A method and device for predicting propagation delay of long-wave timing signal

Abstract

The present invention discloses a method and device for predicting the propagation delay of a long-wave timing signal, which belongs to the field of land-based long-wave timing technology. The method continuously obtains the delay data of the long-wave timing signal in the signal receiving area. Once the receiver of the long-wave timing signal fails and cannot receive the long-wave timing signal, a measurement value data set is constructed, and a long-wave propagation delay mathematical model is constructed at the same time. The long-wave propagation delay prediction model is obtained in combination with the measurement value data set, and subsequent delay data are predicted by the long-wave propagation delay prediction model. The method deeply analyzes the time-frequency characteristics of the propagation delay of the long-wave timing signal, utilizes the daily cycle phenomenon exhibited by the long-wave timing signal in the frequency domain, and constructs a quadratic polynomial prediction model containing a periodic term as a long-wave propagation delay mathematical model. The mathematical model can solve the problem of unified modeling of propagation delay under the influence of multiple complex factors, and helps to more accurately predict the propagation delay of the long-wave timing signal.

Description

Technical Field

[0001] The invention belongs to the technical field of land-based long-wave timing service, and in particular relates to a method and device for predicting the propagation delay of a long-wave timing service signal. Background Art

[0002] In many fields, the demand for precise time synchronization is increasing. Based on this, the satellite-based Beidou global satellite navigation system (referred to as Beidou-3 system) and the land-based Loran-C (Loran C) long-wave timing system were designed and established. After its completion, the Beidou-3 system has achieved global coverage, and its timing accuracy has reached 10ns (nanoseconds), providing high-precision time and location services to users around the world. However, the signals of the satellite navigation system are easily affected by factors such as obstruction, interference and deception, so that in some application scenarios, the reliability of its signals may be threatened.

[0003] In contrast, the long-wave timing system shows significant advantages. Since the long-wave timing signal has strong diffraction ability, it can effectively penetrate large buildings, and its power is much higher than that of satellite signals, which enhances the anti-interference and stability of the long-wave timing system. However, the propagation path of the long-wave timing signal is affected by the complexity of the terrain and meteorological conditions, making its delay difficult to accurately predict. Therefore, accurately studying and predicting the delay characteristics of land-based long-wave timing signals is of great practical significance for improving the reliability and accuracy of the system.

[0004] The propagation delay of long-wave timing signals is affected by many factors, including meteorological conditions (such as temperature, humidity, atmospheric pressure, seasonal changes and diurnal changes), electrical parameters of the propagation medium, and terrain complexity. These factors interact with each other, causing the propagation delay of long-wave timing signals to vary significantly under different conditions. However, existing research on the delay data of long-wave timing signals is mostly based on meteorological factors or radio wave transmission characteristics, which is difficult to fully reflect the overall characteristics of long-wave timing signal propagation. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and provide a method and device for predicting the propagation delay of a long-wave timing signal to solve the problem that it is difficult to fully reflect the propagation characteristics of the long-wave timing signal in the prior art, resulting in inaccurate delay data of the long-wave timing signal.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] The first aspect of the present invention provides a method for predicting the propagation delay of a long-wave timing signal, comprising:

[0008] Continuously receive long-wave timing signals and calculate signal delay data. When no long-wave timing signals are received, construct a measurement value data set with the existing signal delay data.

[0009] Constructing a long wave propagation delay mathematical model, wherein the long wave propagation delay mathematical model is composed of a constant term, a first order term, a second order term, a periodic term and a sum of fitting residuals;

[0010] Based on the measurement data set, a constant term, a linear term coefficient, a quadratic term coefficient, a periodic term parameter and a fitting residual are obtained, and substituted into a long-wave propagation delay mathematical model to obtain a long-wave propagation delay prediction model;

[0011] The signal delay data of the long-wave timing signal at subsequent moments is predicted by the long-wave propagation delay prediction model.

[0012] Preferably, the amount of signal delay data in the measurement value data set is a fixed value, and the signal delay data in the measurement value data set is the signal delay data measured closest to the subsequent moment.

[0013] Preferably, the measurement value data set is updated via a sliding window, and the last data in the sliding window is the signal delay data predicted at a previous moment of a subsequent moment.

[0014] Preferably, during the updating process of the measurement value data set, the amount of signal delay data in the measurement value data set remains unchanged.

[0015] Preferably, during the updating process of the measurement data set, the construction formula of the measurement data set is:

[0016]

[0017] in, Indicates the predicted time, represents a data set of measurements, represents the measured value at the first moment, Indicates The measured value at the moment, Indicates the newly obtained The predicted value at time, Indicates the newly obtained The predicted value at time, l is the length of the sliding window, and represents the sum.

[0018] Preferably, the long wave propagation delay mathematical model is:

[0019]

[0020] In the formula, t For the moment, i is the ordinal number of the periodic term, is the number of periodic terms, for tThe predicted delay at time p Indicates prediction, represent the constant term, the coefficient of the first term, and the coefficient of the second term respectively. is the i-th periodic term, , , Respectively represent The amplitude, frequency and phase of the periodic terms, express t The fitting residual after adding the periodic term at time instant, sin() represents the sine function.

[0021] Preferably, the measurement data set is represented as follows:

[0022]

[0023] in: Indicates the measured value, Indicates The signal delay data at time, T represents the transpose of the matrix, A represents the matrix coefficients, X Represents the parameter to be estimated.

[0024] Preferably, the matrix coefficients A The form is:

[0025]

[0026] in, Indicates time, Indicates The frequency of a periodic term, sin() represents the sine function, and cos() represents the cosine function;

[0027] The parameters to be estimated X The calculation formula is:

[0028]

[0029] in, represent the constant term, the coefficient of the first term, and the coefficient of the second term respectively. Indicates The coefficients of the periodic terms; Indicates The phase of the periodic term, T Represents the transpose of a matrix;

[0030] The parameters to be estimated X Solve the following equation by the least square method;

[0031]

[0032] in, represents the matrix coefficients, Representation Matrix The transposed matrix of Represents a measurement dataset.

[0033] Preferably, the signal delay data in the measurement value data set is processed by a filter.

[0034] The second aspect of the present invention discloses a long-wave timing signal propagation delay prediction device, comprising:

[0035] A data module is used to continuously receive long-wave timing signals and calculate signal delay data. When no long-wave timing signals are received, the existing signal delay data is used to construct a measurement value data set.

[0036] Establishing a module for constructing a long wave propagation delay mathematical model, wherein the long wave propagation delay mathematical model is composed of a constant term, a first order term, a second order term, a periodic term and a sum of fitting residuals;

[0037] A calculation module, used for obtaining a constant term, a linear term coefficient, a quadratic term coefficient, a periodic term parameter and a fitting residual based on the measurement value data set, and substituting them into a long-wave propagation delay mathematical model to obtain a long-wave propagation delay prediction model;

[0038] The prediction module is used to predict the signal delay data of the long-wave timing signal at a subsequent moment through the long-wave propagation delay prediction model.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The invention discloses a method for predicting the propagation delay of a long-wave timing signal. The method continuously obtains the delay data of the long-wave timing signal in the receiving area of ​​the signal. Once the receiver of the receiving signal fails and cannot receive the long-wave timing signal, the delay data is collected to construct a measurement value data set, and a long-wave propagation delay mathematical model is constructed at the same time. The long-wave propagation delay prediction model is obtained in combination with the measurement value data set, and the subsequent delay data is predicted by the long-wave propagation delay prediction model. The method deeply analyzes the time-frequency characteristics of the propagation delay of the long-wave timing signal, makes full use of the daily cycle phenomenon exhibited by the long-wave timing signal in the frequency domain, constructs a quadratic polynomial prediction model of the periodic term as the mathematical model of the long-wave propagation delay, and obtains the long-wave propagation delay prediction model after solving. The mathematical model can solve the problem of unified modeling of propagation delay under the influence of multiple complex factors, and helps to more accurately predict the propagation delay of the long-wave timing signal, thereby improving the engineering realization efficiency of the long-wave timing signal. This not only helps to promote the application of land-based long-wave timing systems, but also has important significance for the development of related fields such as communication, navigation and measurement.

[0041] Furthermore, it was found in practice that the long-wave propagation delay prediction model based on the periodic term performed well in short-term (less than 1 hour) prediction, but the long-term (prediction time greater than 1 hour) prediction effect was not accurate enough. In order to further improve the prediction accuracy, the present invention combines the sliding window technology and the periodic term to continuously update the measurement data set, so that the prediction results are significantly improved in both short-term and long-term prediction accuracy, and the consistency with the measured results is greatly improved. Verification found that the proportion of short-term predictions with an absolute error of less than 50 nanoseconds is as high as 96.22%, and the proportion of 24-hour predictions with an absolute error of less than 50 nanoseconds is as high as 80%. These improvements not only improve the accuracy of the prediction model, but also improve the efficiency of engineering implementation, which is of great significance to promoting the large-scale application of land-based long-wave timing systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of the method for predicting the propagation delay of a long-wave timing signal of the present invention;

[0043] Figure 2 is a block diagram of a long wave propagation delay measurement system of the present invention;

[0044] Figure 3 A diagram of a long-wave timing signal propagation delay prediction device according to the present invention;

[0045] Figure 4 is a block diagram of the long wave propagation delay measurement system in Example 1;

[0046] Figure 5 It is a flow chart of the method for predicting the propagation delay of a long-wave timing signal in Example 1;

[0047] Figure 6 This is a time series diagram of the 4-day long wave delay data of Region 1;

[0048] Figure 7 This is the time series diagram of the 4-day long wave delay data of Region 2;

[0049] Figure 8 This is the spectrum diagram of the 4-day data of the long wave delay in area 1;

[0050] Fig. 9 This is the spectrum diagram of the 4-day data of the long wave delay of Region 2;

[0051] Fig.10 It is the long wave delay model fitting result and the measured data waveform;

[0052] Fig.11 It is the 24-hour prediction result and prediction error diagram of long wave propagation delay with periodic term in area 1;

[0053] Fig.12It is the 24-hour prediction result and prediction error diagram of long wave propagation delay with periodic term in region 2;

[0054] Fig.13 It is the 24-hour prediction result and prediction error diagram of the long wave propagation delay prediction model based on sliding window and periodic term in region 1;

[0055] Fig.14 It is the 24-hour prediction result and prediction error diagram of the long wave propagation delay prediction model based on sliding window and periodic term in region 2.

[0056] Among them, 1. Long wave transmitter; 2. Loran-C long wave receiver; 3. Long wave antenna; 4. GPS receiver; 5. Satellite antenna; 6. Time interval counter; 7. Computer; 8. Satellite. DETAILED DESCRIPTION

[0057] In the following, the terms "first", "second", "third", and "fourth" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, a feature defined as "first", "second", "third", and "fourth" may explicitly or implicitly include one or more of the features.

[0058] The co-shooting method provided in the embodiment of the present application can be applied to terminal devices such as mobile phones, tablet computers, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptop computers, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA), etc. The embodiment of the present application does not impose any restrictions on the specific type of the terminal device.

[0059] It should be noted that the terms "first", "second", etc. in the specification and drawings of the present invention are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] First, some abbreviations of the present invention are explained.

[0061] Loran-C is the abbreviation of Loran C;

[0062] The Loran-C long-wave timing system, short for the Loran-C navigation system, is the abbreviation of the long-range land-based radio navigation and positioning system. It is also known as the Loran-C navigation system or the pulse phase distance hyperbola navigation system.

[0063] GPS is the abbreviation of Global Positioning System.

[0064] Hampel filter is a Hampel filter.

[0065] PPS signal, short for Pulse Per Second, is a signal that generates one pulse per second.

[0066] See also Figure 1 The first aspect of the present invention discloses a method for predicting the propagation delay of a long-wave timing signal, the method comprising the following steps:

[0067] S101, continuously receiving long-wave timing signals and calculating signal delay data, and when no long-wave timing signals are received, constructing a measurement value data set with the existing signal delay data;

[0068] S102, constructing a long wave propagation delay mathematical model, wherein the long wave propagation delay mathematical model is composed of a constant term, a linear term, a quadratic term, a periodic term, and a sum of fitting residuals;

[0069] S103, obtaining a constant term, a linear term coefficient, a quadratic term coefficient, a periodic term parameter and a fitting residual based on the measurement value data set, and substituting them into a long-wave propagation delay mathematical model to obtain a long-wave propagation delay prediction model;

[0070] S104, predicting the signal delay data of the long-wave timing signal at a subsequent time by using the long-wave propagation delay prediction model.

[0071] It should be understood that, in the above steps, S101 and S102 can be adjusted in order with each other, and the long-wave propagation delay mathematical model in S102 can be constructed in advance in the computer 7, and when needed, it can be directly called out from the relevant modules of the computer 7 for use.

[0072] In some implementation modes of the present invention, the signal delay data in the measurement value data set is the measured signal delay data closest to the subsequent time instant.

[0073] For further information, see Figure 2When obtaining the measurement value data set, the 1PPS signal output by the GPS receiver 4 is used as a reference, and a long-wave propagation delay measurement system is set up in the signal receiving area to monitor the propagation delay data of the Loran-C signal emitted by the long-wave transmitter 1 in the signal transmitting area. Each long-wave propagation delay measurement system includes: a Loran-C long-wave receiver 2 and a long-wave antenna 3 connected thereto, a GPS receiver 4 and a satellite antenna 5 connected thereto wirelessly, a time interval counter 6 and a computer 7. After the signal connection is normal, the computer 7 will start to collect the signal delay data. In the receiving area of ​​the computer 7, the time interval counter 6 is used to measure the time difference between the 1PPS signal output by the Loran-C long-wave receiver 2 and the 1PPS signal output by the GPS receiver 4 as the long-wave propagation delay measurement value data; the GPS receiver 4 receives the signal sent by the satellite 8 through the satellite antenna 5. Specifically, after the GPS receiver 4 receives the signal sent by the satellite 8, the GPS receiver 4 outputs the time information and the 1PPS signal. This part uses the 1PPS signal output by the GPS receiver 4 as the reference signal, and the 1PPS signal output by the Loran-C long-wave receiver 2 as the measured signal. The time interval counter 6 is used to measure the time difference data between the two 1PPS signals as the long-wave propagation delay measurement value.

[0074] As a preferred solution, since there are many uncontrollable factors in the process of measuring the signal delay data for a long time by the time interval counter 6, the signal delay data will have abnormal values ​​such as gross errors. If the abnormal values ​​are substituted into the actual measurement operation, it is bound to cause large errors. Therefore, it is necessary to detect and correct the abnormal values ​​before performing subsequent analysis and calculation to make the prediction more accurate. The present invention uses a Hampel filter to pre-process the signal delay data of the collected long-wave timing signal.

[0075] Furthermore, the Hampel filter is a filter based on the median and median absolute deviation ( MAD, The Hampel filter is designed to identify and remove outliers in time series data, and is particularly suitable for time series data. The core of the Hampel filter lies in the calculation of the median and the solution of the median absolute deviation. The median represents the middle value of the data, while the median absolute deviation measures the degree of dispersion between the data points and the median.

[0076] Median: For Dataset with data points W , the median calculation formula is as follows:

[0077] (1)

[0078] Median absolute deviation: is the median absolute deviation of each data point. Wi With the dataset W The median of the absolute differences between the medians of :

[0079] (2)

[0080] in, W i Represents each data point.

[0081] Hampel filter outlier judgment criteria: If the data point W i It is identified as an outlier, and the absolute deviation of the outlier value divided by the median absolute deviation exceeds the threshold, so it is replaced with the data set W otherwise, keep the original value of the data point.

[0082] (3)

[0083] Threshold: Define a threshold, usually a multiple of the absolute deviation from the median (e.g. 2.5 or 3). This threshold is used to determine whether a data point is an outlier.

[0084] In some embodiments of the present invention, based on the characteristics of periodic fluctuation of long-wave delay, the present invention constructs a mathematical model of long-wave propagation delay with periodic terms, which is expressed as follows:

[0085] (4)

[0086] In the formula, t For the moment, i is the ordinal number of the periodic term, is the number of periodic terms, for t The predicted delay at time p Indicates prediction, represent the constant term, the coefficient of the first term, and the coefficient of the second term respectively. is the i-th periodic term, where , , Respectively represent The amplitude, frequency and phase of the periodic term are the parameters of the periodic term. express t The fitting residual after adding the periodic term at time instant, sin() represents the sine function.

[0087] The specific periodic terms are obtained through FT (Fourier Transform) time-frequency analysis.

[0088] In some embodiments of the present invention, formula (4) can be expressed in matrix form as follows:

[0089] (5)

[0090] in: Indicates the measured value, the subscript M indicates measurement, Indicates Signal delay data at time X represents the parameters to be estimated, A represents the coefficient matrix.

[0091] Parameters to be estimated X The calculation formula is:

[0092] (6)

[0093] in, represent the constant term, the coefficient of the first term, and the coefficient of the second term respectively. Indicates The amplitude of the periodic term; Respectively represent The phase of a periodic term.

[0094] Coefficient Matrix A The calculation formula is:

[0095] (7)

[0096] in, Indicates time, Indicates The frequency of a periodic term, sin() represents the sine function, and cos() represents the cosine function.

[0097] To find the parameters to be estimated , the matrix equation (5) can be transformed into equation (8), where Representation Matrix The transposed matrix of .

[0098] (8)

[0099] Formula (8) can be solved by the least squares method to obtain X.

[0100] In some embodiments of the present invention, the long wave propagation delay prediction model described above can theoretically predict the time delay at any time. The corresponding long-wave propagation delay, however, was found in the verification that the short-term prediction results were good and the long-term deviations were large. Therefore, in order to improve the prediction accuracy and make full use of historical data, improvement measures were introduced based on the original model. Specifically, the new prediction value was incorporated into the measured value data set, and the data set was updated through a sliding window. The last data in the sliding window was the signal delay data of the long-wave timing signal predicted at the previous moment of the subsequent moment.

[0101] During the update process, the measurement data set is constructed based on the following principles:

[0102] (9)

[0103] in, Indicates the predicted time, represents a data set of measurements, represents the measured value at the first moment, Indicates The measured value at the moment, Indicates the newly obtained The predicted value at time, Indicates the newly obtained The predicted value at time, l is the length of the sliding window, preferably, l At least 86400 points, and represents the sum.

[0104] By updating the measured value data set, a new coefficient matrix A is obtained, and then new parameters to be estimated are obtained. X and a long wave propagation delay prediction model based on a sliding window and with a periodic term; as the prediction proceeds, the long wave propagation delay prediction model is continuously updated to make further predictions.

[0105] See also Figure 3 The second aspect of the present invention discloses a long-wave timing signal propagation delay prediction device, comprising:

[0106] A data module is used to continuously receive long-wave timing signals and calculate signal delay data. When no long-wave timing signals are received, the existing signal delay data is used to construct a measurement value data set.

[0107] Establishing a module for constructing a long wave propagation delay mathematical model, wherein the long wave propagation delay mathematical model is composed of a constant term, a first order term, a second order term, a periodic term and a sum of fitting residuals;

[0108] A calculation module, used for obtaining a constant term, a linear term coefficient, a quadratic term coefficient, a periodic term parameter and a fitting residual based on the measurement value data set, and substituting them into a long-wave propagation delay mathematical model to obtain a long-wave propagation delay prediction model;

[0109] The prediction module is used to predict the signal delay data of the long-wave timing signal at a subsequent moment through the long-wave propagation delay prediction model.

[0110] The following is further described in conjunction with specific embodiments.

[0111] Example 1

[0112] See also Figure 4 In this embodiment, a signal transmission area and two signal receiving areas are set up, namely area 1 and area 2. Taking the 1PPS signal output by the GPS receiver 4 as a reference, long-wave propagation delay measurement systems are respectively set up in area 1 and area 2 to monitor the propagation delay data of the Loran-C signal from the long-wave transmitter 1 in the signal transmission area. Each long-wave propagation delay measurement system includes: a Loran-C long-wave receiver 2 and a long-wave antenna 3 connected thereto, a GPS receiver 4 and a satellite antenna 5 connected thereto wirelessly, a time interval counter 6 and a computer 7.

[0113] See also Figure 5 The process of using the long-wave propagation delay measurement system to predict the long-wave timing signal propagation delay is as follows:

[0114] S201: The long wave propagation delay measurement systems in regions 1 and 2 are Figure 4 The connection mode shown is connected, and after the signal connection is normal, the computer 7 will start to collect signal delay data. The time difference between the 1PPS signal output by the Loran-C long wave receiver 2 and the 1PPS signal output by the GPS receiver 4 is measured using the time interval counter 6, which is the signal delay data, and these data are transmitted to the computer 7.

[0115] S202, pre-processing the abnormal values ​​in the signal delay data measured by the time interval counter 6.

[0116] In this embodiment, the Hampel filter is used to pre-process the abnormal values. In this embodiment, the threshold is uniformly set to 3, that is, The long-wave propagation delay data collected in Region 1 and Region 2 are pre-processed three times.

[0117] S203, based on the data preprocessing result of S202, respectively perform time-frequency characteristic analysis on the signal delay data of region 1 and region 2 at different long wave propagation distances.

[0118] (1) By visually observing the changing trend of signal delay data in the time domain, it is found that the signal delay data of long-wave timing signals at different propagation distances show a significant daily cycle phenomenon over time. Figure 6 and Figure 7As shown, it can be found that the signal delay data of the long-wave timing signals in Regions 1 and 2 for 4 days have obvious daily cycles in their changes over time, which is consistent with the "daily cycle" phenomenon of long-wave propagation delay.

[0119] (2) In order to gain a deeper understanding of this phenomenon, the time domain signal delay data is further converted into frequency domain data through Fourier Transform (FT) to observe its frequency domain characteristics. The spectrum of the signal delay data between the two locations is as follows: Figure 8 and Fig. 9 The results show that the signal delay data also exhibits a daily cycle phenomenon in the frequency domain. This analysis verifies the consistency of the daily cycle phenomenon at different propagation distances and provides basic data for subsequent modeling and prediction.

[0120] by Figure 8 and Fig. 9 Take point A as an example, the frequency , the corresponding period If the vertical axis amplitude is 10 as the threshold, the periodic items of region 1 are: 24 hours, 12 hours and 8 hours, corresponding to Figure 8 Points A, B, and C in the middle; the periodic items of region 2 are: 24 hours, 12 hours, and 8 hours, corresponding to Fig. 9 Points A, B, and C. From this, we can see that the most obvious one is the 24-hour daily cycle, which corresponds to point A in the two figures, and also conforms to the daily cycle phenomenon of long-wave propagation delay.

[0121] S204, a mathematical model of long wave propagation delay with periodic term is established according to the time-frequency characteristics of long wave propagation delay, as shown in the above formula (4). Formula (4) is expressed in matrix form as formula (5). In order to obtain the estimated parameter in formula (5) , then transform equation (5) into equation (8), and finally convert the preprocessed measurement value in S202 Substitute into formula (8), and obtain the parameters to be estimated through formula (8), and finally obtain the long wave propagation delay prediction model, which is model 1.

[0122] Through the long-wave propagation delay prediction model, the fitted long-wave propagation delay data and measured data are obtained, such as Fig.10 As shown in the figure, the results show that the fitting results of the long-wave propagation delay prediction model with periodic terms are in good agreement with the measured results, and the mean absolute error and root mean square error of the fitting error are 0.005182ns and 8.9701ns respectively. The fitting results show that the long-wave propagation delay prediction model shows good consistency in the actual data fitting and is highly consistent with the measured results.

[0123] Furthermore, through the long wave propagation delay prediction model, it is theoretically possible to predict the The corresponding long wave propagation delay is predicted for 24 hours for region 1 and region 2 respectively. The results are as follows Fig.11 and Fig.12 The results show that the short-term prediction results are good, but the long-term deviations are good.

[0124] S205, based on the analysis of the shortcomings of the existing model in S204, in order to improve the prediction accuracy and make full use of historical data, the new prediction value is included in the measurement value data set, and the prediction is performed in the form of a sliding window. The newly predicted delay data is continuously used as part of the measurement value data set. The window length is 86400. Formula (4) is expressed in matrix form as Formula (5). In order to obtain the parameters to be estimated in Formula (5), , then transform equation (5) into equation (8), and finally convert the preprocessed measurement value in S202 Substituting into formula (8), we finally obtain the long wave propagation delay prediction model, which is based on the sliding window and is model 2.

[0125] S206, predicting the sth time delay, updating the measurement value data set in S202, and further updating various coefficients and parameters in the long wave propagation time delay prediction model.

[0126] S207, generating a long wave delay prediction sequence.

[0127] Then, the long wave propagation delay between Region 1 and Region 2 was predicted for 24 hours. Fig.13 and Fig.14 As shown in Table 1, the results of the two models are shown in Table 1. Compared with the initial model, the model 2 with the sliding window data update has higher prediction accuracy. The short-term and long-term prediction results are in good agreement with the measured results. The proportion of short-term absolute prediction errors within 1 hour less than 50ns is more than 96.22%, and the proportion of 24-hour absolute prediction errors less than 50ns is close to 80%. The more significant the prediction accuracy improvement effect is. It shows that the quadratic polynomial long-wave propagation delay prediction model based on the sliding window and with periodic terms can achieve high-precision prediction of long-wave propagation delay, which will be of great significance to further improve the timing accuracy of the land-based long-wave timing system. After updating the measurement value data set through the sliding window, the long-wave propagation delay prediction model obtained again shows a significant improvement in prediction accuracy, and as the prediction time increases, the improvement effect of prediction accuracy becomes more obvious. This improvement verifies the effectiveness and advantages of the quadratic polynomial model combining sliding window with periodic terms in long-term prediction.

[0128] Table 1. Error statistics of prediction results of two models at different time lengths

[0129]

[0130] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting the propagation delay of a long-wave timing signal, characterized in that: include: Continuously receive long-wave timing signals and calculate signal delay data. When no long-wave timing signals are received, construct a measurement value data set with the existing signal delay data. Constructing a long wave propagation delay mathematical model, wherein the long wave propagation delay mathematical model is composed of a constant term, a first order term, a second order term, a periodic term and a sum of fitting residuals; Based on the measurement data set, a constant term, a linear term coefficient, a quadratic term coefficient, a periodic term parameter and a fitting residual are obtained, and substituted into a long-wave propagation delay mathematical model to obtain a long-wave propagation delay prediction model; The measured value data set is represented as follows: in: Indicates the measured value, Indicates The signal delay data at time, T represents the transpose of the matrix, A represents the matrix coefficients, X represents the parameter to be estimated; The matrix coefficients A The form is: in, Indicates time, Indicates The frequency of a periodic term, sin() represents the sine function, and cos() represents the cosine function; The parameters to be estimated X The calculation formula is: in, represent the constant term, the coefficient of the first term, and the coefficient of the second term respectively. Indicates The coefficients of the periodic terms; Indicates The phase of the periodic term, T Represents the transpose of a matrix; The parameters to be estimated X Solve the following equation by the least square method; in, represents the matrix coefficients, Representation Matrix The transposed matrix of represents a data set of measured values; The signal delay data of the long-wave timing signal at subsequent moments is predicted by the long-wave propagation delay prediction model.

2. A method for predicting the propagation delay of a long-wave timing signal according to claim 1, characterized in that: The amount of signal delay data in the measurement value data set is a fixed value, and the signal delay data in the measurement value data set is the signal delay data measured closest to the subsequent time point.

3. The method for predicting the propagation delay of a long-wave timing signal according to claim 1, characterized in that: The measurement value data set is updated through a sliding window, and the last data in the sliding window is the signal delay data predicted at the previous moment of the subsequent moment.

4. A method for predicting the propagation delay of a long-wave timing signal according to claim 3, characterized in that: During the updating process of the measurement value data set, the amount of signal delay data in the measurement value data set remains unchanged.

5. The method for predicting the propagation delay of a long-wave timing signal according to claim 3, characterized in that: During the updating process of the measurement data set, the construction formula of the measurement data set is: in, Indicates the predicted time, represents a data set of measurements, represents the measured value at the first moment, Indicates The measured value at the moment, Indicates the newly obtained The predicted value at time, Indicates the newly obtained The predicted value at time, l is the length of the sliding window, and represents the sum.

6. The method for predicting the propagation delay of a long-wave timing signal according to claim 1, characterized in that: The long wave propagation delay mathematical model is: In the formula, t For the moment, i is the ordinal number of the periodic term, is the number of periodic terms, for t The predicted delay at time p Indicates prediction, represent the constant term, the coefficient of the first term, and the coefficient of the second term respectively. is the i-th periodic term, , , Respectively represent The amplitude, frequency and phase of the periodic terms, express t The fitting residual after adding the periodic term at time instant, sin() represents the sine function.

7. The method for predicting the propagation delay of a long-wave timing signal according to claim 1, characterized in that: The signal delay data in the measurement value data set is processed by a filter.

8. A long-wave timing signal propagation delay prediction device for implementing the method described in claim 1, characterized in that: include: A data module is used to continuously receive long-wave timing signals and calculate signal delay data. When no long-wave timing signals are received, the existing signal delay data is used to construct a measurement value data set. Establishing a module for constructing a long wave propagation delay mathematical model, wherein the long wave propagation delay mathematical model is composed of a constant term, a first order term, a second order term, a periodic term and a sum of fitting residuals; A calculation module, used for obtaining a constant term, a linear term coefficient, a quadratic term coefficient, a periodic term parameter and a fitting residual based on the measurement value data set, and substituting them into a long-wave propagation delay mathematical model to obtain a long-wave propagation delay prediction model; The prediction module is used to predict the signal delay data of the long-wave timing signal at a subsequent moment through the long-wave propagation delay prediction model.

Images