1. An artificial intelligence prediction method for geocentric motion is characterized by comprising the following steps:

decomposing and reconstructing the time sequence of the original geocentric movement by using an EMD algorithm to obtain a new time sequence;

solving the period and the corresponding amplitude of the new time series; and

based on the period and the corresponding amplitude, the periodic term of the geocentric motion is predicted using a SSA and ARMA combined model.

2. The artificial intelligence prediction method for geocentric motion according to claim 1, wherein the decomposing and reconstructing the time series of the original geocentric motion by using the EMD algorithm to obtain a new time series comprises:

decomposing the time sequence of the original geocentric motion according to the sequence of the frequency from high to low to obtain a plurality of IMF components;

determining a noise term among the plurality of IMF components, and rejecting the noise term; and

and reconstructing the IMF components with the noise items removed to obtain the new time sequence.

3. The method of artificial intelligence for predicting geocentric motion of claim 2, wherein the determining a noise term among the plurality of IMF components comprises:

calculating a correlation coefficient between each IMF component of the plurality of IMF components and the time series of the original geocentric motion;

determining an IMF component of which the relation number meets a preset condition as a boundary IMF component of the noise term and the useful signal; and

determining the demarcation IMF component and a previous IMF as the noise term.

4. The artificial intelligence prediction method of geocentric motion of claim 3, wherein the determining the IMF component of which the correlation coefficient satisfies a predetermined condition as the boundary IMF component of the noise term and the useful signal comprises:

searching a local minimum value in a series of correlation coefficients according to a correlation coefficient local minimum value principle; and

and determining the IMF component corresponding to the local minimum value as a boundary IMF component of the noise term and the useful signal.

5. The method of artificial intelligence prediction of geocentric motion of claim 1, wherein solving for the period and corresponding amplitude of the new time series comprises:

and performing fast Fourier transform on the new time sequence, calculating power spectral density, identifying periods, calculating the contribution rate of each period, and solving the periods and the amplitude corresponding to each period through least square fitting.

6. The artificial intelligence earth-motion prediction method according to claim 1, wherein the predicting the periodic term of the earth-motion by using the SSA and ARMA combined model comprises:

carrying out extrapolation prediction on the long-term and the periodic term of the geocentric movement by utilizing SSA, and predicting a residual term by utilizing ARMA; and

and adding the result values of the prediction to be used as the prediction value of the final time series.

7. An artificial intelligence prediction system for earth movement, comprising:

the decomposition and reconstruction module is used for decomposing and reconstructing the time sequence of the original geocentric movement by utilizing an EMD algorithm to obtain a new time sequence;

the information solving module is used for solving the period and the corresponding amplitude of the new time sequence; and

and the prediction module is used for predicting the periodic term of the geocentric motion by utilizing an SSA and ARMA combined model based on the period and the corresponding amplitude.

8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the artificial intelligence prediction method of earth movement according to any one of claims 1-6 when executing the computer program.

9. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor, performs the steps of the artificial intelligence prediction method of geocentric motion as recited in any one of claims 1-6.

Background

The earth Reference System is an important foundation for research in earth science, and the International earth Reference System (ITRS) is the one with the highest precision and the most widely used in the world. The International earth Reference Frame (ITRF) is a specific implementation of ITRS, which has released 13 versions since its establishment in 1988, the most recent version being ITRF 2014. The ITRF2014 is realized by reprocessing technical solutions by four space geodetic means of GNSS, SLR, VLBI and DORIS, and meanwhile, the number of stations is increased and the processing of nonlinear motion is improved.

According to the technical specifications published by the International Earth Rotation and Reference Systems Service (IERS), the origin of ITRS is the Center of Mass of the Earth System (CM) including the Mass of the solid Earth, the ocean and the atmosphere, while the origin of ITRF appears as CM only on a long time scale and on a short time scale, the origin of ITRF is approximately at a point that is a fixed offset from the Center of the Earth's globe. The redistribution of mass in the earth system causes the center of shape of the solid earth to change constantly with respect to the center of mass, known as geocentric motion, which is directly related to the origin of the ITRF and has become a major source of error in the earth's frame of reference for millimeter-scale accuracy.

The earth mass center movement influences the accurate realization of the origin of the reference frame, so that the construction of the millimeter-scale reference frame with real-time and high precision is also influenced to a certain extent. However, the geocentric motion provided by the current CSR and other mechanisms is a lunar solution, the intervals are sparse, the motion of the earth centroid cannot be accurately estimated in real time, and researchers have made relevant research on the dynamic prediction of the geocentric motion to obtain certain results. However, due to the weak regularity of the geocentric motion, the adopted mathematical model also has certain defects, which results in low prediction accuracy of the geocentric motion.

Disclosure of Invention

In order to solve the problems in the prior art, the invention provides an artificial intelligence prediction method and system for geocentric motion, electronic equipment and a storage medium.

In a first aspect, the present invention provides an artificial intelligence prediction method for geocentric motion, including:

decomposing and reconstructing the time sequence of the original geocentric movement by using an EMD algorithm to obtain a new time sequence;

solving the period and the corresponding amplitude of the new time series; and

based on the period and the corresponding amplitude, the periodic term of the geocentric motion is predicted using a SSA and ARMA combined model.

Further, the decomposing and reconstructing the time series of the original geocentric motion by using the EMD algorithm to obtain a new time series includes:

decomposing the time sequence of the original geocentric motion according to the sequence of the frequency from high to low to obtain a plurality of IMF components;

determining a noise term among the plurality of IMF components, and rejecting the noise term; and

and reconstructing the IMF components with the noise items removed to obtain the new time sequence.

Further, the determining a noise term among the plurality of IMF components comprises:

calculating a correlation coefficient between each IMF component of the plurality of IMF components and the time series of the original geocentric motion;

determining an IMF component of which the relation number meets a preset condition as a boundary IMF component of the noise term and the useful signal; and

determining the demarcation IMF component and a previous IMF as the noise term.

Further, the determining an IMF component of which the correlation coefficient satisfies a predetermined condition as a boundary IMF component of the noise term and the useful signal includes:

searching a local minimum value in a series of correlation coefficients according to a correlation coefficient local minimum value principle; and

and determining the IMF component corresponding to the local minimum value as a boundary IMF component of the noise term and the useful signal.

Further, said solving for a period and corresponding amplitude of said new time series comprises:

and performing fast Fourier transform on the new time sequence, calculating power spectral density, identifying periods, calculating the contribution rate of each period, and solving the periods and the amplitude corresponding to each period through least square fitting.

Further, the predicting the periodic term of the geocentric motion by using the SSA and ARMA combined model comprises the following steps:

carrying out extrapolation prediction on the long-term and the periodic term of the geocentric movement by utilizing SSA, and predicting a residual term by utilizing ARMA; and

and adding the result values of the prediction to be used as the prediction value of the final time series.

In a second aspect, the present invention provides an artificial intelligence prediction system for a geocentric motion, comprising:

the decomposition and reconstruction module is used for decomposing and reconstructing the time sequence of the original geocentric movement by utilizing an EMD algorithm to obtain a new time sequence;

the information solving module is used for solving the period and the corresponding amplitude of the new time sequence; and

and the prediction module is used for predicting the periodic term of the geocentric motion by utilizing an SSA and ARMA combined model based on the period and the corresponding amplitude.

In a third aspect, the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the artificial intelligence prediction method for geocarding motion according to any one of the first aspect.

In a fourth aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the steps of the artificial intelligence prediction method for geocentric motion according to any one of the first aspect.

According to the method, the EMD algorithm is adopted to decompose and reconstruct the time sequence of the geocentric motion, so that the periodic term of the geocentric motion can be effectively identified; in addition, the prediction precision of the geocentric movement of 1mm is realized by constructing an artificial intelligent geocentric movement prediction model, namely performing long-term, medium-term and short-term prediction on periodic items of the geocentric movement by using an SSA and ARMA combined model.

Drawings

FIG. 1 is a flowchart of an artificial intelligence prediction method for geocentric motion according to an embodiment of the present invention;

fig. 2(a) to 2(c) are exploded views of the geocentric motion time sequence EMD provided in the embodiment of the present invention in different directions;

FIG. 3 is a diagram of correlation coefficients of IMF components and an original sequence according to an embodiment of the present invention;

FIG. 4 is a diagram of power spectral densities in various directions for EMD reconstruction timing provided by an embodiment of the present invention;

FIG. 5 is a diagram of a SSA and ARMA combined model predictive technology path provided by an embodiment of the present invention;

FIG. 6 is a schematic structural diagram of an artificial intelligence prediction system for geocentric motion according to an embodiment of the present invention; and

fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

According to the present invention, geocentric motion prediction involves geocentric motion time series analysis and geocentric motion artificial intelligence prediction, the process of which will be described below.

1. Analysis of earth's heart movement time series

The time series is a time-varying sequence, and the task of time series analysis is to extract representative information of the time series, such as period, amplitude, and long-term variation of the time series, so as to provide support for prediction of geocentric motion. However, different time series have different properties, wherein, mixed noise is different, in order to extract information more effectively, a high-precision empirical mode decomposition analysis method is adopted, and specific contents of information such as a periodic item and an amplitude item of geocentric motion are identified as follows:

(1) the essence of Empirical Mode Decomposition (EMD) is to smooth a non-stationary signal, decompose the signal step by step according to fluctuations of different scales, and generate a series of sequences representing different features, i.e., an Intrinsic Mode Function (IMF) component and a trend term in which the feature scales increase from small to large. Each IMF component is a nonlinear stationary sequence with localized features in time and must satisfy two conditions: the number of extreme points of the whole sequence is equal to the number of zero-crossing points or at most one difference; the average value of the envelope composed of the local maximum value or the local minimum value is 0.

(2) According to a detailed decomposition process, the EMD algorithm does not need any prior value, and is sequentially decomposed from high frequency to low frequency according to the data, IMF components containing high-frequency components are firstly separated, and the whole filtering process is completed in a self-adaptive mode.

2. Artificial intelligence forecast for geocentric movement

The geocentric motion time sequence is composed of various periodic terms, trend terms and noise terms, is complex in property and has the characteristics of nonlinearity and non-stability. The ARMA method is more suitable for short-term prediction of stationary sequences, while the SSA method is better able to process and predict signals with periodic oscillations. Therefore, the combination of the SSA method and the ARMA method and the prediction can make up for the defects of the single method and improve the prediction precision.

The main idea of the SSA + ARMA prediction is to extrapolate and predict the long term and the period term by SSA, predict the residual term by ARMA, and add the predicted values of the two as the predicted value of the final time series.

Embodiments of the present invention will be described in detail below by way of examples.

Fig. 1 shows a flowchart of an artificial intelligence prediction method for geocentric motion according to an embodiment of the present invention. Referring to fig. 1, the prediction method includes:

step S101: decomposing and reconstructing the time sequence of the original geocentric movement by using an EMD algorithm to obtain a new time sequence;

step S103: solving the period and the corresponding amplitude of the new time sequence; and

step S105: based on the period and the corresponding amplitude, the periodic term of the geocentric motion is predicted by using the SSA and ARMA combined model.

Specifically, step S101 (decomposing and reconstructing the time series of the original geocentric motion by using the EMD algorithm to obtain a new time series) is as follows:

firstly, an EMD algorithm is utilized to decompose an original geocentric motion time sequence to obtain a plurality of IMF components and trend terms.

Fig. 2(a) to fig. 2(c) are exploded views of the geocentric motion time sequence EMD in different directions (taking the geocentric motion of 2007 and 2017 as an example), which show the original time sequence (Signal) in three directions and each decomposed IMF component and trend term (res), wherein fig. 2(a) represents an exploded view in the X direction, fig. 2(b) represents an exploded view in the Y direction, and fig. 2(c) represents an exploded view in the Z direction.

Then, noise terms in the plurality of IMF components are determined and culled. Since the EMD decomposition process is performed from high to low in frequency, the noise signal usually exhibits high frequency characteristics, and therefore, it is necessary to determine the eigenmode component of the boundary of the noise and the useful signal, and then the eigenmode component of the boundary and the IMF component before the boundary can be determined as the noise term (high frequency term). The noise term may be determined by computing the correlation of each IMF component with the original sequence to determine the dominant strength of the noise to the original sequence to determine the boundary eigenmode components.

The steps for determining the noise term are as follows:

(1) calculating the correlation coefficient of all IMF components with the original sequence (as shown in Table 1 and FIG. 3);

(2) determining the IMF component meeting the preset condition as a boundary eigenmode component; and

(3) the boundary eigenmode component and its preceding IMF component are determined as noise terms.

In the present invention, the calculation formula of the correlation coefficient is as follows:

wherein r is a correlation coefficient, σ_ijIs the covariance, σ, of the IMF component and the original sequence_iIs the standard deviation, σ, of the IMF component_jIs the standard deviation of the original sequence.

In the present invention, the predetermined condition may mean that the correlation coefficient is a local minimum value, and accordingly, determining the IMF component satisfying the predetermined condition as the boundary eigenmode component may mean the following case: searching a local minimum value in a series of correlation coefficients according to a correlation coefficient local minimum value principle; and determining the IMF component corresponding to the local minimum value as a boundary IMF component of the noise term and the useful signal.

Table 1 below shows the correlation coefficient of all IMF components with the original sequence in each direction, and fig. 3 more clearly shows the variation of the correlation coefficient of all IMF components with the original sequence in each direction.

TABLE 1 statistical table of correlation coefficients of IMF components in three directions

As can be seen from table 1 and fig. 3, the boundary layers of the high-frequency term and the low-frequency term in the three directions are all concentrated in the 2 nd IMF component; the 4 th IMF component with better consistency with the original sequence in the X direction has a correlation coefficient of nearly 0.9; the components of 4 th and 7 th IMF with better consistency with the original sequence in the Y and Z directions have the correlation coefficient of more than 0.6; and the correlation coefficients of the three directions corresponding to the same IMF component are also different, which also indicates that the method is to decompose according to the self characteristics of the original sequence. It can be seen from this that, in the EMD algorithm, the boundary value of the high-frequency term is determined without setting a priori values such as basis functions in the decomposition and reconstruction process, and the determination is performed adaptively according to the characteristics of the time series.

Here, the boundary IMF component is determined with the local minimum value as the predetermined condition (for example, the 2 nd IMF component is determined as the boundary IFM component, and thus the 1 st IMF component and the 2 nd IMF component are determined as the noise term), but it should be noted that in the present invention, the predetermined condition is not limited to the above condition, and may refer to other conditions. Alternatively, the predetermined condition may be that the magnitude of the correlation coefficient is smaller than a set threshold value or the like.

And finally, reconstructing the residual IMF components after the noise items are removed to obtain a relatively clean geocentric motion time sequence.

Specifically, step S103 (solving for the period of the new time series and the corresponding amplitude) is as follows:

and performing fast Fourier transform on the reconstructed relatively clean geocentric motion time sequence in 2007-2017 years, calculating power spectral density, drawing a spectrogram, performing power spectral analysis, identifying periods, calculating the contribution rate of each period, and finally solving the long-term variation and the amplitude corresponding to each period through least square fitting.

The power spectral density is the ratio of the spectral density corresponding to a period to the sum of the spectral densities of the periods. Fig. 4 shows X, Y and a power spectral density map in the Z direction (sampling frequency 7 d). Table 2 shows the statistics of the contribution rates of the respective periods before and after the reconstruction of the time series of changes in geocentric motion using the EMD method.

TABLE 2 periodic contribution statistics before and after reconstructing a time series using EMD

It can be seen from figure 4 that there is a higher energy present at 0.019Hz (corresponding to a one year period) in all three directions, indicating that there is a significant annual change in all three directions. However, the Y and Z directions have larger energy in the low frequency part, which shows that the two directions have more obvious long-term change terms, and the Z direction is more obvious. It can be seen from table 2 that after the time series is reconstructed by using the EMD method, the contribution rates of the main periods are all improved, and through calculation, the contribution rates of the periods in the X, Y and Z directions are improved by 12.3%, 16.7% and 6.3% on average, which indicates that the EMD method can suppress high-frequency information, improve the period contribution rate, and make the identification of the periodic characteristics more obvious and accurate. In addition, it can be seen that the trend of the long period variation in the Z direction is more significant than in the X and Y directions, and the contribution rate of the trend term is the highest among the contribution rates of the periods. And on the basis of the identification period, fitting the amplitude and the long-term change value corresponding to each period of the geocentric motion time sequence by using least square. Table 3 shows the amplitude and phase values.

TABLE 3 amplitude and phase of the cyclic variation of the Earth's cardiac motion

As can be seen from table 3, the annual and semiannual terms have amplitudes in mm, and the annual term has an amplitude greater than the semiannual amplitude, which is substantially consistent with the previous results. X, Y and Z-direction annual amplitude are maximum values of amplitude corresponding to each period, and are respectively 2.32mm, 1.89mm and 2.07 mm. However, the cycle contribution rate in the Z direction is 4004 days, which is close to the study time span of the paper, and shows that the trend term in the Z direction is more obvious, which is consistent with the phenomenon of high spectral density at low frequency in the power spectral density map. However, the annual change contribution rate in the Z direction is different from the contribution rate in the long-term tendency by only about 1.5%, and it is considered that there is a relatively significant annual change also in the Z direction. Also, as shown in tables 2-3, the half-year term was weak in all three directions, with 4-6 months of concussion in the X and Y directions.

Specifically, step S105 (predicting the periodic term of the geocentric motion using the SSA and ARMA combined model based on the period and corresponding amplitude) is as follows:

the geocentric motion time sequence is composed of various periodic terms, trend terms and noise terms, is complex in property and has the characteristics of nonlinearity and non-stability. The ARMA method is more suitable for short-term prediction of stationary sequences, while the SSA method can better process and predict signals with periodic oscillation. Therefore, the combination of the SSA method and the ARMA method and the prediction can make up for the defects of the single method and improve the prediction precision.

The main idea of the SSA + ARMA prediction is to extrapolate and predict the long term and the period term by SSA, predict the residual term by ARMA, and add the predicted values of the two as the predicted value of the final time series.

Referring to fig. 5, the specific implementation steps are as follows:

(1) if a time sequence with the length of N is to be predicted, adding N0 values to the end of the original sequence (with the length of N) to form a new sequence with the length of N + N;

(2) performing SSA decomposition on the new sequence (with the length of N + N), selecting N values at the end of the first component (RC1) as predicted values, replacing the original sequence to form a new sequence, and repeating the process until the difference between the two decomposed RC1 is smaller than a threshold value;

(3) after the step (2) is completed, performing SSA decomposition on the time sequence, obtaining a new sequence by overlapping n values at the tail ends of RC1 and RC2 as predicted values, and repeating the process until the difference between the two decomposed RC1+ RC2 is smaller than a threshold value;

(4) repeating the above process until the predicted data is replaced by k RCs, wherein n values at the end of the new sequence are the predicted values RCP of the SSA;

(5) the residual RCs contain other weaker periodic terms and noise information, and the ARMA algorithm is used for predicting the residual RCs; and

(6) and adding the SSA predicted value and the ARMA predicted value to obtain a final predicted value.

Fig. 6 is an artificial intelligence prediction system for geocentric movement according to an embodiment of the present invention. Referring to fig. 6, the system 600 includes:

a decomposition reconstruction module 601, configured to perform decomposition reconstruction on the time sequence of the original geocentric motion by using an EMD algorithm to obtain a new time sequence;

an information solving module 603 for solving a period and a corresponding amplitude of the new time series; and

a prediction module 605 configured to predict a periodic term of the geocentric motion using the SSA and ARMA combined model based on the period and the corresponding amplitude.

From the above, the respective modules 601 to 605 of the system 600 may respectively perform the respective steps of the artificial intelligence prediction method for geocentric motion described with reference to the above embodiments, and details thereof will not be described here.

According to the method, the EMD algorithm is adopted to decompose and reconstruct the time sequence of the geocentric motion, so that the periodic term of the geocentric motion can be effectively identified; in addition, the prediction precision of the geocentric movement of 1mm is realized by constructing an artificial intelligent geocentric movement prediction model, namely performing long-term, medium-term and short-term prediction on periodic items of the geocentric movement by using an SSA and ARMA combined model.

In another aspect, the present invention provides an electronic device. As shown in fig. 7, electronic device 700 includes a processor 701, a memory 702, a communication interface 703, and a communication bus 704.

The processor 701, the memory 702 and the communication interface 703 complete mutual communication through a communication bus 704;

the processor 701 is used for calling the computer program in the memory 702, and the steps of the artificial intelligence prediction method for geocentric motion provided by the embodiment of the invention as described above are realized when the computer program is executed by the processor 701.

Further, the computer program in the memory may be implemented in the form of a software functional unit and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on such understanding, the technical solution of the present invention or a part thereof, which essentially contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium and includes several computer programs to make a computer device (which may be a personal computer, a server, or a network device) execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

In another aspect, the present invention provides a non-transitory computer readable storage medium, having a computer program stored thereon, where the computer program is executed by a processor to implement the steps of the artificial intelligence prediction method for geocardiac motion provided by the embodiments of the present invention as described above.

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.

Finally, it should be noted that: the above examples are only for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.