A method for predicting the ultra-short-term earth rotation parameters

Abstract

A method for predicting ultra-short-term earth rotation parameters includes the following steps: S1, obtaining basic data of UT1-UTC and LOD, and obtaining the latest basic data of LOD derived from fast and low-latency observations of GNSS technology; S2, removing the leap seconds and tides of UT1-UTC to obtain basic data of UT1R-TAI, and removing the tides of LOD to obtain basic data of LODR; S3, first calculating the first-order difference sequence of UT1R-TAI, then fitting and predicting the periodic term and linear term of LODR using the least squares model, and at the same time, calculating the residual term and the residual term of LODR, then fitting the residual term of LODR using the multivariate autoregressive model, and then calculating the predicted value of UT1R-TAI using the differential recovery operator; S4, adding the leap seconds and tides to obtain the predicted value of UT1-UTC. This design can significantly improve the accuracy of ultra-short-term prediction of UT1-UTC.

Description

Technical Field

[0001] The invention relates to the field of geodesy and satellite navigation technology, and in particular to an ultra-short-term earth rotation parameter prediction method, which is mainly suitable for improving the accuracy of UT1-UTC ultra-short-term prediction. Background Art

[0002] Earth rotation parameters (ERPs), such as UT1-UTC, are necessary for the conversion between celestial and terrestrial reference systems. However, UT1-UTC cannot be observed by global navigation satellite system (GNSS) technology, which is the most difficult to determine in the real-time estimation process. In order to obtain the observed UT1-UTC, the complexity and time-consuming data processing of other space geodetic techniques (such as very long baseline interferometry (VLBI)) always cause time delays. Therefore, the real-time orbit determination of GNSS satellites is completely dependent on the predicted UT1-UTC. In near-real-time and real-time high-precision GNSS applications, the ultra-short-term prediction of UT1-UTC (1 to 10 days) directly affects and limits the performance of rapid and ultra-rapid orbit determination of GNSS satellites. For example, the IGS (International GNSS Service) and iGMAS (International GNSS Monitoring and Assessment System) analysis centers usually use the International Earth Rotation and Reference System Service (IERS) Bulletin-A (such as UT1-UTC) updated weekly to generate rapid and ultra-rapid orbits of GNSS satellites. Since the accuracy of UT1-UTC prediction is much lower than that of the observation part, the prediction method of UT1-UTC deserves to be improved for real-time applications related to reference systems, such as GNSS satellite orbit determination.

[0003] How to improve the accuracy of UT1-UTC ultra-short-term prediction has always been a hot topic of concern and research in the field of geodesy and satellite navigation. At present, in the ultra-short-term prediction methods of UT1-UTC, the relevant information between LOD and UT1-UTC has not been fully explored and utilized, and the characteristics of GNSS technology that can quickly and low-latency derive LOD parameters have not been fully utilized, resulting in low ultra-short-term prediction accuracy of UT1-UTC. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects and problems of low prediction accuracy in the prior art and to provide an ultra-short-term Earth rotation parameter prediction method with high prediction accuracy.

[0005] To achieve the above purpose, the technical solution of the present invention is: a method for predicting ultra-short-term earth rotation parameters, the method comprising the following steps:

[0006] S1. Obtain the UT1-UTC basic data and LOD basic data provided by IERS through files or the Internet, and obtain the latest LOD basic data derived from the fast and low-latency observations of GNSS technology;

[0007] S2. Remove the leap seconds and tides of the UT1-UTC basic data to obtain the UT1R-TAI basic data, and remove the tides of the LOD basic data to obtain the LODR basic data. The time series of the UT1R-TAI basic data and the LODR basic data are divided into periodic terms, linear terms and residual terms.

[0008] S3. Calculate the forecast value of UT1R-TAI basic data at time t+1 The specific steps include:

[0009] S31. Calculate the first-order difference sequence of UT1R-TAI basic data:

[0010]

[0011] Where Δf t ut1 is the first-order difference sequence of UT1R-TAI basic data at time t, f t ut1 is the UT1R-TAI basic data at time t, is the UT1R-TAI basic data at time t-1;

[0012] S32. Fit and predict Δf using the least squares model t ut1 The periodic and linear terms of , the periodic and linear terms of LODR, and at the same time, calculate Δf t ut1 The residual term dΔf t ut1 The residual term dLODR of LODR t ;

[0013] S33. Fitting Δf using a multivariate autoregressive model t ut1 The residual term dΔf t ut1 The residual term dLODR of LODR t , and predict Δf t ut1 The residual term of

[0014] S34. Comprehensive forecast Δf t ut1 The periodic term, linear term and residual term of Δf are calculated t ut1 The forecast value at time t+1

[0015] S35. Calculate the forecast value of UT1R-TAI basic data at time t+1 using the differential recovery operator

[0016]

[0017] S4. Increase The leap second and tide of UT1-UTC are used to obtain the forecast value of UT1-UTC basic data at time t+1.

[0018] In step S1, the latest LOD basic data is obtained by the following method: downloading the LOD basic data of the international GNSS service with a delay of 3 to 9 hours or downloading the LOD basic data of the domestic iGMAS with a delay of 3 to 9 hours or using the LOD basic data of the latest ultra-fast product derived by GNSS technology.

[0019] In step S1, the latest LOD basic data is obtained by ultra-fast LOD parameter solution based on the original GNSS observation data:

[0020] a. Real-time acquisition of the latest 24-hour observation value files, navigation files, a priori earth rotation parameter files and general configuration files for precise determination of GNSS earth rotation parameters;

[0021] b. Establish GNSS observation equation based on GNSS observation data:

[0022]

[0023] In the formula, is the theoretical distance, ρ is the observed distance, Δρ is the observation error, X s is the coordinate of the satellite in the earth-fixed coordinate system, X r is the coordinate of the measuring station in the Earth-fixed coordinate system, P, N, S, and U are the rotation matrices of precession, nutation, rotation, and polar motion from the Earth-fixed coordinate system to the inertial coordinate system, respectively;

[0024] The GNSS observation equation is linearized as follows:

[0025]

[0026] Where LOD0 is the a priori LOD parameter, θ is the Greenwich mean true sidereal time, γ≈1.0027379093, t is the Julian day at the observation time, and t0 is the Julian day at the selected reference time;

[0027] c. After linearization, the GNSS observation equation is used to solve the correction number of the prior LOD parameter LOD0 according to the least squares method, and finally the latest fast and low-latency LOD basic data derived from GNSS technology is obtained.

[0028] In step S2, when the forecast duration is greater than 2 days, the latest LOD basic data is used.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] In the ultra-short-term earth rotation parameter prediction method of the present invention, the relevant information between LOD and UT1-UTC is fully excavated and utilized, and the characteristic that the GNSS technology can quickly and low-latency derive LOD parameters is fully utilized, thereby significantly improving the accuracy of the ultra-short-term prediction of UT1-UTC. Therefore, the present invention improves the accuracy of the ultra-short-term prediction of UT1-UTC. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 The present invention is a flow chart of a method for predicting ultra-short-term earth rotation parameters.

[0032] Figure 2 It is a flow chart for predicting ultra-short-term Earth rotation parameters using the traditional LS / WLS+MAR method and the traditional LS / WLS+AR method.

[0033] Figure 3 is the correlation coefficient statistically calculated in the embodiment of the present invention.

[0034] Figure 4 is the correlation coefficient calculated by traditional statistical methods.

[0035] Figure 5 It is a schematic diagram of the prediction error results of the present invention and the traditional method. DETAILED DESCRIPTION

[0036] The present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0037] See also Figure 1 , a method for predicting ultra-short-term earth rotation parameters, the method comprising the following steps:

[0038] S1. Obtain the UT1-UTC basic data and LOD basic data provided by IERS through files or the Internet, and obtain the latest LOD basic data derived from the fast and low-latency observations of GNSS technology;

[0039] S2. Remove the leap seconds and tides of the UT1-UTC basic data to obtain the UT1R-TAI basic data, and remove the tides of the LOD basic data to obtain the LODR basic data. The time series of the UT1R-TAI basic data and the LODR basic data are divided into periodic terms, linear terms and residual terms.

[0040] S3. Calculate the forecast value of UT1R-TAI basic data at time t+1 The specific steps include:

[0041] S31. Calculate the first-order difference sequence of UT1R-TAI basic data:

[0042]

[0043] Where Δf t ut1 is the first-order difference sequence of UT1R-TAI basic data at time t, f t ut1 is the UT1R-TAI basic data at time t, is the UT1R-TAI basic data at time t-1;

[0044] S32. Fit and predict Δf using the least squares model t ut1 The periodic and linear terms of , the periodic and linear terms of LODR, and at the same time, calculate Δf t ut1 The residual term dΔf t ut1 The residual term dLODR of LODR t ;

[0045] S33. Fitting Δf using a multivariate autoregressive model t ut1 The residual term dΔf t ut1 The residual term dLODR of LODR t , and predict Δf t ut1 The residual term of

[0046] S34. Comprehensive forecast Δf t ut1 The periodic term, linear term and residual term of Δf are calculated t ut1 The forecast value at time t+1

[0047] S35. Calculate the forecast value of UT1R-TAI basic data at time t+1 using the differential recovery operator

[0048]

[0049] S4. Increase The leap second and tide of UT1-UTC are used to obtain the forecast value of UT1-UTC basic data at time t+1.

[0050] In step S1, the latest LOD basic data is obtained by the following method: downloading the LOD basic data of the international GNSS service with a delay of 3 to 9 hours or downloading the LOD basic data of the domestic iGMAS with a delay of 3 to 9 hours or using the LOD basic data of the latest ultra-fast product derived by GNSS technology.

[0051] In step S1, the latest LOD basic data is obtained by ultra-fast LOD parameter solution based on the original GNSS observation data:

[0052] a. Real-time acquisition of the latest 24-hour observation value files, navigation files, a priori earth rotation parameter files and general configuration files for precise determination of GNSS earth rotation parameters;

[0053] b. Establish GNSS observation equation based on GNSS observation data:

[0054]

[0055] In the formula, is the theoretical distance, ρ is the observed distance, Δρ is the observation error, X s is the coordinate of the satellite in the Earth-fixed coordinate system, X r is the coordinate of the measuring station in the Earth-fixed coordinate system, P, N, S, and U are the rotation matrices of precession, nutation, rotation, and polar motion from the Earth-fixed coordinate system to the inertial coordinate system, respectively;

[0056] The GNSS observation equation is linearized as follows:

[0057]

[0058] Where LOD0 is the a priori LOD parameter, θ is the Greenwich mean true sidereal time, γ≈1.0027379093, t is the Julian day at the observation time, and t0 is the Julian day at the selected reference time;

[0059] c. After linearization, the GNSS observation equation is used to solve the correction number of the prior LOD parameter LOD0 according to the least squares method, and finally the latest fast and low-latency LOD basic data derived from GNSS technology is obtained.

[0060] In step S2, when the forecast duration is greater than 2 days, the latest LOD basic data is used.

[0061] The principle of the present invention is described as follows:

[0062] Compared with the traditional LS+MAR technology, this design first obtains the first-order difference sequence of UT1R-TAI basic data, and then restores it to the non-difference UT1R-TAI sequence after LS+MAR forecast. The UT1R-TAI basic data is subjected to first-order difference operation to significantly improve the correlation with LODR. After the correlation is stronger, the information of LODR can be better used in the MAR model to improve the ultra-short-term forecast accuracy of UT1-UTC; at the same time, the LOD data calculated by fast and low-latency observation of GNSS technology is fully utilized to obtain the improved LODR fitting residual, and then the ultra-short-term forecast accuracy of UT1-UTC is improved based on the improved LODR fitting residual.

[0063] According to the definition of LOD = -d(UT1-UTC) / dt, the correlation information between LOD and UT1-UTC is fully explored and utilized, that is, the first-order difference of LOD and UT1-UTC is associated in the MAR model instead of associating LOD with UT1-UTC.

[0064] Example:

[0065] See also Figure 1 , a method for predicting ultra-short-term earth rotation parameters, the method comprising the following steps:

[0066] S1. Obtain the UT1-UTC basic data and LOD basic data provided by IERS through files or the Internet, and obtain the latest LOD basic data derived from the fast and low-latency observations of GNSS technology;

[0067] A. The latest LOD basic data can be obtained by the following methods: downloading the LOD basic data of international GNSS services with a delay of 3 to 9 hours, or downloading the LOD basic data of domestic iGMAS with a delay of 3 to 9 hours, or using the LOD basic data of the latest ultra-fast products derived by GNSS technology;

[0068] B. The latest LOD basic data is obtained by ultra-fast LOD parameter calculation based on the original GNSS observation data:

[0069] a. Real-time acquisition of the latest 24-hour observation value files, navigation files, a priori earth rotation parameter files and general configuration files for precise determination of GNSS earth rotation parameters;

[0070] b. Establish GNSS observation equation based on GNSS observation data:

[0071]

[0072] In the formula, is the theoretical distance, ρ is the observed distance, Δρ is the observation error, X s is the coordinate of the satellite in the Earth-fixed coordinate system, X r is the coordinate of the measuring station in the Earth-fixed coordinate system, P, N, S, and U are the rotation matrices of precession, nutation, rotation, and polar motion from the Earth-fixed coordinate system to the inertial coordinate system, respectively;

[0073] It is necessary to linearize the parameters of GNSS satellite orbits and stations in a regular manner. The daily changing LOD parameters are implicit in the rotation matrix S, which is linearized as follows:

[0074]

[0075] Where LOD0 is the a priori LOD parameter, θ is the Greenwich mean true sidereal time, γ≈1.0027379093, t is the Julian day at the observation time, and t0 is the Julian day at the selected reference time;

[0076] c. After linearization, the GNSS observation equation is used to solve the correction number of the prior LOD parameter LOD0 according to the least squares method, and finally the latest fast and low-latency LOD basic data derived from GNSS technology is obtained;

[0077] S2. Remove the leap seconds and tides of the UT1-UTC basic data to obtain the UT1R-TAI basic data, and remove the tides of the LOD basic data to obtain the LODR basic data. The time series of the UT1R-TAI basic data and the LODR basic data are divided into periodic terms, linear terms and residual terms.

[0078] When the forecast duration is longer than 2 days, the latest LOD basic data is used;

[0079] S3. Calculate the forecast value of UT1R-TAI basic data at time t+1 The specific steps include:

[0080] S31. Calculate the first-order difference sequence of UT1R-TAI basic data:

[0081]

[0082] Where Δf t ut1 is the first-order difference sequence of UT1R-TAI basic data at time t, f t ut1 is the UT1R-TAI basic data at time t, is the UT1R-TAI basic data at time t-1;

[0083] S32. Fit and predict Δf using the least squares (LS) model t ut1The periodic and linear terms of , the periodic and linear terms of LODR, and at the same time, calculate Δf t ut1 The residual term dΔf t ut1 The residual term dLODR of LODR t ;

[0084] The correlation coefficient r is:

[0085]

[0086] In the formula, is Δf t ut1 The mean value of the residual term, is the average value of the residual term of LODR;

[0087] The calculation of the correlation coefficient is to facilitate intuitive testing and verification that this design does indeed mine and utilize the relevant information between LOD and UT1-UTC; Figure 3 is the correlation coefficient of this design statistic, Figure 4 is the correlation coefficient of traditional statistical methods, from Figure 3 , Figure 4 It can be seen that this design does fully explore and utilize the relevant information between LOD and UT1-UTC;

[0088] S33. Fitting Δf using the multivariate autoregression (MAR) model t ut1 The residual term dΔf t ut1 The residual term dLODR of LODR t , and predict Δf t ut1 The residual term of

[0089] S34. Comprehensive forecast Δf t ut1 The periodic term, linear term and residual term of Δf are calculated t ut1 The forecast value at time t+1

[0090] S35. Calculate the forecast value of UT1R-TAI basic data at time t+1 using the differential recovery operator

[0091]

[0092] S4. Increase The leap second and tide of UT1-UTC are used to obtain the forecast value of UT1-UTC basic data at time t+1.

[0093] The above method can also be generalized and improved by WLS+MAR, that is, the WLS model is used in step S32 instead of the LS model.

[0094] See also Figure 5 , this embodiment is the statistical result obtained by performing 210 forecasts, each forecast is 1 to 10 days, where MAE represents the mean absolute error of the forecast UT1-UTC, LS+AR and LS+MAR represent the forecast error results of the traditional method, Bulletin-A is the forecast error result of the Bulletin-A product released by IERS, and the present invention represents the forecast error result of this design. Therefore, the results show that the ultra-short-term prediction result of UT1-UTC obtained by the ultra-short-term earth rotation parameter prediction method of this design taking into account the correlation of earth rotation parameters and the assistance of GNSS derived parameters is significantly improved in forecast accuracy performance compared with the prediction results of the traditional methods of LS+AR and LS+MAR, and is also significantly better than the forecast accuracy of the Bulletin-A product released by IERS. Since there are many weighting methods for WLS, and in theory, LS can be regarded as a special WLS, the implementation example of this design extended to WLS+MAR and the comparison with the traditional WLS+AR / WLS+MAR method will not be repeated here.

Claims

1. A method for predicting ultra-short-term earth rotation parameters, characterized in that: The method comprises the following steps: S1. Obtain the UT1-UTC basic data and LOD basic data provided by IERS through files or the Internet, and obtain the latest LOD basic data derived from the fast and low-latency observations of GNSS technology; S2. Remove the leap seconds and tides of the UT1-UTC basic data to obtain the UT1R-TAI basic data, and remove the tides of the LOD basic data to obtain the LODR basic data. The time series of the UT1R-TAI basic data and the LODR basic data are divided into periodic terms, linear terms and residual terms. S3. Calculate the forecast value of UT1R-TAI basic data at time t+1 The specific steps include: S31. Calculate the first-order difference sequence of UT1R-TAI basic data: Where Δf t ut1 is the first-order difference sequence of UT1R-TAI basic data at time t, f t ut1 is the UT1R-TAI basic data at time t, is the UT1R-TAI basic data at time t-1; S32. Fit and predict Δf using the least squares model t ut1 The periodic and linear terms of , the periodic and linear terms of LODR, and at the same time, calculate Δf t ut1 The residual term dΔf t ut1 and the residual term dLODRt of LODR; S33. Fitting Δf using a multivariate autoregressive model t ut1 The residual term dΔf t ut1 The residual term dLODR of LODR t , and predict Δf t ut1 The residual term of S34. Comprehensive forecast Δf t ut1 The periodic term, linear term and residual term of Δf are calculated t ut1 The forecast value at time t+1 S35. Calculate the forecast value of UT1R-TAI basic data at time t+1 using the differential recovery operator S4. Increase The leap second and tide of UT1-UTC are used to obtain the forecast value of UT1-UTC basic data at time t+1.

2. The method for predicting ultra-short-term earth rotation parameters according to claim 1, characterized in that: In step S1, the latest LOD basic data is obtained by the following method: downloading the LOD basic data of the international GNSS service with a delay of 3 to 9 hours or downloading the LOD basic data of the domestic iGMAS with a delay of 3 to 9 hours or using the LOD basic data of the latest ultra-fast product derived by GNSS technology.

3. The method for predicting ultra-short-term earth rotation parameters according to claim 1, characterized in that: In step S1, the latest LOD basic data is obtained by ultra-fast LOD parameter solution based on the original GNSS observation data: a. Real-time acquisition of the latest 24-hour observation value files, navigation files, a priori earth rotation parameter files and general configuration files for precise determination of GNSS earth rotation parameters; b. Establish GNSS observation equation based on GNSS observation data: In the formula, is the theoretical distance, ρ is the observed distance, Δρ is the observation error, X s is the coordinate of the satellite in the Earth-fixed coordinate system, X r is the coordinate of the measuring station in the Earth-fixed coordinate system, P, N, S, and U are the rotation matrices of precession, nutation, rotation, and polar motion from the Earth-fixed coordinate system to the inertial coordinate system, respectively; The GNSS observation equation is linearized as follows: Where LOD0 is the a priori LOD parameter, θ is the Greenwich mean true sidereal time, γ≈1.0027379093, t is the Julian day at the observation time, and t0 is the Julian day at the selected reference time; c. After linearization, the GNSS observation equation is used to solve the correction number of the prior LOD parameter LOD0 according to the least squares method, and finally the latest fast and low-latency LOD basic data derived from GNSS technology is obtained.

4. The method for predicting ultra-short-term earth rotation parameters according to claim 1, characterized in that: In step S2, when the forecast duration is greater than 2 days, the latest LOD basic data is used.

Images