Earth rotation parameter forecasting method based on least square and long short-term memory network

Abstract

The invention relates to the technical field of geophysics, in particular to an earth rotation parameter forecasting method based on a least square and long short-term memory network, which comprises the following steps: acquiring and preprocessing earth rotation basic sequence data and effective angular momentum data; converting the preprocessed earth rotation basic sequence data into geodetic measurement angular momentum; fitting the effective angular momentum data and the geodetic measurement angular momentum to respectively obtain a fitting sequence and a residual sequence of the effective angular momentum and the geodetic measurement angular momentum; inputting the effective angular momentum and the geodetic measurement angular momentum residual error sequence into the long short-term memory network model to obtain an effective angular momentum error term and a geodetic measurement angular momentum error term; and integrating the effective angular momentum fitting sequence, the fitting sequence, the effective angular momentum and the geodetic measurement angular momentum error term to obtain an earth rotation parameter forecasting result. Therefore, the problem that a traditional earth rotation parameter prediction method is difficult to capture complex nonlinear characteristics in rotation parameter changes is solved.

Description

Technical Field

[0001] The invention relates to the technical field of geophysics, and in particular to a method for predicting earth rotation parameters based on least squares and long short-term memory network. Background Art

[0002] High-precision Earth rotation parameters are the basic parameters for establishing and maintaining the Earth reference frame, and are also the key parameters for realizing the conversion between the International Terrestrial Reference Frame (ITRF) and the International Celestial Reference Frame (ICRF). They play an important role in applications such as deep space exploration of spacecraft, precise orbit determination of artificial satellites, navigation positioning, and high-precision timing. For modern geodetic surveying, aerospace autonomous positioning, and satellite navigation systems, accurate prediction of Earth rotation parameters (ERPs) is essential. The effective operation of these fields depends on the long-term prediction of ERPs. Therefore, improving the accuracy of ERPs prediction is crucial to improving the accuracy of satellite orbit determination and spacecraft positioning.

[0003] Although researchers can obtain a variety of observation data provided by different geodetic techniques, ERPs still cannot be provided in real time due to the complexity of the processing process. Traditional prediction methods usually rely on simple linear regression analysis of historical data, which is difficult to capture the complex nonlinear characteristics of the changes in rotation parameters. Therefore, a new prediction method is urgently needed to capture the complex nonlinear characteristics of the changes in rotation parameters, thereby improving the prediction accuracy of the Earth's rotation parameters. Summary of the invention

[0004] The present invention provides a method for predicting the earth's rotation parameters based on least squares and long short-term memory networks to solve the problems that traditional earth's rotation parameter prediction methods usually rely on simple linear regression analysis of historical data and are difficult to capture the complex nonlinear characteristics of rotation parameter changes.

[0005] The first aspect of the present invention provides a method for predicting earth rotation parameters based on least squares and long short-term memory network, comprising the following steps:

[0006] Collecting earth rotation basic sequence data and effective angular momentum data; preprocessing the earth rotation basic sequence data and the effective angular momentum data to obtain preprocessed earth rotation basic sequence data and preprocessed effective angular momentum data; converting the preprocessed earth rotation basic sequence data into geodetic angular momentum; fitting the effective angular momentum data and the geodetic angular momentum to obtain an effective angular momentum fitting sequence, an effective angular momentum residual sequence, a geodetic angular momentum fitting sequence and a geodetic angular momentum residual sequence; inputting the effective angular momentum residual sequence and the geodetic angular momentum residual sequence into a pre-constructed long short-term memory network model to obtain an effective angular momentum error term and a geodetic angular momentum error term; integrating the effective angular momentum fitting sequence, the geodetic angular momentum fitting sequence, the effective angular momentum error term and the geodetic angular momentum error term to obtain a prediction result of the total angular momentum of the earth, and converting the prediction result of the total angular momentum of the earth into a prediction result of the earth rotation parameter.

[0007] Optionally, the earth rotation basic sequence data includes polar motion and earth rotation rate; the effective angular momentum data includes atmospheric angular momentum, ocean angular momentum, land water angular momentum and sea level angular momentum.

[0008] Optionally, the preprocessing of the earth rotation basic sequence data and the effective angular momentum data to obtain preprocessed earth rotation basic sequence data and preprocessed effective angular momentum data includes:

[0009] The effective angular momentum data are interpolated to the same interval and the same time as the earth rotation basic sequence data to obtain the preprocessed effective angular momentum data; leap second deduction and earth solid tide correction are performed on the earth rotation basic sequence data to obtain the preprocessed earth rotation basic sequence data.

[0010] Optionally, converting the preprocessed earth rotation basic sequence data into geodetic angular momentum comprises:

[0011] The pre-processed earth rotation basic sequence data is converted from the polar shift domain to the excitation domain through the Liouville equation to obtain the geodetic angular momentum.

[0012] Optionally, the fitting of the effective angular momentum data and the geodetic angular momentum to obtain an effective angular momentum fitting sequence, an effective angular momentum residual sequence, a geodetic angular momentum fitting sequence and a geodetic angular momentum residual sequence comprises:

[0013] Performing data fitting on the effective angular momentum data and the geodetic angular momentum through a least squares model to obtain the effective angular momentum fitting sequence and the geodetic angular momentum fitting sequence;

[0014] The effective angular momentum residual sequence is calculated according to the preprocessed effective angular momentum data and the effective angular momentum fitting sequence; and the geodetic angular momentum residual sequence is calculated according to the geodetic angular momentum and the geodetic angular momentum fitting sequence.

[0015] Optionally, converting the Earth's total angular momentum prediction result into an Earth rotation parameter prediction result comprises:

[0016] The prediction result of the total angular momentum of the Earth is converted from the excitation domain to the polar shift domain to obtain the prediction result of the Earth rotation parameter.

[0017] The second aspect of the present invention provides an earth rotation parameter prediction device based on least squares and long short-term memory network, comprising:

[0018] The acquisition module is used to acquire the earth rotation basic sequence data and the effective angular momentum data; the preprocessing module is used to preprocess the earth rotation basic sequence data and the effective angular momentum data to obtain the preprocessed earth rotation basic sequence data and the preprocessed effective angular momentum data; the conversion module is used to convert the preprocessed earth rotation basic sequence data into geodetic angular momentum; the fitting module is used to fit the effective angular momentum data and the geodetic angular momentum to obtain the effective angular momentum fitting sequence, the effective angular momentum residual sequence, the geodetic angular momentum A fitting sequence and a geodetic angular momentum residual sequence; a training module, used to train a pre-constructed long short-term memory network model using the effective angular momentum residual sequence and the geodetic angular momentum residual sequence to obtain an effective angular momentum error term and a geodetic angular momentum error term; a prediction module, used to integrate the effective angular momentum fitting sequence, the geodetic angular momentum fitting sequence, the effective angular momentum error term and the geodetic angular momentum error term to obtain a total angular momentum prediction result of the earth, and convert the total angular momentum prediction result of the earth into a rotation parameter prediction result of the earth.

[0019] A third aspect of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for predicting the earth's rotation parameters based on least squares and long short-term memory network as described in the above embodiment.

[0020] The fourth aspect of the present invention provides a computer program product, which, when executed by a processor, implements the above-mentioned method for predicting the earth's rotation parameters based on least squares and long short-term memory network.

[0021] The fifth aspect of the present invention provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned method for predicting the earth's rotation parameters based on least squares and long short-term memory network.

[0022] The method for predicting the earth rotation parameters based on least squares and long short-term memory network in the embodiment of the present invention, by adopting the LS+LSTM combined model, not only makes full use of the interpretability of the linear model, but also introduces the powerful modeling ability of deep learning for complex time series data, thereby effectively overcoming the defect of poor prediction accuracy of the random part of the earth rotation parameters by the traditional linear model, and at the same time considers the influence of physical excitation factors on ERPs, providing higher prediction accuracy and stability than traditional methods. This method realizes high-precision prediction of earth parameters in the medium and long term, and provides a new research method for the prediction of earth rotation parameters.

[0023] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0025] Figure 1 A flowchart of a method for predicting earth rotation parameters based on least squares and long short-term memory network provided by an embodiment of the present invention;

[0026] Figure 2 A schematic diagram of a basic sequence of earth rotation parameters from EOP 20C04 provided in an embodiment of the present invention;

[0027] Figure 3 Schematic diagrams of AAM, OAM, HAM and SLAM after interpolation provided in an embodiment of the present invention, wherein (a) is AAM, (b) is OAM, (c) is HAM, and (d) is SLAM;

[0028] Figure 4 A schematic diagram of the equatorial components χ1, χ2 and the axial component χ3 after the effective angular momentum is combined provided in an embodiment of the present invention;

[0029] Figure 5 A schematic diagram of the UT1-UTC prediction process provided by an embodiment of the present invention;

[0030] Figure 6 A schematic diagram of error statistics of LSTM model parameter training based on PMX sequence provided in an embodiment of the present invention;

[0031] Figure 7 A schematic block diagram of an earth rotation parameter prediction device based on least squares and long short-term memory network provided by an embodiment of the present invention;

[0032] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0033] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0034] The following describes the method for predicting the earth rotation parameters based on the least squares and long short-term memory network of the embodiment of the present invention with reference to the accompanying drawings. In view of the problems mentioned in the above background technology center, the present application takes into account that the LSTM model can establish implicit but unknown connections for different training windows and different features. At the same time, the LSTM model has a high degree of parameterization and a long time span, and more information can be used to improve the prediction accuracy. Therefore, the LSTM model can be used to capture the nonlinear structure between different periods in the ERPs time series. In addition, since the atmosphere, ocean, terrestrial hydrosphere, etc. affect the rotation of the earth, it is necessary to consider these physical excitation factors when predicting ERPs. Therefore, for the problem of ERPs prediction, a discretized Liouville equation is established, and the geodetic angular momentum is obtained by solving it, and then the geodetic angular momentum and the effective angular momentum are comprehensively applied to the ERPs prediction. Therefore, the present application provides a method for combining the physical excitation factors affecting the rotation of the earth with the LS+LSTM model to predict the earth's rotation parameters, so as to improve the accuracy and stability of the prediction of the earth's rotation parameters.

[0035] Specifically, Figure 1 A flowchart of a method for predicting earth rotation parameters based on least squares and long short-term memory network provided in an embodiment of the present invention.

[0036] like Figure 1 As shown, the method for predicting the earth rotation parameters based on least squares and long short-term memory network includes the following steps:

[0037] In step S101, the earth rotation basic sequence data and effective angular momentum data are collected.

[0038] In the actual implementation process, Figure 2 and 3As shown in the figure, the physical excitation factors related to the earth's rotation parameters are collected as effective angular momentum data, which mainly include atmospheric angular momentum (AAM), oceanic angular momentum (OAM), hydrospheric angular momentum (HAM) and sea-level angular momentum (SLAM). At the same time, the basic sequence of earth rotation parameters with a time interval of 1 day is downloaded from the official website of the International Earth Rotation and Reference Systems Service (IERS), including polar motion (PM) and earth rotation rate (Universal Time-Coordinated Universal Time, UT1-UTC).

[0039] In step S102, the earth rotation basic sequence data and the effective angular momentum data are preprocessed to obtain preprocessed earth rotation basic sequence data and preprocessed effective angular momentum data.

[0040] In some embodiments, preprocessing the earth rotation basic sequence data and the effective angular momentum data to obtain preprocessed earth rotation basic sequence data and preprocessed effective angular momentum data includes:

[0041] Interpolate the effective angular momentum data to the same interval and time as the basic sequence data of the earth's rotation to obtain the preprocessed effective angular momentum data;

[0042] The leap second deduction and the Earth's solid tide correction are performed on the Earth's rotation basic sequence data to obtain the preprocessed Earth's rotation basic sequence data.

[0043] In the actual implementation process, Figure 4 and 5 As shown, the effective angular momentum data related to the polar motion are merged together and interpolated, and the interpolation time interval is consistent with the ERPs sequence provided by IERS, that is, the collected effective angular momentum data are interpolated to the same interval and the same time as the basic sequence of the Earth's rotation parameters, and the preprocessed Earth rotation basic sequence data and the preprocessed effective angular momentum data are obtained.

[0044] It should be noted that before preprocessing the UT1-UTC sequence, it is necessary to deduct the leap second and correct the earth's solid tide. The leap second can be corrected by the value provided by the earth's rotation center. The earth's solid tide correction model is:

[0045]

[0046] Among them, δUT1 is the Earth solid tide correction term, B i ,G i They are given in the 7th and 8th columns of Table 8.1 of the 2010 IERS Convention Standards, α ij is the α of the ith tide in Table 8.1 j Integer multiples, from 1 to 5 columns, α j is the sun and moon nutation parameter.

[0047] In step S103, the preprocessed earth rotation basic sequence data is converted into geodetic angular momentum.

[0048] In some embodiments, converting the preprocessed Earth rotation basic sequence data into geodetic angular momentum includes:

[0049] The preprocessed Earth rotation basic sequence data are converted from the polar shift domain to the excitation domain through the Liouville equation to obtain the geodetic angular momentum.

[0050] In the actual implementation process, the Earth's rotation parameters are transferred to the excitation domain to obtain the geodetic angular momentum. For the polar motion, it is converted from the polar motion domain to the excitation domain through the Liouville equation, and the formula is as follows:

[0051]

[0052] Among them, χ(t) represents the excitation function, σ CW represents the Chandler oscillation frequency, where σ CW =2π(1+i / 2Q) / T ch , Q and T ch denote the quality factor and the Chandler oscillation period respectively. Δt is the sampling interval of the polar shift sequence. The polar shift is converted into an excitation function using the PM at times t-Δt, t, and t+Δt.

[0053] In step S104, the effective angular momentum data and the geodetic angular momentum are fitted to obtain an effective angular momentum fitting sequence, an effective angular momentum residual sequence, a geodetic angular momentum fitting sequence and a geodetic angular momentum residual sequence.

[0054] In some embodiments, fitting the effective angular momentum data and the geodetic angular momentum to obtain an effective angular momentum fitting sequence, an effective angular momentum residual sequence, a geodetic angular momentum fitting sequence, and a geodetic angular momentum residual sequence includes:

[0055] The effective angular momentum data and the geodetic angular momentum are fitted by the least squares model to obtain the effective angular momentum fitting sequence and the geodetic angular momentum fitting sequence;

[0056] Calculate the effective angular momentum residual sequence according to the preprocessed effective angular momentum data and the effective angular momentum fitting sequence;

[0057] A geodetic angular momentum residual sequence is calculated based on the geodetic angular momentum and the geodetic angular momentum fitting sequence.

[0058] In the actual implementation process, the effective angular momentum data and the geodetic angular momentum are used to construct a linear LS model through least squares to obtain the effective angular momentum fitting sequence and the geodetic angular momentum fitting sequence, wherein the linear LS model is constructed according to the following formula:

[0059]

[0060] Among them, a k represents a constant, a k+1 represents the linear term, k is [0,7], i is the segment number of LS fitting, C i and D i is the coefficient corresponding to the periodic term, R i represents the periodic term. The periodic terms considered in the fitting process include annual swing, semi-annual swing, 1 / 3 annual swing and 13.7-day swing. The values ​​are usually 1 year, 0.5 year, 1 / 3 year and 13.7 days. Please note that it is calculated in years. i represents zero-mean white noise, and t represents UTC time.

[0061] Furthermore, the preprocessed effective angular momentum is subtracted from the effective angular momentum fitting sequence to obtain the effective angular momentum residual, and the geodetic angular momentum residual sequence is calculated according to the geodetic angular momentum and the geodetic angular momentum fitting sequence. Finally, the effective angular momentum residual sequence and the geodetic angular momentum residual sequence are used as part of the input of the pre-constructed long short-term memory network LSTM model.

[0062] In step S105, the effective angular momentum residual sequence and the geodetic angular momentum residual sequence are input into a pre-constructed long short-term memory network model to obtain the effective angular momentum error term and the geodetic angular momentum error term.

[0063] In the actual implementation process, in the prediction of the earth's rotation parameters, before model training, the earth's rotation parameters are first converted into the excitation domain according to step S103 to obtain the geodetic angular momentum, and the geodetic angular momentum fitting sequence and fitting residual sequence are obtained by the least squares fitting method. Then, the effective angular momentum residual sequence and the geodetic angular momentum residual sequence are respectively used as the input values ​​of the LSTM model for model parameter training. First, the output result c(t) of the LSTM model cell is calculated, and the error term of the LSTM cell is calculated. According to the specific error term, the weight gradient is solved, and the gradient optimization algorithm is used to re-weight.

[0064] In step S106, the effective angular momentum fitting sequence, the geodetic angular momentum fitting sequence, the effective angular momentum error term and the geodetic angular momentum error term are integrated to obtain the Earth's total angular momentum prediction result, and the Earth's total angular momentum prediction result is converted into the Earth's rotation parameter prediction result.

[0065] In some embodiments, converting the Earth's total angular momentum prediction result into the Earth's rotation parameter prediction result includes:

[0066] The prediction results of the Earth's total angular momentum are converted from the excitation domain to the polar shift domain to obtain the prediction results of the Earth's rotation parameters.

[0067] In the actual implementation process, the effective angular momentum fitting sequence, the geodetic angular momentum fitting sequence, the effective angular momentum error sequence predicted by the LSTM model, and the geodetic angular momentum error sequence are added together to obtain the prediction sequence by the LS+LSTM model, that is, the total angular momentum prediction result of the Earth.

[0068] Furthermore, the prediction results of the total angular momentum of the Earth in the excitation domain are replaced by the prediction results of the Earth's rotation parameters through the Liouville equation. It should be noted that for the UT1-UTC prediction sequence, the prediction values ​​of the UT1-TAI sequence are obtained through the Liouville equation. On this basis, the influence of leap seconds and the Earth's solid tides should be added to obtain the final UT1-UTC prediction results.

[0069] The angular momentum can be converted from the excitation domain to the polar shift domain by the following formula:

[0070]

[0071] Among them, P(t) represents the polar shift, and the meanings of other parameters are the same as in the above formula.

[0072] For UT1-UTC,

[0073]

[0074] Among them, UT1 represents universal time, TAI represents atomic time, UT1-TAI is the result obtained by preprocessing UT1-UTC, Ψ is the geodetic angular momentum function, and Ω is the average angular velocity of the earth, which is 7.292115×10 -5 rands -1 .

[0075] It should be noted that the embodiment of the present invention also uses the mean absolute error MAE to perform accuracy statistics on the prediction results of the earth rotation parameters, and compares them with the accuracy of the forecast sequence Bulletin A provided by the International Earth Rotation and Reference System Service. The accuracy evaluation standard of the earth rotation parameters is through the following formula:

[0076]

[0077] Among them, MAE j is the accuracy of the jth time series, n is the length of the time series, i is the length of the predicted time series, P i is the i-th forecast value, X i is the i-th true value.

[0078] Among them, when conducting MAE progress evaluation, the true value selected is the IERS EOP 20C04 sequence, and the prediction accuracy of the Earth's rotation parameters is compared with the prediction accuracy of Bulletin A, which is generally recognized as the ERPs sequence with the highest prediction accuracy internationally, to further verify the effectiveness of the linear model and deep learning combined model used in this experiment.

[0079] The working process of the earth rotation parameter prediction method based on least squares and long short-term memory network proposed in the embodiment of the present invention is further explained below.

[0080] Step 1: Collect the basic sequence data of the Earth's rotation parameters and effective angular momentum data, such as Figure 2 and Figure 3 shown.

[0081] Step 2: Before building the model, the effective angular momentum data and UT1-UTC data are first preprocessed.

[0082] When processing effective angular momentum data, the equatorial and axial components of AAM, OAM, HAM and SLAM are first interpolated, and the interpolated time is consistent with the IERS EOP 20C04 sequence. Then, the equatorial and axial components of AAM, OAM, HAM and SLAM after interpolation are merged together, as shown in Figure 4When predicting the UT1-UTC sequence, first deduct the leap second and the part affected by the earth's solid tide. The UT1-UTC preprocessed sequence is the UT1-TAI sequence, as shown in Figure 5 shown.

[0083] Step three, transfer the Earth's rotation parameters to the excitation domain according to the Liouville equation. In this process, it should be noted that the discrete Liouville equation is used to convert the Earth's rotation parameters into geodetic angular momentum.

[0084] Step 4: construct LS models according to the effective angular momentum and geodetic angular momentum sequences respectively, and then obtain the fitting sequence and residual sequence of the effective angular momentum and geodetic angular momentum.

[0085] Step 5: The residual sequence of effective angular momentum and geodetic angular momentum obtained in step 4 is used as the input sequence for constructing the LSTM model. The two residual sequences are used as basic sequences for model parameter training to obtain the optimal training parameters. Taking the X component of polar motion as an example, the error statistics after LSTM model parameter training are as follows: Figure 6 shown.

[0086] Step 6: The extrapolated results of the effective angular momentum and geodetic angular momentum fitting sequence obtained by extrapolation of the LS model are added to the results predicted by the LSTM model for the effective angular momentum and geodetic angular momentum residual sequence to obtain the total angular momentum prediction result. At this time, it is necessary to convert it into the earth rotation parameter through the Liouville equation to obtain the prediction result of the earth rotation parameter. For the UT1-UTC sequence, the UT1R-TAI sequence is predicted, and the influence of the leap second and the earth's solid tide needs to be added to obtain the final prediction result.

[0087] In step seven, the prediction results of the earth's rotation parameters of LS+LSTM taking into account the physical excitation factors are statistically analyzed by MAE, and the prediction accuracy is compared with the prediction accuracy of Bulletin A, which further verifies the effectiveness and reliability of the accuracy improvement of the earth's rotation parameter prediction method proposed in the present invention.

[0088] The method for predicting earth rotation parameters based on least squares and long short-term memory network proposed in the embodiment of the present invention has the following beneficial effects:

[0089] (1) By adopting the LS+LSTM combined model, the advantages of both can be fully utilized, which not only fully utilizes the interpretability of the linear model, but also introduces the powerful modeling ability of deep learning for complex time series data;

[0090] (2) The combined model can better capture the seasonal changes and long-term trends in the Earth's rotation parameters, improving the accuracy of the prediction;

[0091] (3) The method can be widely used in the prediction of different types of Earth rotation parameters and has high versatility and flexibility;

[0092] (4) By utilizing the online learning capability of LSTM, rapid response and real-time update of new observation data can be achieved, providing timely support for relevant decision-making;

[0093] (5) The prediction results of the Earth's rotation parameters can provide a scientific basis for fields such as weather forecasting, navigation systems, and earthquake monitoring, and promote intelligent decision-making and applications in various fields.

[0094] Next, the device for predicting earth rotation parameters based on least squares and long short-term memory network proposed in accordance with an embodiment of the present invention will be described with reference to the accompanying drawings.

[0095] Figure 7 It is a block diagram of an earth rotation parameter prediction device based on least squares and long short-term memory network according to an embodiment of the present invention.

[0096] like Figure 7 As shown, the earth rotation parameter prediction device 70 based on least squares and long short-term memory network includes: an acquisition module 701, a preprocessing module 702, a conversion module 703, a fitting module 704, a training module 705 and a prediction module 706.

[0097] Among them, the acquisition module 701 is used to acquire the earth rotation basic sequence data and the effective angular momentum data. The preprocessing module 702 is used to preprocess the earth rotation basic sequence data and the effective angular momentum data to obtain the preprocessed earth rotation basic sequence data and the preprocessed effective angular momentum data. The conversion module 703 is used to convert the preprocessed earth rotation basic sequence data into geodetic angular momentum. The fitting module 704 is used to fit the effective angular momentum data and the geodetic angular momentum to obtain the effective angular momentum fitting sequence, the effective angular momentum residual sequence, the geodetic angular momentum fitting sequence and the geodetic angular momentum residual sequence. The training module 705 is used to train the pre-constructed long short-term memory network model using the effective angular momentum residual sequence and the geodetic angular momentum residual sequence to obtain the effective angular momentum error term and the geodetic angular momentum error term. The prediction module 706 is used to integrate the effective angular momentum fitting sequence, the geodetic angular momentum fitting sequence, the effective angular momentum error term and the geodetic angular momentum error term to obtain the total angular momentum prediction result of the earth, and convert the total angular momentum prediction result of the earth into the prediction result of the earth rotation parameters.

[0098] It should be noted that the aforementioned explanation of the embodiment of the method for predicting the earth's rotation parameters based on least squares and long short-term memory network is also applicable to the earth's rotation parameter prediction device based on least squares and long short-term memory network of this embodiment, and will not be repeated here.

[0099] The device for predicting earth rotation parameters based on least squares and long short-term memory network proposed in the embodiment of the present invention has the following beneficial effects:

[0100] (1) By adopting the LS+LSTM combined model, the advantages of both can be fully utilized, which not only fully utilizes the interpretability of the linear model, but also introduces the powerful modeling ability of deep learning for complex time series data;

[0101] (2) The combined model can better capture the seasonal changes and long-term trends in the Earth's rotation parameters, improving the accuracy of the prediction;

[0102] (3) The method can be widely used in the prediction of different types of Earth rotation parameters and has high versatility and flexibility;

[0103] (4) By utilizing the online learning capability of LSTM, rapid response and real-time update of new observation data can be achieved, providing timely support for relevant decision-making;

[0104] (5) The prediction results of the Earth's rotation parameters can provide a scientific basis for fields such as weather forecasting, navigation systems, and earthquake monitoring, and promote intelligent decision-making and applications in various fields.

[0105] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device may include:

[0106] A memory 801 , a processor 802 , and a computer program stored in the memory 801 and executable on the processor 802 .

[0107] When the processor 802 executes the program, the method for predicting the earth rotation parameters based on least squares and long short-term memory network provided in the above embodiment is implemented.

[0108] Furthermore, the electronic device further comprises:

[0109] The communication interface 803 is used for communication between the memory 801 and the processor 802 .

[0110] The memory 801 is used to store computer programs that can be executed on the processor 802 .

[0111] The memory 801 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0112] If the memory 801, the processor 802 and the communication interface 803 are implemented independently, the communication interface 803, the memory 801 and the processor 802 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0113] Optionally, in a specific implementation, if the memory 801, the processor 802 and the communication interface 803 are integrated on a chip, the memory 801, the processor 802 and the communication interface 803 can communicate with each other through an internal interface.

[0114] The processor 802 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0115] An embodiment of the present invention also provides a computer program product, which, when executed by a processor, implements the above-mentioned method for predicting the earth rotation parameters based on least squares and long short-term memory network.

[0116] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for predicting the earth rotation parameters based on least squares and long short-term memory network.

[0117] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. 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 may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0118] In addition, the terms "first" and "second" 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, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0119] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0120] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0121] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0122] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0123] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0124] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A method for predicting earth rotation parameters based on least squares and long short-term memory network, characterized in that: include: Collect basic sequence data of earth rotation and effective angular momentum data; Preprocessing the earth rotation basic sequence data and the effective angular momentum data to obtain preprocessed earth rotation basic sequence data and preprocessed effective angular momentum data; Converting the preprocessed earth rotation basic sequence data into geodetic angular momentum; Fitting the effective angular momentum data and the geodetic angular momentum to obtain an effective angular momentum fitting sequence, an effective angular momentum residual sequence, a geodetic angular momentum fitting sequence and a geodetic angular momentum residual sequence; Inputting the effective angular momentum residual sequence and the geodetic angular momentum residual sequence into a pre-constructed long short-term memory network model to obtain an effective angular momentum error term and a geodetic angular momentum error term; The effective angular momentum fitting sequence, the geodetic angular momentum fitting sequence, the effective angular momentum error term and the geodetic angular momentum error term are integrated to obtain a prediction result of the total angular momentum of the Earth, and the prediction result of the total angular momentum of the Earth is converted into a prediction result of the Earth's rotation parameters.

2. The method for predicting earth rotation parameters based on least squares and long short-term memory network according to claim 1, characterized in that: The basic sequence data of the earth's rotation include polar motion and the earth's rotation rate; the effective angular momentum data include atmospheric angular momentum, ocean angular momentum, land water angular momentum and sea level angular momentum.

3. The method for predicting earth rotation parameters based on least squares and long short-term memory network according to claim 1, characterized in that: The preprocessing of the earth rotation basic sequence data and the effective angular momentum data to obtain the preprocessed earth rotation basic sequence data and the preprocessed effective angular momentum data includes: Interpolating the effective angular momentum data to the same interval and time as the earth rotation basic sequence data to obtain the pre-processed effective angular momentum data; The earth rotation basic sequence data is subjected to leap second subtraction and earth solid tide correction to obtain the preprocessed earth rotation basic sequence data.

4. The method for predicting earth rotation parameters based on least squares and long short-term memory network according to claim 1, characterized in that: The step of converting the pre-processed earth rotation basic sequence data into geodetic angular momentum comprises: The pre-processed earth rotation basic sequence data is converted from the polar shift domain to the excitation domain through the Liouville equation to obtain the geodetic angular momentum.

5. The method for predicting earth rotation parameters based on least squares and long short-term memory network according to claim 1, characterized in that: The step of fitting the effective angular momentum data and the geodetic angular momentum to obtain an effective angular momentum fitting sequence, an effective angular momentum residual sequence, a geodetic angular momentum fitting sequence and a geodetic angular momentum residual sequence comprises: Performing data fitting on the effective angular momentum data and the geodetic angular momentum through a least squares model to obtain the effective angular momentum fitting sequence and the geodetic angular momentum fitting sequence; Calculating the effective angular momentum residual sequence according to the preprocessed effective angular momentum data and the effective angular momentum fitting sequence; The geodetic angular momentum residual sequence is calculated according to the geodetic angular momentum and the geodetic angular momentum fitting sequence.

6. The method for predicting earth rotation parameters based on least squares and long short-term memory network according to claim 1, characterized in that: The converting the Earth's total angular momentum prediction result into the Earth's rotation parameter prediction result comprises: The prediction result of the total angular momentum of the Earth is converted from the excitation domain to the polar shift domain to obtain the prediction result of the Earth rotation parameter.

7. An earth rotation parameter prediction device based on least squares and long short-term memory network, characterized in that: include: Acquisition module, used to collect basic sequence data of earth rotation and effective angular momentum data; A preprocessing module, used for preprocessing the earth rotation basic sequence data and the effective angular momentum data to obtain preprocessed earth rotation basic sequence data and preprocessed effective angular momentum data; A conversion module, used for converting the preprocessed earth rotation basic sequence data into geodetic angular momentum; A fitting module, used for fitting the effective angular momentum data and the geodetic angular momentum to obtain an effective angular momentum fitting sequence, an effective angular momentum residual sequence, a geodetic angular momentum fitting sequence and a geodetic angular momentum residual sequence; A training module, used for training a pre-built long short-term memory network model using the effective angular momentum residual sequence and the geodetic angular momentum residual sequence to obtain an effective angular momentum error term and a geodetic angular momentum error term; A prediction module is used to integrate the effective angular momentum fitting sequence, the geodetic angular momentum fitting sequence, the effective angular momentum error term and the geodetic angular momentum error term to obtain a prediction result of the total angular momentum of the earth, and convert the prediction result of the total angular momentum of the earth into a prediction result of the earth rotation parameters.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for predicting the earth rotation parameters based on least squares and long short-term memory network as described in any one of claims 1 to 6.

9. A computer program product, characterized in that When the computer program / instruction is executed by a processor, the method for predicting the earth rotation parameters based on least squares and long short-term memory network described in any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for predicting earth rotation parameters based on least squares and long short-term memory network as described in any one of claims 1 to 6.

Images