Low earth orbit satellite orbit forecasting method based on feature screening and related device

Abstract

The invention discloses a low earth orbit satellite orbit forecasting method based on feature screening and a related device. The method comprises the following steps: defining a pseudo resistance coefficient according to an atmospheric resistance perturbation formula; an SHAP method is adopted to analyze influence values of different space environment characteristics on a pseudo resistance coefficient prediction result of the target prediction model, and an influence bee colony graph is generated; screening out an optimal feature combination based on the influence bee colony graph, and predicting a pseudo resistance coefficient according to the optimal feature combination and the target prediction model to obtain a pseudo resistance prediction value; and performing prediction calculation on the orbit state of the low-orbit satellite according to the pseudo resistance prediction value to obtain an orbit prediction result. The process does not depend on orbit forecast errors and is not influenced by uncertainty of error divergence directions. Therefore, the technical problem that the prediction result is lack of accuracy and reliability due to the fact that the existing orbit prediction depends on the prediction error but the error is uncertain relative to the divergence direction of the real atmospheric resistance can be solved.

Description

Technical Field

[0001] This application relates to the field of satellite technology, and particularly to a low-earth orbit satellite orbit prediction method based on feature screening and related devices. Background Art

[0002] With the rapid development of space technology and the increasing complexity of satellite missions, orbit prediction technology is of great significance in maintaining the safe operation of space targets and optimizing orbit management. The current orbit prediction method for space targets is based on the framework of orbit determination and prediction. In the orbit determination process, observation data is used to obtain accurate historical orbit data and estimate relevant parameters. However, in the orbit prediction stage, observation data is not available, and the same physical model as the orbit determination process and the estimated atmospheric drag coefficient must be used to achieve orbit prediction.

[0003] Due to the low orbit altitude, low-earth orbit satellites are affected by the combined action of conservative forces such as the earth's gravity and the third-body gravity, as well as non-conservative forces such as atmospheric drag. The accuracy of orbit prediction is extremely vulnerable to the interference of environmental uncertainties. In particular, atmospheric drag is severely affected by the space environment and is the main source of error. In medium- and long-term orbit prediction, the error will accumulate significantly over time.

[0004] However, the existing orbit prediction methods based on machine learning modeling usually require the premise that the orbit prediction error follows a certain rule, such as the amplitude fluctuation of the prediction error or the divergence direction of the error; while the orbit situation of low-earth orbit satellites is difficult to accurately predict in the real space environment, so the divergence direction of the error relative to the real atmospheric drag is uncertain, and the prediction result lacks accuracy and reliability. Summary of the Invention

[0005] This application provides a low-earth orbit satellite orbit prediction method based on feature screening and related devices, which is used to solve the technical problem that the existing orbit prediction depends on the prediction error, but the divergence direction of the error relative to the real atmospheric drag is uncertain, resulting in the lack of accuracy and reliability of the prediction result.

[0006] In view of this, the first aspect of this application provides a low-earth orbit satellite orbit prediction method based on feature screening, including:

[0007] Define a pseudo-drag coefficient according to the atmospheric drag perturbation formula;

[0008] Use the SHAP method to analyze the influence value of different space environment characteristics on the pseudo-drag coefficient prediction result of the target prediction model, and generate an influence swarm plot;

[0009] Based on the influence swarm plot, screen out the optimal feature combination, and predict the pseudo-drag coefficient according to the optimal feature combination and the target prediction model to obtain the pseudo-drag prediction value;

[0010] Predict and calculate the orbital state of the LEO satellite based on the predicted pseudo-drag value to obtain the orbit prediction result.

[0011] Preferably, before using the SHAP method to analyze the influence value of different space environment characteristics on the predicted pseudo-drag coefficient result of the target prediction model and generating an influence swarm plot, it further includes:

[0012] Combine the components of multiple measured space environment characteristics into a spliced feature vector;

[0013] Construct a pseudo-drag coefficient prediction model based on the spliced feature vector and the preset pseudo-drag coefficient to obtain the target prediction model.

[0014] Preferably, before constructing a pseudo-drag coefficient prediction model based on the spliced feature vector and the preset pseudo-drag coefficient to obtain the target prediction model, it further includes:

[0015] Use numerical integration to predict the orbital state of the LEO satellite to obtain the predicted ephemeris;

[0016] Calculate the traditional prediction error based on the orbit dynamics model and the predicted ephemeris;

[0017] Through the golden section search algorithm, perform iterative optimization analysis according to the traditional prediction error to obtain the preset pseudo-drag coefficient.

[0018] Preferably, predicting and calculating the orbital state of the LEO satellite based on the predicted pseudo-drag value to obtain the orbit prediction result includes:

[0019] Calculate the atmospheric drag perturbation value based on the predicted pseudo-drag value;

[0020] Substitute the atmospheric drag perturbation value into the orbit dynamics model formula to predict and calculate the orbital state of the LEO satellite to obtain the orbit prediction result.

[0021] The second aspect of the present application provides a LEO satellite orbit prediction device based on feature screening, including:

[0022] A coefficient definition unit for defining a pseudo-drag coefficient according to the atmospheric drag perturbation formula;

[0023] An influence analysis unit for using the SHAP method to analyze the influence value of different space environment characteristics on the predicted pseudo-drag coefficient result of the target prediction model and generating an influence swarm plot;

[0024] A drag prediction unit for screening out the optimal feature combination based on the influence swarm plot and predicting the pseudo-drag coefficient according to the optimal feature combination and the target prediction model to obtain the predicted pseudo-drag value;

[0025] An orbit prediction unit, configured to perform prediction calculations on the orbit state of a low-earth orbit satellite based on the predicted pseudo-drag value to obtain an orbit prediction result.

[0026] Preferably, it further includes:

[0027] A feature processing unit, configured to combine the components of multiple measured space environment features into a spliced feature vector;

[0028] A model construction unit, configured to construct a pseudo-drag coefficient prediction model based on the spliced feature vector and a preset pseudo-drag coefficient to obtain a target prediction model.

[0029] Preferably, it further includes:

[0030] An ephemeris prediction unit, configured to predict the orbit state of a low-earth orbit satellite by means of numerical integration to obtain a predicted ephemeris;

[0031] An error calculation unit, configured to calculate a traditional prediction error based on an orbit dynamics model and the predicted ephemeris;

[0032] An iterative calculation unit, configured to perform iterative optimization analysis based on the traditional prediction error through a golden section search algorithm to obtain a preset pseudo-drag coefficient.

[0033] Preferably, the orbit prediction unit is specifically configured to:

[0034] Calculate the atmospheric drag perturbation value based on the predicted pseudo-drag value;

[0035] Substitute the atmospheric drag perturbation value into the orbit dynamics model formula to perform prediction calculations on the orbit state of a low-earth orbit satellite to obtain an orbit prediction result.

[0036] A third aspect of the present application provides a low-earth orbit satellite orbit prediction device based on feature screening, and the device includes a processor and a memory;

[0037] The memory is used to store program code and transmit the program code to the processor;

[0038] The processor is used to execute the low-earth orbit satellite orbit prediction method based on feature screening described in the first aspect according to the instructions in the program code.

[0039] A fourth aspect of the present application provides a computer-readable storage medium, and the computer-readable storage medium is used to store program code, and the program code is used to execute the low-earth orbit satellite orbit prediction method based on feature screening described in the first aspect.

[0040] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0041] In this application, a low-Earth orbit satellite orbit prediction method based on feature screening is provided, including: defining a pseudo-drag coefficient according to the atmospheric drag perturbation formula; using the SHAP method to analyze the influence values of different space environment features on the pseudo-drag coefficient prediction results of the target prediction model, and generating an influence swarm plot; screening out the optimal feature combination based on the influence swarm plot, and predicting the pseudo-drag coefficient according to the optimal feature combination and the target prediction model to obtain a pseudo-drag prediction value; predicting and calculating the low-Earth orbit satellite orbit state based on the pseudo-drag prediction value to obtain an orbit prediction result.

[0042] The low-Earth orbit satellite orbit prediction method based on feature screening provided in this application defines the concept of pseudo-drag coefficient, and converts the problem of low-Earth orbit satellite orbit prediction into a pseudo-drag coefficient prediction problem; then uses the SHAP method to analyze the pseudo-drag coefficient prediction process based on space environment parameters and prediction models. This process can determine the influence values of different space environment features on the model prediction results, and based on these influence values, a space environment feature combination that is more helpful for the prediction results can be screened out. The pseudo-drag prediction value predicted according to this feature combination is more in line with the characteristics of the actual environment, which can ensure the accuracy and reliability of the low-Earth orbit satellite orbit results predicted on this basis. And this process does not depend on orbit prediction errors, so it will not be affected by the uncertainty of the error divergence direction. Therefore, this application can solve the technical problem that the existing orbit prediction depends on prediction errors, but the divergence direction of the error relative to the true atmospheric drag is uncertain, resulting in the lack of accuracy and reliability of the prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic flow chart of a low-Earth orbit satellite orbit prediction method based on feature screening provided by an embodiment of this application;

[0044] Figure 2 It is a schematic structural diagram of a low-Earth orbit satellite orbit prediction device based on feature screening provided by an embodiment of this application;

[0045] Figure 3 It is a schematic diagram of the radial error of the orbit prediction based on the pseudo-drag coefficient and the traditional numerical method provided by an embodiment of this application;

[0046] Figure 4 It is a schematic diagram of the track error of the orbit prediction based on the pseudo-drag coefficient and the traditional numerical method provided by an embodiment of this application;

[0047] Figure 5 It is a schematic diagram of the normal error of the orbit prediction based on the pseudo-drag coefficient and the traditional numerical method provided by an embodiment of this application;

[0048] Figure 6 It is an influence swarm plot drawn based on the Shapley value provided by an embodiment of this application. Detailed implementation manners

[0049] To enable those skilled in the art to better understand the solutions of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part rather than all of the embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the scope of protection of this application.

[0050] For ease of understanding, please refer to Figure 1 , an embodiment of a low-Earth orbit satellite orbit prediction method based on feature screening provided by this application, includes:

[0051] Step 101: Define a pseudo-drag coefficient according to the atmospheric drag perturbation formula.

[0052] The atmospheric drag perturbation formula is expressed as:

[0053]

[0054] where a drag represents the atmospheric drag perturbation, which refers to the acceleration caused by the atmospheric drag, is the mass of the current satellite, is the windward area, is the area-to-mass ratio, is the atmospheric drag coefficient, is the atmospheric density, which is related to the orbital altitude, represents the modulus of v e , and v e represents the velocity vector of the spacecraft relative to the atmosphere. All the parameters on the right side of the equation have uncertainties. Most studies focus on the modeling of the atmospheric density , which has a complex dependence relationship with space environment parameters such as geomagnetic index and solar radiation flux index. The atmospheric drag coefficient is estimated together with the state variables of the spacecraft in traditional precise orbit determination and prediction, and is used to absorb the errors of the atmospheric density model, satellite area-to-mass ratio, and satellite attitude change. The partial derivative of the atmospheric drag perturbation a drag with respect to is expressed as:

[0055]

[0056] According to the above two formulas, it can be seen that the change of space environment parameters will affect the calculation of the atmospheric density , and further will also affect the calculation of the atmospheric drag coefficient . Moreover, in the case of medium- and long-term orbit prediction or obvious changes in the space environment, the atmospheric density It is difficult to accurately estimate, which in turn affects the calculation of the atmospheric drag coefficient , resulting in a decrease in the accuracy of orbit prediction and an increase in error.

[0057] It can be found that in this embodiment, space environment parameters are used to calculate the atmospheric density ; then, using the real orbit data, the changes in the space environment are introduced into the atmospheric drag coefficient . At this time, the space environment parameters for calculating the atmospheric density are set to fixed values, and the atmospheric drag coefficient no longer requires orbit determination to solve, and at the same time, it loses its original physical meaning. Therefore, in this embodiment, it is defined as the pseudo drag coefficient .

[0058] Step 102: Use the SHAP method to analyze the influence values of different space environment characteristics on the prediction results of the pseudo drag coefficient of the target prediction model, and generate an influence swarm plot.

[0059] Furthermore, before step 102, it also includes:

[0060] Combine the components of multiple measured space environment characteristics into a spliced feature vector;

[0061] Based on the spliced feature vector and the preset pseudo drag coefficient, construct a pseudo drag coefficient prediction model to obtain the target prediction model.

[0062] Furthermore, before constructing the pseudo drag coefficient prediction model based on the spliced feature vector and the preset pseudo drag coefficient to obtain the target prediction model, it also includes:

[0063] Use numerical integration to predict the orbit state of a low-earth orbit satellite to obtain a predicted ephemeris;

[0064] Calculate the traditional prediction error based on the orbit dynamics model and the predicted ephemeris;

[0065] Through the golden section search algorithm, perform iterative optimization analysis according to the traditional prediction error to obtain the preset pseudo drag coefficient.

[0066] It should be noted that the target prediction model is pre-constructed according to the space environment characteristics and the preset pseudo drag coefficient, and this model can predict the pseudo drag coefficient based on the space environment characteristics. The preset pseudo drag coefficient here is calculated based on the traditional method; for the specific process, please refer to Figure 3 . In this embodiment, the changes in the space environment are introduced into the atmospheric drag coefficient using the real orbit data. At this time, the space environment parameters for calculating the atmospheric density are fixed, and the pseudo drag coefficient can be calculated . And the pseudo drag coefficient The solution requires a high-precision orbital dynamics model and real satellite orbit data; the orbital dynamics model adopted in this embodiment is in the Earth-Centered Inertial (ECI) coordinate system:

[0067]

[0068] Among them, and respectively represent the position and velocity of the space target in the inertial system, , represents the Euclidean norm of the vector; is the Earth's gravitational constant; represents the perturbation acceleration vector, which consists of non-spherical perturbation , three-body gravitational perturbation , atmospheric drag perturbation , solar radiation pressure perturbation and other unmodeled perturbations and so on. Except for the atmospheric drag perturbation, the modeling of the other perturbations is accurate enough or has little impact on low-Earth orbit satellites, so they will not be specifically discussed.

[0069] Set the space environment parameters for calculating the atmospheric density as fixed values, given the initial epoch and the initial orbit state of the low-Earth orbit satellite , , the orbit state at any future epoch can be regarded as a single-variable function of the pseudo-drag coefficient , specifically as the following formula:

[0070]

[0071] can be solved by numerical integration calculation using the above-mentioned orbital dynamics model to obtain the predicted ephemeris.

[0072] Using satellite measurement data as the real satellite orbit data, which is called the precise ephemeris. Based on the precise ephemeris and with the principle of minimizing the prediction error e(t), the objective function for minimizing the prediction error can be constructed:

[0073]

[0074] Among them, x true (t) is the precise ephemeris data in the inertial system, and e(t) is the error value in the orbit coordinate system along the track, normal, and radial directions, that is, the traditional prediction error.

[0075] Pseudo-drag coefficient The optimization calculation process can be regarded as a problem of finding the minimum value of a single variable function within a certain interval. The golden section search is a method of finding the extreme value of a function by searching within a known interval through the golden ratio and continuously reducing the maximum value of the function within the known interval. Therefore, this embodiment selects the golden section search algorithm to perform iterative optimization calculation according to the objective function of the forecast error to obtain the pseudo resistance coefficient when the forecast error is the smallest. It can be understood that the pseudo drag coefficient calculated by the above method can be recorded as the preset pseudo drag coefficient , which can also be called the historical pseudo-drag coefficient, because this coefficient is calculated based on the real satellite orbit data.

[0076] From the process of solving the pseudo-drag coefficient, it can be seen that the determination of the pseudo-drag coefficient is related to the forecast duration. This embodiment considers improving the medium- and long-term orbit forecast of low-orbit satellites, and the forecast duration is selected as 10 days. The corresponding pseudo-drag coefficient can be solved based on this. Taking a low-orbit satellite as an example, the orbit altitude is about 400km. Under the same conditions, the preset pseudo-drag coefficient forecast and the traditional method forecast are performed respectively. The orbit forecast error in the RTN coordinate system is shown in the following figure. Figure 3 , Figure 4 and Figure 5 .

[0077] By comparing the forecast errors, it can be found that the forecast based on the preset pseudo-drag coefficient in this embodiment has obvious improvement in the total position error compared with the traditional forecast, which is mainly reflected in the along-track direction. The forecast results based on the preset pseudo-drag coefficient are more consistent with those of the traditional forecast.

[0078] The prediction error comparison in this embodiment is based on a unified normalized scaling operation, using the pseudo drag coefficient The maximum prediction error of is unit quantity, and the scaled error expression is:

[0079]

[0080] in, is the error after scaling, is the selected pseudo drag coefficient The forecast error, is other forms of forecast error, represent radial, track and normal directions respectively.

[0081] After analysis, it can be seen that the orbit prediction error of low-orbit satellites is mainly manifested in the direction of the track, and the magnitude of the track error is much greater than that of the radial and normal directions; the pseudo-drag coefficient It can effectively absorb the orbit prediction error, which mainly affects the track error. The radial error is not as large as the track error but can still be improved, and the normal error has basically no effect; preset pseudo drag coefficient The prediction can approximately represent the prediction results of traditional methods, and also illustrate the scientificity and rationality of the pseudo-drag coefficient proposed in this embodiment. Concept.

[0082] Taking the space environment characteristics as the model input vector and the preset pseudo-drag coefficient as the output vector, a prediction model of the pseudo-drag coefficient, that is, the target prediction model, can be constructed. Since this embodiment needs to analyze the influence of different space environment characteristics on the model output, the feature vector input to the model is an overall vector composed of the components of multiple measured space environment characteristics, that is, the spliced feature vector; and the model can analyze the importance of different partial feature components, and then screen out the features corresponding to the optimal output prediction result as the optimal feature combination.

[0083] It should be noted that this embodiment uses the SHAP method to quantify the importance of the components of different space environment characteristics to the model output, specifically manifested as the influence value, that is, the Shapley value. The SHAP (SHapley Additive exPlanations) method is used to explain the prediction of machine learning models, based on the Shapley value in game theory. The Shapley value is a fair distribution of the payoffs of a game, used to determine the contribution of each player to the overall result of the game. The players in this embodiment are the various different space environment characteristics.

[0084] For the features in the feature set The Shapley value is calculated as follows:

[0085]

[0086] Where is the set of all features, is any feature subset of that does not contain the feature; represents the number of features in the set , represents the contribution of the feature set to the model prediction output, represents the contribution of the feature set containing the feature to the model prediction output.

[0087] In this embodiment, the Shapley value is used to quantify the influence value of the components of different space environment characteristics on the model output. The specific space environment characteristics include but are not limited to the solar radiation flux index, the geomagnetic index Kp, the full-day planetary magnetic index Cp, the Bartels solar rotation period, and the international sunspot number; the specific space environment characteristics can be obtained and selected from the Space Environment Prediction Center - SEPC. This is only an example here and is not limited. As long as the characteristic parameters related to the prediction of the low-Earth orbit satellite orbit are considered, the influence value on the model output can be analyzed; and then an influence swarm plot is generated. Please refer to Figure 6 . In the figure, the dark dots indicate that the characteristic value has a positive impact on the model prediction in this observation; the light dots indicate that the characteristic value has a negative impact on the model prediction in this observation; the horizontal axis is the SHAP value, which shows the magnitude of the influence value. The farther the dot is from the central zero line, the greater the influence of the characteristic on the model output, and vice versa; in the vertical arrangement direction, the characteristics are sorted in descending order of influence from top to bottom. The characteristics above have a greater overall impact on the model output, while the characteristics below have a smaller impact; the characteristics in the middle show dots of two different depths of color, but the distribution of the dots is more concentrated and the influence is relatively small; the characteristics at the bottom have the smallest impact on the model, and most of the impacts are relatively close to zero, indicating that these characteristics contribute less to the model prediction.

[0088] Step 103: Screen out the optimal feature combination based on the influence swarm plot, and predict the pseudo-drag coefficient according to the optimal feature combination and the target prediction model to obtain the pseudo-drag prediction value.

[0089] Based on the influence swarm plot, the optimal feature combination can be screened out, and based on this feature combination, a more realistic pseudo-drag coefficient, that is, the pseudo-drag prediction value, can be predicted. The original space environment data is used in this embodiment, and these data do not need to be preprocessed. Through model analysis, a richer and more effective optimal feature combination can be screened out. This method can ensure the accuracy and reliability of the pseudo-drag prediction value.

[0090] Step 104: Predict and calculate the orbit state of the low-Earth orbit satellite based on the pseudo-drag prediction value to obtain the orbit prediction result.

[0091] Further, step 104 includes:

[0092] Calculate the atmospheric drag perturbation value based on the pseudo-drag prediction value;

[0093] Substitute the atmospheric drag perturbation value into the orbit dynamics model formula to predict and calculate the orbit state of the low-Earth orbit satellite to obtain the orbit prediction result.

[0094] The atmospheric drag perturbation value can be calculated according to the atmospheric drag perturbation formula. Substitute the obtained atmospheric drag perturbation value into the formula of the orbital dynamics model to calculate the specific position and velocity of the satellite in the orbit, and the state prediction result of the satellite orbit can be obtained. Since the pseudo-drag prediction value is predicted based on the optimized screening method, its reliability can be ensured. Therefore, the orbit prediction result calculated on this basis has accuracy and reliability.

[0095] The low-Earth orbit satellite orbit prediction method based on feature screening provided by the embodiments of the present application defines the concept of pseudo-drag coefficient, and converts the problem of low-Earth orbit satellite orbit prediction into a pseudo-drag coefficient prediction problem; then uses the SHAP method to analyze the pseudo-drag coefficient prediction process based on space environment parameters and prediction models. This process can determine the influence value of different space environment characteristics on the model prediction result, and based on this influence value, a combination of space environment characteristics that is more helpful for the prediction result can be screened out. The pseudo-drag prediction value predicted according to this characteristic combination is more in line with the characteristics of the actual environment, which can ensure the accuracy and reliability of the low-Earth orbit satellite orbit result predicted on this basis. And this process does not depend on the orbit prediction error, so it will not be affected by the uncertainty of the error divergence direction. Therefore, the embodiments of the present application can solve the technical problem that the existing orbit prediction depends on the prediction error, but the divergence direction of the error relative to the true atmospheric drag is uncertain, resulting in the lack of accuracy and reliability of the prediction result.

[0096] For ease of understanding, please refer to Figure 2 , an embodiment of a low-Earth orbit satellite orbit prediction device based on feature screening provided by the present application includes:

[0097] A coefficient definition unit 201 for defining a pseudo-drag coefficient according to the atmospheric drag perturbation formula;

[0098] An influence analysis unit 202 for using the SHAP method to analyze the influence value of different space environment characteristics on the pseudo-drag coefficient prediction result of the target prediction model and generating an influence swarm plot;

[0099] A drag prediction unit 203 for screening out the optimal feature combination based on the influence swarm plot and predicting the pseudo-drag coefficient according to the optimal feature combination and the target prediction model to obtain a pseudo-drag prediction value;

[0100] An orbit prediction unit 204 for predicting and calculating the low-Earth orbit satellite orbit state according to the pseudo-drag prediction value to obtain an orbit prediction result.

[0101] Furthermore, it further includes:

[0102] A feature processing unit 205 for combining the components of multiple measured space environment characteristics into a spliced feature vector;

[0103] A model construction unit 206, configured to construct a pseudo drag coefficient prediction model based on the spliced feature vector and a preset pseudo drag coefficient to obtain a target prediction model.

[0104] Further, it further includes:

[0105] An ephemeris prediction unit 207, configured to predict the orbit state of a low-earth orbit satellite by means of numerical integration to obtain a predicted ephemeris;

[0106] An error calculation unit 208, configured to calculate a traditional prediction error based on an orbit dynamics model and a predicted ephemeris;

[0107] An iterative calculation unit 209, configured to perform iterative optimization analysis according to the traditional prediction error through a golden section search algorithm to obtain a preset pseudo drag coefficient.

[0108] Further, the orbit prediction unit 204 is specifically configured to:

[0109] Calculate an atmospheric drag perturbation value according to a pseudo drag prediction value;

[0110] Substitute the atmospheric drag perturbation value into the formula of the orbit dynamics model to perform prediction calculation on the orbit state of the low-earth orbit satellite to obtain an orbit prediction result.

[0111] The present application further provides a low-earth orbit satellite orbit prediction device based on feature screening. The device includes a processor and a memory;

[0112] The memory is used to store program codes and transmit the program codes to the processor;

[0113] The processor is configured to execute the low-earth orbit satellite orbit prediction method based on feature screening in the above method embodiments according to the instructions in the program codes.

[0114] The present application further provides a computer-readable storage medium. The computer-readable storage medium is used to store program codes, and the program codes are used to execute the low-earth orbit satellite orbit prediction method based on feature screening in the above method embodiments.

[0115] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces. The indirect coupling or communication connection of the devices or units can be in an electrical, mechanical or other form.

[0116] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0117] In addition, each functional unit in various embodiments of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0118] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (English full name: Read-Only Memory, English abbreviation: ROM), random access memories (English full name: Random Access Memory, English abbreviation: RAM), magnetic disks, or optical discs and other various media that can store program codes.

[0119] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of various embodiments of the present application.

Claims

1. A method for predicting low-orbit satellite orbits based on feature screening, characterized in that: include: The pseudo drag coefficient is defined according to the atmospheric drag perturbation formula; The SHAP method is used to analyze the impact of different spatial environmental characteristics on the pseudo drag coefficient prediction results of the target prediction model and generate an impact bee swarm diagram. Based on the influence bee swarm diagram, an optimal feature combination is selected, and a pseudo resistance coefficient is predicted according to the optimal feature combination and the target prediction model to obtain a pseudo resistance prediction value; The orbit state of the low-orbit satellite is predicted and calculated according to the pseudo-drag prediction value to obtain an orbit prediction result.

2. The method for predicting low-orbit satellite orbits based on feature screening according to claim 1, characterized in that: The SHAP method is used to analyze the influence of different spatial environment characteristics on the prediction results of the pseudo drag coefficient of the target prediction model, and generate an influence bee swarm diagram, which also includes: Combining components of multiple measured spatial environmental features into a concatenated feature vector; A pseudo-drag coefficient prediction model is constructed based on the spliced ​​feature vector and the preset pseudo-drag coefficient to obtain a target prediction model.

3. The method for predicting low-orbit satellite orbits based on feature screening according to claim 2, characterized in that: The pseudo drag coefficient prediction model is constructed based on the spliced ​​feature vector and the preset pseudo drag coefficient to obtain a target prediction model, and the method further includes: The orbit state of low-orbit satellites is predicted by numerical integration to obtain the predicted ephemeris; Calculating a traditional prediction error based on the orbital dynamics model and the predicted ephemeris; The preset pseudo drag coefficient is obtained by performing iterative optimization analysis based on the traditional prediction error through the golden section search algorithm.

4. The method for predicting low-orbit satellite orbits based on feature screening according to claim 1, characterized in that: The step of predicting and calculating the orbit state of the low-orbit satellite according to the pseudo-drag prediction value to obtain an orbit prediction result includes: Calculating an atmospheric resistance perturbation value according to the pseudo resistance prediction value; The atmospheric drag perturbation value is substituted into the orbital dynamics model formula to predict and calculate the orbital state of the low-orbit satellite to obtain an orbit prediction result.

5. A low-orbit satellite orbit prediction device based on feature screening, characterized in that: include: A coefficient definition unit is used to define a pseudo drag coefficient according to the atmospheric drag perturbation formula; An impact analysis unit is used to analyze the impact of different spatial environmental characteristics on the pseudo drag coefficient prediction results of the target prediction model using the SHAP method, and generate an impact bee swarm diagram; a resistance prediction unit, configured to select an optimal feature combination based on the influence bee swarm diagram, and predict a pseudo resistance coefficient according to the optimal feature combination and the target prediction model to obtain a pseudo resistance prediction value; The orbit prediction unit is used to predict and calculate the orbit state of the low-orbit satellite according to the pseudo-drag prediction value to obtain an orbit prediction result.

6. The low-orbit satellite orbit prediction device based on feature screening according to claim 5, characterized in that: Also includes: A feature processing unit, used to combine components of multiple measured spatial environment features into a spliced ​​feature vector; The model building unit is used to build a pseudo drag coefficient prediction model based on the spliced ​​feature vector and the preset pseudo drag coefficient to obtain a target prediction model.

7. The low-orbit satellite orbit prediction device based on feature screening according to claim 6, characterized in that: Also includes: An ephemeris prediction unit is used to predict the orbit state of the low-orbit satellite by numerical integration to obtain a predicted ephemeris; An error calculation unit, used for calculating a traditional prediction error based on an orbital dynamics model and the predicted ephemeris; The iterative calculation unit is used to perform iterative optimization analysis according to the traditional prediction error through a golden section search algorithm to obtain a preset pseudo drag coefficient.

8. The low-orbit satellite orbit prediction device based on feature screening according to claim 5, characterized in that: The orbit prediction unit is specifically used for: Calculating an atmospheric resistance perturbation value according to the pseudo resistance prediction value; The atmospheric drag perturbation value is substituted into the orbital dynamics model formula to predict and calculate the orbital state of the low-orbit satellite to obtain an orbit prediction result.

9. A low-orbit satellite orbit prediction device based on feature screening, characterized in that: The device comprises a processor and a memory; The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the low-orbit satellite orbit prediction method based on feature screening as described in any one of claims 1-4 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the low-orbit satellite orbit prediction method based on feature screening as described in any one of claims 1-4.

Images