Real-time sliding ultra-short-term forecasting model algorithm based on frequency data and phase data

Abstract

The present application relates to the field of satellite navigation systems. Disclosed is a real-time sliding ultra-short-term forecasting model algorithm based on frequency data and phase data, the algorithm comprising the following steps: S1, converting clock error phase data into frequency data; S2, using a frequency abnormality detection function to process the frequency data, so as to eliminate an abnormal value, wherein the frequency abnormality detection function determines, on the basis of the standard deviation and threshold of the frequency data, whether the frequency data is an abnormal point; and S3, performing real-time sliding clock error forecasting, and using fitting phase data and a threshold range to eliminate the abnormal value and update a forecasting epoch. In the present invention, the elimination of an abnormal value from clock data and the calculation of a clock error prediction value are linked, so that the detection and elimination of the abnormal value from the clock data are rolled out in real time; and the calculation of a real-time clock error prediction value is performed, so that a clock frequency deviation is corrected. The present invention can improve the accuracy, real-time performance and data stability of forecasting, and is flexible and applicable to various types of satellite orbits.

Description

A real-time sliding ultra-short-term forecasting model algorithm based on frequency and phase data Technical Field

[0001] The present invention relates to the technical field of satellite navigation systems, in particular to a real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data. Background Art

[0002] Satellite navigation systems play a vital role in modern society, widely used in aviation, maritime navigation, and vehicle navigation. They provide precise positioning and navigation services for these applications, enabling us to accurately find our destinations in unfamiliar environments. Satellite clock error, a crucial parameter in satellite navigation systems, directly impacts positioning and navigation accuracy. Therefore, accurately predicting changes in satellite clock error is crucial to ensuring the high accuracy and reliability of navigation systems.

[0003] Traditional satellite clock error prediction methods primarily rely on statistical analysis or mathematical modeling of historical data. These methods, through analysis and modeling of historical data, can provide a certain degree of forecast accuracy. However, these methods have several limitations. First, they typically require long periods of data accumulation and offline processing, failing to meet real-time requirements. Real-time performance is crucial in applications requiring instant navigation and positioning. Second, due to environmental and system changes, the statistical characteristics of historical data may change, resulting in a decrease in the accuracy of forecast results. This uncertainty can pose potential risks and safety hazards to navigation systems.

[0004] Furthermore, traditional forecasting methods are relatively weak in handling abnormal data. In real-world applications, satellite clock error data may be affected by various noise and interference factors, including measurement errors, signal attenuation, and atmospheric disturbances. These abnormal values ​​can adversely affect forecast results and reduce their reliability.

[0005] Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention provides a real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data, aiming to improve the accuracy, real-time performance, data stability and flexibility of satellite clock error prediction. Based on the real-time received satellite clock error phase data, it is converted into frequency data, and the frequency anomaly detection function is used to eliminate outliers, and real-time sliding clock error prediction is performed.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data, comprising the following steps:

[0008] Convert clock difference phase data into frequency data;

[0009] Use the frequency anomaly detection function to process the frequency data to remove outliers. The frequency anomaly detection function determines whether it is an outlier based on the standard deviation and threshold of the frequency data.

[0010] Perform real-time sliding clock error forecasts, use fitted phase data and threshold ranges to remove outliers and update forecast epochs.

[0011] Preferably, the calculation formula for converting the clock difference phase data into frequency data is: i =y i -y i-1

[0012] Among them, f i Represents the frequency data of the i-th epoch, y i represents the phase data of the i-th epoch, y i-1 Represents the phase data for the previous epoch.

[0013] Preferably, updating the forecast epoch comprises the following steps:

[0014] Import the data of the first 40 epochs for preliminary fitting and obtain 40 fitting residuals;

[0015] The frequency data is replaced by the fitted phase data, and an iterative process is performed to remove gross errors in the fitted data and update the threshold;

[0016] Calculate the root mean square error of the fitting residual as the basis for setting the threshold;

[0017] Determine whether the fitting residual of the forecast epoch exceeds the threshold range. If it exceeds, the data of the epoch is eliminated. If it does not exceed, slide forward one epoch and perform fitting and forecasting again.

[0018] Preferably, the step of replacing the frequency data with the fitted phase data and performing iterative processing to eliminate gross errors in the fitted data and update the threshold specifically includes:

[0019] Calculate the standard deviation of all frequency data. The calculation formula for calculating the standard deviation of all frequency data is:

[0020] Among them, ave is the average value of all frequency data, and n is the total number of frequency data;

[0021] Remove the frequency data from f m0 The maximum value that satisfies the condition f m0 =MAX(fabs(f n -ave)), calculate the updated sigma again;

[0022] Different thresholds are designed based on different satellite orbit types and in conjunction with the updated sigma;

[0023] Determine fabs(f m0 -ave) is greater than the threshold, if so, it is considered that f m0 is the frequency anomaly point;

[0024] Continue to iterate the next frequency maximum point f m1 , until no epoch exceeds the threshold.

[0025] Preferably, the calculation formula of the threshold is:

[0026] Among them, μ is a constant, set to 3; η orbit It is an empirical threshold set according to the satellite orbit type.

[0027] Preferably, for MEO / IGSO / GEO satellite orbits, the empirical thresholds are 0, 0.016*10 -9 , 0.033*10 -9 .

[0028] Preferably, the calculation formula for calculating the root mean square error of the fitting residual as the basic step for setting the threshold is:

[0029] Among them, y i Indicates the actual phase data, x i represents the fitted phase data, and n1 represents the number of data points.

[0030] The present invention also provides a device for a real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data, comprising:

[0031] A data receiving module is used to receive satellite clock error phase data;

[0032] A frequency conversion module, used to convert clock difference phase data into frequency data;

[0033] Frequency anomaly detection module, used to determine whether it is an anomaly point based on the standard deviation and threshold of the frequency data;

[0034] The prediction module is used to perform real-time sliding clock error prediction, including the calculation of fitting phase data and the elimination of outliers.

[0035] The present invention also provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method when executing the computer program.

[0036] The present invention also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the method described above is implemented.

[0037] The present invention provides a real-time sliding ultra-short-term forecasting model algorithm based on frequency data and phase data. It has the following beneficial effects:

[0038] 1. The present invention links the elimination of clock data outliers with the calculation of clock error prediction values ​​to achieve the detection and elimination of real-time rolling clock data outliers, and calculates real-time clock error prediction values ​​to correct clock frequency deviations. This can improve the accuracy, real-time performance and data stability of the forecast, and at the same time has flexibility and is applicable to various satellite orbit types.

[0039] 2. The present invention improves the accuracy of the forecast by fitting phase data and judging the threshold range, eliminating data points with large fitting residuals, and adopting a real-time sliding forecasting method, which can update the forecast results in time and adapt to dynamically changing data. In addition, the threshold is set according to the satellite orbit type, so that the forecast model can adapt to different types of satellite orbits, improving the adaptability and flexibility of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] FIG1 is a schematic diagram of a process flow of the present invention;

[0041] FIG2 is a second schematic diagram of the process of the present invention;

[0042] FIG3 is a schematic diagram of the structure of the device of the present invention;

[0043] FIG4 is a schematic diagram of the computer device structure of the present invention.

[0044] Among them, 100 is a data receiving module; 200 is a frequency conversion module; 300 is a frequency anomaly detection module; 400 is a forecast module; 40 is a computer device; 41 is a processor; 42 is a memory; and 43 is a storage medium. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the present specification. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] To overcome the limitations of traditional methods, this paper proposes a real-time sliding ultra-short-term prediction model designed to improve the accuracy, real-time performance, data stability, and flexibility of satellite clock error prediction. This model, based on real-time satellite clock error phase data, converts it into frequency data, uses a frequency anomaly detection function to remove outliers, and then performs real-time sliding clock error prediction. Compared with traditional methods, this model is more adaptable to dynamically changing data and provides more accurate and real-time satellite clock error prediction results.

[0047] Referring to Figures 1 and 2 , an embodiment of the present invention provides a real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data, comprising steps S1 to S3;

[0048] Wherein, step S1, converting the clock difference phase data into frequency data;

[0049] Specifically, the calculation formula for converting clock error phase data into frequency data in this embodiment is: i =y i -y i-1

[0050] Among them, f i Represents the frequency data of the i-th epoch, y i represents the phase data of the i-th epoch, y i-1 Represents the phase data of the previous epoch;

[0051] Converting clock error phase data into frequency data makes the data easier to process and analyze, and can provide information on the rate of change of clock errors, which is of great significance for forecasting.

[0052] Step S2: Process the frequency data using a frequency anomaly detection function to remove outliers, wherein the frequency anomaly detection function determines whether an outlier is present based on the standard deviation and threshold of the frequency data;

[0053] Specifically, the frequency data is processed using a frequency anomaly detection function to remove outliers. Based on the standard deviation of the frequency data and a preset threshold, if the frequency data of a certain epoch exceeds the threshold, it is identified as an outlier and removed.

[0054] Eliminate outliers in frequency data, reduce noise interference, and improve data accuracy and reliability. Through anomaly detection, abnormal data caused by equipment failure, signal interference or other factors can be eliminated to ensure the accuracy of the forecast model.

[0055] Step S3: Perform real-time sliding clock error prediction, use the fitted phase data and threshold range to remove outliers and update the forecast epoch;

[0056] Specifically, the specific steps of step S3 are as follows:

[0057] Step S31: import the data of the first 40 epochs for preliminary fitting to obtain 40 fitting residuals;

[0058] Step S32: Replace the frequency data with the fitted phase data, and perform iterative processing to eliminate gross errors in the fitted data and update the threshold;

[0059] Specifically, the specific steps of step S32 in this embodiment are as follows:

[0060] Step S321: Calculate the standard deviation of all frequency data. The formula for calculating the standard deviation of all frequency data is:

[0061] Among them, ave is the average value of all frequency data, and n is the total number of frequency data;

[0062] Step S322: Remove the frequency data from the m0 The maximum value that satisfies the condition f m0 =MAX(fabs(f n -ave)), calculate the updated sigma again;

[0063] Step S323: Different thresholds are designed according to different satellite orbit types and in conjunction with the updated sigma. The threshold formula in this embodiment is as follows:

[0064] Among them, μ is a constant, set to 3; η orbit It is an empirical threshold set according to the satellite orbit type.

[0065] Specifically, for MEO / IGSO / GEO satellite orbits, the empirical thresholds are 0, 0.016*10 -9 , 0.033*10 -9 .

[0066] Step S324, determine fabs(f m0 -ave) is greater than the threshold, if so, it is considered that f m0 is the frequency anomaly point;

[0067] Step S325, continue to iterate the next frequency maximum point f m1 , until no epoch exceeds the threshold.

[0068] S33. Calculate the root mean square error of the fitting residual as the basis for setting the threshold. The root mean square error of the fitting residual calculated in this embodiment is expressed as follows:

[0069] Among them, yi Indicates the actual phase data, x i represents the fitted phase data, n1 represents the number of data points;

[0070] S34. Determine whether the fitting residual of the forecast epoch exceeds a threshold range. If so, remove the data of the epoch. If not, slide forward one epoch and perform fitting and forecasting again.

[0071] The embodiment of the present invention improves the accuracy of the forecast by fitting phase data and judging the threshold range, eliminating data points with large fitting residuals, and adopting a real-time sliding forecasting method to timely update the forecast results and adapt to dynamically changing data. In addition, the threshold is set according to the satellite orbit type, so that the forecast model can adapt to different types of satellite orbits, thereby improving the adaptability and flexibility of the algorithm.

[0072] In general, the present invention links the elimination of clock data outliers with the calculation of clock error prediction values ​​to achieve the detection and elimination of real-time rolling clock data outliers, and calculates real-time clock error prediction values ​​to correct clock frequency deviations. This can improve the accuracy, real-time performance and data stability of the forecast, and at the same time has flexibility and is applicable to various satellite orbit types.

[0073] The device for the real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data described below and the real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data described above can correspond to each other.

[0074] Referring to FIG. 3 , the present invention provides a device for a real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data, comprising:

[0075] Data receiving module 100, which is used to receive satellite clock phase data. This data can be obtained through a satellite navigation system or other related equipment and transmitted to a subsequent processing module for analysis and prediction;

[0076] Frequency conversion module 200, which converts clock phase data into frequency data. By processing and calculating the clock phase data, corresponding frequency data can be obtained. This conversion provides information about the rate of change of the clock error, providing input for subsequent prediction modules.

[0077] Frequency anomaly detection module 300, which is used to determine whether a point is an outlier based on the standard deviation and threshold of the frequency data. By performing statistical analysis on the frequency data, the standard deviation of the frequency data can be calculated and compared with a pre-set threshold. If the frequency data exceeds the threshold range, it is determined to be an outlier and requires further processing;

[0078] Prediction module 400 is used to perform real-time sliding clock error forecasting. This module includes two main steps: calculating fitted phase data and removing outliers. First, by fitting the phase data, the trend and variation of the clock error can be determined. Then, based on a pre-set threshold and the fitting results, data points with large fitting residuals are removed to improve forecast accuracy. This forecast module can update forecast results in real time to adapt to dynamically changing data.

[0079] The device of this embodiment can be used to execute the above method embodiment, and its principles and technical effects are similar, so they will not be repeated here.

[0080] Please refer to FIG4 . The present invention further provides a computer device 40 , including a processor 41 and a memory 42 . The memory 42 stores a computer program executable by the processor. When the computer program is executed by the processor, the above method is performed.

[0081] The present invention further provides a storage medium 43 on which a computer program is stored. When the computer program is run by the processor 41 , the above method is executed.

[0082] Among them, the storage medium 43 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 read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0083] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data, characterized in that, The following steps are involved: Convert clock difference phase data into frequency data; The frequency data is processed using a frequency anomaly detection function to remove outliers, wherein the frequency anomaly detection function determines whether it is an outlier based on the standard deviation and threshold of the frequency data; Perform real-time sliding clock error forecast, use the fitted phase data and threshold range to remove outliers and update the forecast epoch.

2. The real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data according to claim 1, characterized in that, The calculation formula for converting clock difference phase data into frequency data is: f i = y i - y i-1 Among them, f i represents the frequency data of the i-th epoch, and y i represents the phase data of the i-th epoch, and y i-1 represents the phase data of the previous epoch.

3. A real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data according to claim 1, characterized in that, The update of the forecast epoch includes the following steps: Import the data of the first 40 epochs for preliminary fitting and obtain 40 fitting residuals; The frequency data is replaced with the fitted phase data, and an iterative process is performed to eliminate the gross errors in the fitted data and update the threshold; Calculate the root mean square error of the fitting residuals as the basis for setting the threshold; Determine whether the fitting residual of the forecast epoch exceeds the threshold range. If so, remove the forecast epoch. If the data of the epoch is not exceeded, slide forward one epoch and perform fitting and forecasting again.

4. A real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data according to claim 3, characterized in that The step of replacing the frequency data with the fitting phase data and performing iterative processing to eliminate the gross errors in the fitting data and update the threshold specifically includes: Calculate the standard deviation of all frequency data, and the calculation formula for the standard deviation of all frequency data is: Among them, ave is the average value of all frequency data, and n is the total number of frequency data; Remove the value in the frequency data that is farthest from f m0 with the maximum value, satisfying the condition f m0 = MAX(fabs(f n - ave)), and recalculate the updated sigma; Different thresholds are designed according to different satellite orbit types and in combination with the updated sigma; Judge whether fabs(f m0 - ave) is greater than the threshold. If it holds, then f m0 is considered as a frequency anomaly point; Continue to iterate to the next maximum frequency point f m1 until no epoch exceeds the threshold value 5. The real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data according to claim 4, characterized in that, The calculation formula for the threshold value is as follows: where μ is a constant, set to 3; η orbit is an empirical threshold set according to the satellite orbit type.

6. The real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data according to claim 5, characterized in that, For MEO / IGSO / GEO satellite orbits, the empirical thresholds are 0, 0.016*10 -9 , 0.033*10 -9 .

7. A real-time sliding ultra-short-term prediction model algorithm based on frequency data and phase data according to claim 3, characterized in that, The calculation formula for taking the root mean square error of the calculated fitting residuals as the basic step for setting the threshold is as follows: Among them, y i represents the actual phase data, and x i represents the fitted phase data, and n1 represents the number of data points.

Images