A method for predicting the high-latitude ionosphere based on multi-source heterogeneous data

Abstract

This invention belongs to the field of high-latitude ionospheric technology, and specifically relates to a high-latitude ionospheric forecasting method based on multi-source heterogeneous data. The steps are as follows: S1, establish a time-series dataset of ground-based ionospheric measurement data and space weather index sets; S2, establish a time-series dataset of space-based high-energy particle deposition flux measurement data; S3, normalize all time-series datasets to establish an input dataset based on multi-source heterogeneous data; S4, establish an LSTM prediction neural network, and obtain the forecast results of high-latitude ionospheric parameters based on the input dataset. This method, based on ground-based ionospheric measurement data and space weather indices, introduces space-based high-energy particle deposition flux measurement data to obtain a normalized time-series dataset, thereby establishing an input dataset based on multi-source heterogeneous data. Then, an LSTM prediction neural network is used to obtain the forecast results of high-latitude ionospheric parameters.

Description

Technical Field

[0001] This invention belongs to the field of high-latitude ionospheric technology, and specifically relates to a high-latitude ionospheric forecasting method based on multi-source heterogeneous data. Background Technology

[0002] Ionospheric parameter forecasting is crucial for the operation and maintenance of communication and navigation systems in high-latitude regions. This region is significantly affected by geomagnetic activity and polar particle deposition, especially during geomagnetic storms, when high-energy particles penetrate the atmosphere, causing a sharp increase in ionospheric electron density, leading to problems such as shortwave communication attenuation and a surge in GNSS positioning errors. Existing ionospheric parameter forecasting models largely rely on the solar radiation index (F10.7) and geomagnetic indices (Kp, Dst) as inputs, but fail to fully incorporate direct observational data on high-energy particle deposition flux, resulting in insufficient predictive ability for sudden ionospheric disturbances.

[0003] Current mainstream methods fall into two categories: one is the International Reference Ionospheric Model (IRI), based on empirical formulas, which predicts electron density distribution through historical statistical relationships, but suffers from severe lag in its dynamic response to particle deposition events; the other is machine learning models based on time series algorithms such as LSTM, which can capture parameter changes over time, but their training data is limited to geomagnetic indices and historical ionospheric values, failing to consider the physical causal relationship between particle deposition and ionospheric disturbances. For example, NOAA's real-time ionospheric correction model (STORM-E) showed a 52% prediction error for the TEC peak during the strong geomagnetic storm event in November 2022, even when deposition flux data was ignored.

[0004] The limitations of existing technologies are mainly reflected in: (1) High-energy particle deposition flux is the direct driving factor of sudden changes in ionospheric electron density, and its data has not been included in the input feature set of mainstream forecast models; (2) Satellite-measured particle deposition data (such as POES and Swarm satellite detection results) and ground ionospheric observations have a spatiotemporal resolution mismatch problem, which restricts the efficiency of multi-source data fusion.

[0005] Therefore, the aforementioned problems urgently need to be addressed. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a high-latitude ionospheric forecasting method based on multi-source heterogeneous data. This method, based on ground-based ionospheric measurement data and space weather indices, incorporates space-based high-energy particle deposition flux measurement data to obtain a normalized time-series dataset. This dataset forms the input dataset for the multi-source heterogeneous data. Finally, an LSTM prediction neural network is used to obtain the forecast results for high-latitude ionospheric parameters.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for predicting the high-latitude ionosphere based on multi-source heterogeneous data, comprising the following steps:

[0008] S1, establish a time-series dataset of ground-based ionospheric measurement data and space weather index set;

[0009] S10, ground-based ionospheric measurement data, refers to TEC data obtained by ground receivers, which is the total number of free electrons in a unit cross-sectional area cylinder perpendicular to the propagation path.

[0010] S11 is a set of space weather indices, including the Dst index, the IMF-Bz index, and the F10.7 index.

[0011] TEC: Total Electron Content; Dst Index: Geomagnetic Storm Circulation Current Index; IMF-Bz Index: North-South Component Index of Interplanetary Magnetic Field; F10.7 Index: 10.7 cm Solar Radiocurrent Index.

[0012] S2, Establish a time-series dataset of space-based high-energy particle sedimentation flux measurement data;

[0013] S20 satellites are in continuous motion and cannot perform continuous measurements on a single point. However, high-energy particle precipitation mainly occurs in the auroral zone above magnetic latitude 60°, and all satellite measurements of high-energy particle precipitation flux in this region can be recorded.

[0014] S3, normalizes all time-series datasets and establishes an input dataset based on multi-source heterogeneous data;

[0015] S4. Establish an LSTM prediction neural network to obtain prediction results of high-latitude ionospheric parameters based on the input dataset.

[0016] Further, step S10 includes:

[0017] Using ionospheric TEC data measured by a ground-based GNSS receiver located in a high-latitude region as ground-based measurement data, a time-series dataset of TEC data was established according to a certain time resolution, and S... TEC express:

[0018] Formula 1;

[0019] in, Indicates t i The values ​​of ionospheric TEC data at time t, where t is time, t0 is the start time, and the subscript i is the time series number, N i Where Δt is the total number of data points, and Δt is the time resolution.

[0020] Step S11 includes:

[0021] Time series datasets for the Dst index, IMF-Bz index, and F10.7 index were constructed according to a certain time resolution, and were respectively used with S... Dst S Bz and S F107 express:

[0022] Formula 2;

[0023] Establish a time series dataset of space weather indices ;

[0024] Among them, Dst , , They represent t respectively i The values ​​of the Dst exponent, IMF-Bz exponent, and F10.7 exponent at time t, where t is time, t0 is the start time, and the subscript i is the time series number, N i Δt represents the total number of data points and the time resolution.

[0025] Further, step S20 includes:

[0026] Averaging all measurements in S1 using time resolution yields a time-series dataset of high-energy particle sedimentation flux measurements. P express:

[0027] Formula 3;

[0028] Among them, P k This represents the measured value of high-energy particle sedimentation flux, where k is the measurement value number and φ is the magnetic latitude value corresponding to the high-energy particle sedimentation flux. Indicates t i The value of high-energy particle sedimentation flux at time t, where t is time, t0 is the starting time, and the subscript i is the time sequence number, N i Δt represents the total number of data points, and Δt represents the time resolution. 条件 This is an indicator function that takes the value 1 when the condition is met, and 0 otherwise.

[0029] Further, step S3 includes:

[0030] S31, introduces local time series data, using This means that, similarly, normalization is performed on it to obtain a local time sine-normalized time series dataset. Local time cosine value normalized time series dataset :

[0031] Formula 4;

[0032] S32, normalize the time-series datasets of ground-based ionospheric measurement data, space weather index sets, and space-based high-energy particle precipitation flux measurement data; obtain the normalized time-series dataset of TEC data. Normalized time series dataset of Dst exponent Normalized time series dataset of IMF-Bz index Normalized time series dataset of F10.7 exponent Normalized time series dataset of high-energy particle sedimentation flux measurement data :

[0033] Formula 5;

[0034] S33, Five consecutive time series data points are slidably selected from the normalized time series dataset to establish neural network input data samples, which are then processed using... , , , , , and The input dataset is represented as follows: , , , , , and :

[0035] Formula 6.

[0036] Further, step S4 includes:

[0037] S41, Build an LSTM prediction neural network. The input neurons use the input dataset, and the output layer has only one neuron. That is, in advance The ionospheric TEC value corresponding to the time forecast;

[0038] S42, based on the trained prediction neural network, input samples from the five moments before the moment to be predicted are input to obtain the high-latitude ionospheric prediction results;

[0039] S43 uses two parameters to evaluate forecast errors, including RMSE and PD:

[0040] RMSE stands for Root Mean Square Error, and it is calculated as follows:

[0041] Formula 7;

[0042] PD represents the total relative error, and is calculated as follows:

[0043] Formula 8;

[0044] Where j is the sample number used for error statistics, and N j The total number of samples used for error statistics.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] This invention presents a high-latitude ionospheric forecasting method based on multi-source heterogeneous data. By combining high-energy particle deposition flux data with a traditional LSTM model, the accuracy of existing methods for ionospheric forecasting is improved. High-energy particle deposition flux data from the POES / MetOp satellites is introduced into the LSTM model, forming multi-dimensional input features along with ionospheric historical parameters and geomagnetic indices. Experimental results show that this forecasting method reduces the root mean square error of the forecast from 4.25 to 2.59, and the overall relative error from 0.45 to 0.17. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating a high-latitude ionospheric forecasting method based on multi-source heterogeneous data according to the present invention.

[0049] Figure 2 This is a schematic diagram illustrating the changes and correlations of mep0e1, AE index, and TEC data in Embodiment 1 of the present invention;

[0050] Figure 3 This is a schematic diagram illustrating the correlation between the predicted and actual TEC values ​​obtained in Embodiment 1 of the present invention.

[0051] Figure 4 This is a schematic diagram showing the correlation between the TEC forecast value and the actual value obtained by the present invention using existing methods. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Therefore, the following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention.

[0054] Based on the embodiments described in this invention, all other embodiments obtained by those skilled in the art without inventive effort prior to this specification are within the scope of protection of this invention.

[0055] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0056] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0057] like Figure 1 As shown:

[0058] A method for predicting the high-latitude ionosphere based on multi-source heterogeneous data includes the following steps:

[0059] S1, establish a time-series dataset of ground-based ionospheric measurement data and space weather index set;

[0060] TEC: Total Electron Content; Dst Index: Geomagnetic Storm Circulation Current Index; IMF-Bz Index: North-South Component Index of Interplanetary Magnetic Field; F10.7 Index: 10.7 cm Solar Radiocurrent Index.

[0061] S10, ground-based ionospheric measurement data, refers to TEC data obtained by ground receivers, which is the total number of free electrons in a unit cross-sectional area cylinder perpendicular to the propagation path (usually referring to the path from the satellite to the receiver in a 1 square meter cross-section).

[0062] Step S10 includes:

[0063] Using ionospheric TEC data measured by a ground-based GNSS receiver located in a high-latitude region as ground-based measurement data, a time-series dataset of TEC data was established according to a certain time resolution, and S... TEC express:

[0064] Formula 1;

[0065] in, Indicates t i The values ​​of ionospheric TEC data at time t, where t is time, t0 is the start time, and the subscript i is the time series number, N i Where Δt is the total number of data points, and Δt is the time resolution.

[0066] S11 is a set of space weather indices, including the Dst index, the IMF-Bz index, and the F10.7 index.

[0067] Step S11 includes:

[0068] Time series datasets for the Dst index, IMF-Bz index, and F10.7 index were constructed according to a certain time resolution, and were respectively used with S... Dst S Bz and S F107 express:

[0069] Formula 2;

[0070] Establish a time series dataset of space weather indices ;

[0071] Among them, Dst , , They represent t respectively i The values ​​of the Dst exponent, IMF-Bz exponent, and F10.7 exponent at time t, where t is time, t0 is the start time, and the subscript i is the time series number, N i Where Δt is the total number of data points, and Δt is the time resolution.

[0072] The Dst index can be downloaded from websites such as the World Geomagnetic Data Center, the IMF-Bz index can be downloaded from websites such as OMNIWeb, and the F10.7 index can be downloaded from websites such as the Canadian Space Weather Agency.

[0073] S2, Establish a time-series dataset of space-based high-energy particle sedimentation flux measurement data;

[0074] S20 satellites are in continuous motion and cannot perform continuous measurements on a single point. However, high-energy particle precipitation mainly occurs in the auroral zone above magnetic latitude 60°, and all satellite measurements of high-energy particle precipitation flux in this region can be recorded.

[0075] Step S20 includes:

[0076] Averaging all measurements in S1 using time resolution yields a time-series dataset of high-energy particle sedimentation flux measurements. P express:

[0077] Formula 3;

[0078] Among them, P k This represents the measured value of high-energy particle sedimentation flux, where k is the measurement value number and φ is the magnetic latitude value corresponding to the high-energy particle sedimentation flux. Indicates t i The value of high-energy particle sedimentation flux at time t, where t is time, t0 is the starting time, and the subscript i is the time sequence number, N i Δt represents the total number of data points, and Δt represents the time resolution. 条件 This is an indicator function that takes the value 1 when the condition is met, and 0 otherwise.

[0079] S3, normalizes all time-series datasets and establishes an input dataset based on multi-source heterogeneous data;

[0080] Step S3 includes:

[0081] S31, introduces local time series data, using This means that, similarly, normalization is performed on it to obtain a local time sine-normalized time series dataset. Local time cosine value normalized time series dataset :

[0082] Formula 4;

[0083] S32, normalize the time-series datasets of ground-based ionospheric measurement data, space weather index sets, and space-based high-energy particle precipitation flux measurement data; obtain the normalized time-series dataset of TEC data. Normalized time series dataset of Dst exponent Normalized time series dataset of IMF-Bz index Normalized time series dataset of F10.7 exponent Normalized time series dataset of high-energy particle sedimentation flux measurement data :

[0084] Formula 5;

[0085] S33, Five consecutive time series data points are slidably selected from the normalized time series dataset to establish neural network input data samples, which are then processed using... , , , , , and The input dataset is represented as follows: , , , , , and :

[0086] Formula 6.

[0087] S4. Establish an LSTM prediction neural network and obtain prediction results of high-latitude ionospheric parameters based on the input dataset;

[0088] The S4 step includes:

[0089] S41, Build an LSTM prediction neural network. The input neurons use the input dataset, and the output layer has only one neuron. That is, in advance The ionospheric TEC value corresponding to the time forecast;

[0090] S42, based on the trained prediction neural network, input samples from the five moments before the moment to be predicted are input to obtain the high-latitude ionospheric prediction results;

[0091] S43 uses two parameters to evaluate forecast errors, including RMSE and PD:

[0092] RMSE stands for Root Mean Square Error, and it is calculated as follows:

[0093] Formula 7;

[0094] PD (Percentage Deviation) is the total relative error, calculated as follows:

[0095] Formula 8;

[0096] Where j is the sample number used for error statistics, and N j The total number of samples used for error statistics.

[0097] Example 1:

[0098] The forecasting method of this invention uses ionospheric TEC data (2014) measured by a ground-based GNSS receiver at the Eureka Weather Station (EURC) in Canada, with a geomagnetic latitude of 88.5°N. Data for the corresponding Dst, IMF-Bz, and F10.7 indices were obtained. High-energy particle deposition flux measurements were also obtained using mep0e1 data, which represents the deposition electron flux in the >30 keV channel of the POES / MetOp satellite's MEPED instrument.

[0099] Figure 2 The changes in mep0e1, AE index, and TEC data as of January 4, 2014, show a certain correlation between the high-energy particle deposition flux measurement data (mep0e1) and the TEC changes. Furthermore, the cross-correlation coefficient between mep0e1 and TEC in 2014 was calculated to be 0.45, indicating a moderate correlation between them.

[0100] Figure 3 The figures show the correlation statistics between the predicted and actual TEC values ​​obtained using the forecasting method of this invention. For comparison, Figure 4 The correlation statistics between the predicted and actual TEC values ​​obtained by the existing method (without incorporating high-energy particle sedimentation flux) are presented.

[0101] The above comparison shows that the root mean square error (RMSE) of the parameter prediction results obtained by the prediction method of the present invention has decreased from 4.25 to 2.59 compared with the existing methods, and the overall relative error (PD) has decreased from 0.45 to 0.17. This fully demonstrates that the introduction of high-energy particle sedimentation flux in the prediction method of the present invention can improve the accuracy of TEC prediction results.

[0102] The present invention has been described above by way of example with reference to the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above embodiments. Those skilled in the art can make various modifications or variations to the present invention without departing from the technical concept of the present invention, and such modifications or variations naturally fall within the protection scope of the present invention.

Claims

1. A method for predicting the high-latitude ionosphere based on multi-source heterogeneous data, characterized in that: Includes the following steps: S1, establish a time-series dataset of ground-based ionospheric measurement data and space weather index set; S10, ground-based ionospheric measurement data, refers to TEC data obtained by ground receivers, which is the total number of free electrons in a unit cross-sectional area cylinder perpendicular to the propagation path. Step S10 includes: Using ionospheric TEC data measured by a ground-based GNSS receiver located in a high-latitude region as ground-based measurement data, a time-series dataset of TEC data was established according to a certain time resolution, and S... TEC express: Official 1; in, Indicates t i The values ​​of ionospheric TEC data at time t, where t is time, t0 is the start time, and the subscript i is the time series number, N i Where Δt is the total number of data points, and Δt is the time resolution. S11 is a set of space weather indices, including the Dst index, the IMF-Bz index, and the F10.7 index. Step S11 includes: Time series datasets for the Dst index, IMF-Bz index, and F10.7 index were constructed according to a certain time resolution, and were analyzed using S... Dst S Bz and S F107 express: Official 2; Establish a time-series dataset of space weather indices ; Among them, Dst , , They represent t respectively i The values ​​of the Dst exponent, IMF-Bz exponent, and F10.7 exponent at time t, where t is time, t0 is the start time, and the subscript i is the time series number, N i Where Δt is the total number of data points, and Δt is the time resolution. TEC: Total Electron Content; Dst Index: Geomagnetic Storm Circulation Current Index; IMF-Bz Index: North-South Component Index of Interplanetary Magnetic Field; F10.7 Index: 10.7 cm Solar Radiocurrent Index. S2, Establish a time-series dataset of space-based high-energy particle sedimentation flux measurement data; S20: The satellite is in continuous motion and cannot make continuous measurements of a single point. For high-energy particle precipitation occurring in the auroral zone above magnetic latitude 60°, record all satellite measurements of the high-energy particle precipitation flux in the region. S3, normalizes all time-series datasets and establishes an input dataset based on multi-source heterogeneous data; S4. Establish an LSTM prediction neural network to obtain prediction results of high-latitude ionospheric parameters based on the input dataset.

2. The high-latitude ionospheric forecasting method based on multi-source heterogeneous data according to claim 1, characterized in that: Step S20 includes: Averaging all measurements in S1 using time resolution yields a time-series dataset of high-energy particle sedimentation flux measurements. P express: Official 3; Among them, P k This represents the measured value of high-energy particle sedimentation flux, where k is the measurement value number. This represents the magnetic latitude value corresponding to the high-energy particle precipitation flux. Indicates t i The value of high-energy particle sedimentation flux at time t, where t is time, t0 is the starting time, and the subscript i is the time sequence number, N i Δt represents the total number of data points, and Δt represents the time resolution. 条件 This is an indicator function that takes the value 1 when the condition is met, and 0 otherwise.

3. The high-latitude ionospheric forecasting method based on multi-source heterogeneous data according to claim 2, characterized in that: Step S3 includes: S31, introduces local time time series data, using This means that, similarly, normalization is performed on it to obtain a local time sine-normalized time series dataset. Local time cosine value normalized time series dataset : Official 4; S32, normalize the time-series datasets of ground-based ionospheric measurement data, space weather index sets, and space-based high-energy particle precipitation flux measurement data; obtain the normalized time-series dataset of TEC data. Normalized time series dataset of Dst exponent Normalized time series dataset of IMF-Bz index Normalized time series dataset of F10.7 exponent Normalized time series dataset of high-energy particle sedimentation flux measurement data : Official 5; S33, Five consecutive time series data points are slidably selected from the normalized time series dataset to establish neural network input data samples, which are then processed using... , , , , , and The input dataset is represented as , , , , , and : Formula 6.

4. The high-latitude ionospheric forecasting method based on multi-source heterogeneous data according to claim 3, characterized in that: The S4 step includes: S41, Build an LSTM prediction neural network. The input neurons use the input dataset, and the output layer has only one neuron. That is, in advance The ionospheric TEC value corresponding to the time forecast; S42, based on the trained prediction neural network, input samples from the five moments before the moment to be predicted are input to obtain the high-latitude ionospheric prediction results; S43 uses two parameters to evaluate forecast errors, including RMSE and PD: RMSE stands for Root Mean Square Error, and it is calculated as follows: Official 7; PD represents the total relative error, and is calculated as follows: Official 8; Where j is the sample number used for error statistics, and N j The total number of samples used for error statistics.

Images