initial value, and normalizing the weight coefficient component to obtain the weight.

6. The real-time atomic time calculation method according to claim 5, wherein, The expression of the predicted clock error data is: Among them, is the initial phase, is the real-time prediction rate of the i th atomic clock, is the real-time prediction frequency drift of the i th atomic clock; The expression of the absolute phase prediction error is: Among them, is the phase comparison data, indicating the absolute error of the dynamically updated phase prediction; The expression of the adaptive forgetting factor is: The expression of the weight coefficient component is: Among them, is the exponential filtering coefficient, is the initial value of the weight coefficient component parameter iteration; The expression of the weight is: Wherein, N is the number of atomic clocks.

7. The real-time atomic time calculation method according to claim 6, characterized in that Calculating the real-time atomic time through weighted fusion by integrating the weight coefficients of all atomic clocks, clock error prediction data, and the phase comparison data of each atomic clock: Among them, is the phase comparison result of every two atomic clocks.

Images