Ionized layer irregular body identification method and system based on machine learning

Abstract

The invention discloses an ionized layer irregular body identification method and system based on machine learning. The method comprises the following steps: receiving a GNSS observation file and an ephemeris file, and preprocessing the GNSS observation file and the ephemeris file; extracting electron total content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics according to the preprocessed GNSS observation file and ephemeris file; and according to the extracted electron total content change rate index characteristic, the time integral, the total electron content index characteristic and the Doppler coefficient characteristic, carrying out ionized layer irregular body identification, and outputting an identification result. According to the method provided by the invention, in the process of identifying the ionosphere irregular body, the situation that whether an ionosphere irregular body event occurs or not is judged by setting a single threshold value is avoided, and data missing caused by cycle slip detection is effectively avoided, so that an EPB event and a TID event are effectively identified.

Description

Technical Field

[0001] The present invention relates to the technical field of GNSS and ionosphere, and in particular to a method and system for identifying ionospheric irregularities based on machine learning. Background Art

[0002] The ionosphere is a vital region in the Earth's atmosphere, located at an altitude of about 60 to 1,000 kilometers. The atmosphere in this region is ionized by solar radiation (mainly ultraviolet rays and X-rays) and cosmic rays, forming a large number of free electrons and positive ions, which makes the atmosphere in this region conductive. The existence of the ionosphere has an extremely important influence on the propagation of radio waves. It can refract, reflect, scatter and absorb radio waves, thus affecting the performance of radio information systems such as satellite navigation, communication, and radar.

[0003] Equatorial Plasma Bubbles (EPBs) are large-scale ionospheric irregular structures that occur at night and occur at the magnetic equator and low latitudes. These structures are generally believed to be generated by the bottom disturbance of the ionosphere F layer above the magnetic equator as a seed factor, under the action of Rayleigh-Taylor (RT) instability, and gradually rise to higher altitudes and map to low latitudes along magnetic field lines. The existence of equatorial plasma bubbles not only affects radio communication and navigation systems, but also has important significance for space weather research. They can cause ionospheric scintillation and affect the stability of communication and navigation systems. Therefore, the study of the occurrence law, morphological characteristics and evolution process of equatorial plasma bubbles has important scientific and application value for understanding and predicting space weather phenomena.

[0004] Traveling Ionospheric Disturbances (TIDs) are a periodic electron concentration fluctuation structure in the ionosphere, which is closely related to the movement of acoustic gravity waves in the upper atmosphere. TIDs have many effects on satellite communications by affecting the electron density distribution and signal propagation characteristics of the ionosphere, including changes in signal propagation paths, Doppler effects, ionospheric second-order effects, and signal interference. These effects may reduce communication quality, increase the complexity of signal processing, and in some cases interfere with satellite navigation. The study of TIDs is of great significance for understanding the dynamic changes of the ionosphere and space weather phenomena.

[0005] At present, there are two main problems in the identification method of ionospheric irregularities: first, in the process of equatorial plasma bubble detection, the disturbance is large and cycle slips are prone to occur. If a single threshold is set, it is easy to make it difficult to identify EPB (equatorial plasma bubble); second, it is difficult to identify equatorial plasma bubbles and traveling ionospheric disturbances using the ROTI coefficient alone, which leads to a relatively low identification efficiency of traveling ionospheric disturbances. Summary of the invention

[0006] In order to overcome the problem that the existing ionospheric irregular body recognition technology is difficult to identify EPB events due to cycle slip detection, and the distinction between EPB events and TID events is not high enough, the purpose of the present invention is to propose an ionospheric irregular body recognition method and system based on machine learning, avoid using a single threshold to determine whether an EPB event occurs, can effectively distinguish EPB events, and improve the accuracy of EPB event and TID event detection.

[0007] To achieve the purpose of the present invention, the present invention adopts the following technical solutions:

[0008] A method for identifying ionospheric irregularities based on machine learning, the method comprising the following steps:

[0009] Receiving GNSS observation files and ephemeris files, and preprocessing the GNSS observation files and ephemeris files;

[0010] Extract the total electron content change rate index characteristics, time integral and total electron content index characteristics, and Doppler coefficient characteristics based on the preprocessed GNSS observation files and ephemeris files;

[0011] Ionospheric irregularities are identified based on the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics, and Doppler coefficient characteristics, and the identification results are output.

[0012] In the above technical scheme, by preprocessing the received GNSS observation files and ephemeris files, the reliability and practicality of the data can be improved, and the efficiency of subsequent data processing can be improved; the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics can avoid using a single threshold to determine whether an EPB event has occurred during the process of ionospheric irregularity identification, and effectively avoid data loss caused by cycle slip detection, thereby effectively identifying EPB events and improving the accuracy of EPB event and TID event detection.

[0013] Furthermore, the process of preprocessing the GNSS observation file and the ephemeris file includes:

[0014] Perform gross error detection and abnormal data elimination on the received GNSS observation files;

[0015] Perform data integrity check, data error analysis and data correction processing on the received ephemeris files.

[0016] In the above technical scheme, the GNSS observation files are subjected to gross error detection and abnormal data elimination, and the ephemeris files are subjected to data integrity check, data error analysis and data correction. Finally, the GNSS observation files and ephemeris files are subjected to noise reduction, which can improve the reliability and practicality of the data and improve the efficiency of subsequent data processing.

[0017] Furthermore, the process of extracting the total electron content change rate index feature according to the preprocessed GNSS observation file includes:

[0018] Extract total electron content data based on the preprocessed GNSS observation files;

[0019] Extracting the change rate characteristics of the total electron content according to the extracted total electron content data;

[0020] The total electron content change rate index feature is extracted based on the change rate feature of the total electron content.

[0021] Furthermore, the time derivative between two consecutive time periods is calculated based on the extracted total electron content data to extract the change rate characteristics of the total electron content. The expression is:

[0022]

[0023] The expression for extracting the index characteristic of the total electron content change rate based on the change rate characteristic of the total electron content is:

[0024]

[0025] Among them, t i represents the timestamp, k represents the satellite serial number, STEC represents the original calculated total electron content data, and <> represents the average value within a time interval.

[0026] In the above technical scheme, the total electron content data is extracted according to the preprocessed GNSS observation file. The total electron content data can be used to monitor typical indicators of ionospheric irregularities within a short time window, thereby extracting the time derivative between two consecutive time periods to extract the change rate characteristics of the total electron content.

[0027] Furthermore, the process of extracting the time integral and total electron content index features according to the preprocessed GNSS observation file includes:

[0028] Extract the detrended total electron content based on the preprocessed GNSS observation files;

[0029] Extracting the integral of the detrended total electron content data in a fixed time window based on the detrended total electron content;

[0030] The extracted total electron content change rate index feature is combined with the integral of the detrended total electron content data to obtain the time integral and total electron content index features.

[0031] Furthermore, the moving average method is used to extract the total electron content without trend, and the expression is:

[0032]

[0033] SG filtering is used to extract the background value of the total electron content data. The expression is:

[0034] DTEC=MF×(STEC-SGTEC)

[0035] Based on the extracted detrended total electron content and the background value of the total electron content data, the integral of the detrended total electron content data in a fixed time window is extracted, and the expression is:

[0036]

[0037] The total electron content change rate index characteristic is combined with the integral of the total electron content data without trend to obtain the time integral and total electron content index characteristics, which are expressed as:

[0038] IROTI=IDTEC·ROTI

[0039] Where N is a natural number, MF is the satellite altitude angle, STEC is the total electron content data obtained by original calculation, SGTEC is the background total electron content data obtained after SG filtering and fitting, |DTEC(i)| is the absolute value of DTEC data at time point t, Δt is the time interval of DTEC data, n is the number of DTEC data in the time window, the unit of IDTEC value is TECU, and the unit of IROTI is generally expressed as TECU. 2 .

[0040] In the above technical scheme, based on the extracted total electron content without trend and the background value of the total electron content data, the integral of the total electron content data without trend in a fixed time window is extracted, and then the total electron content change rate index characteristic is combined with the integral of the total electron content data without trend to obtain the time integral and total electron content index characteristics, which can effectively maintain the communication quality and reduce the complexity of signal processing and the interference to satellite navigation in the process of ionospheric traveling disturbance detection, so as to improve the accuracy of TID detection.

[0041] Furthermore, the process of extracting the Doppler coefficient feature according to the preprocessed ephemeris file includes:

[0042] Calculate the Doppler measurement at two frequencies using the expression:

[0043]

[0044] According to the Doppler measurement values ​​at two frequencies, the differential frequency ΔD of the Doppler measurement value is calculated, and the expression is:

[0045]

[0046] Based on the differential frequency ΔD of the Doppler measurement value, the sliding window method is used to calculate the Doppler coefficient characteristics of ΔD in a specific time period. The expression is:

[0047]

[0048] Where i represents the frequency number, D i represents the Doppler measurement value, λ i represents the wavelength, represents the geometric distance rate between the satellite and the receiver, c represents the speed of light, and are the clock errors of the satellite and receiver respectively, represents the ionospheric delay rate, represents the tropospheric delay rate, ε i Indicates noise.

[0049] Furthermore, the process of identifying ionospheric irregularities according to the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics includes:

[0050] A machine learning model is set up, and the total electron content change rate index feature, time integral and total electron content index feature and Doppler coefficient feature are used to iteratively optimize the machine learning model. The process is as follows:

[0051] Based on the total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics, a decision tree is iteratively constructed, and the expression is:

[0052]

[0053] The regularization method is used to construct the objective function to iteratively optimize the machine learning model. The expression is:

[0054]

[0055] Set the loss function to optimize the parameters of the machine learning model during the iterative optimization process. When the round of iterative optimization training ends or the loss function converges, the trained machine learning model is obtained.

[0056] The trained machine learning model is used to identify ionospheric irregularities in the GNSS observation files and ephemeris files to be identified. The expression is:

[0057]

[0058] Design a binary classifier to classify the recognition results and convert the output of the logic function into probability. The expression is:

[0059]

[0060] The identification results include equatorial plasma bubble events, ionospheric traveling disturbance events and quiet periods, K t represents the number of leaf nodes in the tth tree, represents the weight of the kth leaf node, represents an indicator function, x represents sample data, Represents the region of the kth leaf node in the tth tree n represents the number of samples, y i represents the true label of the i-th sample, represents the model's predicted value for the i-th sample, Ω(f t ) represents the regularization term.

[0061] Furthermore, in the process of iteratively optimizing the machine learning model by constructing the objective function using the regularization method, the number of leaf nodes in the tree and the sum of the squares of the leaf node weights are calculated through the regularization term, and the expression is:

[0062]

[0063] Among them, γ and λ are regularization parameters.

[0064] In the above technical solution, the use of thresholds to determine whether an ionospheric irregularity event occurs is avoided, and the total electron content change rate index characteristics, time integral and total electron content index characteristics, and Doppler coefficient characteristics are used to perform iterative parameter optimization training on the set machine learning model, so that the machine learning model can effectively grasp the data changes in the GNSS observation file and the ephemeris file, thereby avoiding the use of a single threshold to determine whether an EPB event occurs during the ionospheric irregularity identification process, and effectively avoiding data loss caused by cycle slip detection, thereby effectively identifying EPB events and, at the same time, improving the accuracy of EPB event and TID event detection.

[0065] An ionospheric irregularity recognition system based on machine learning, the system comprising:

[0066] A data receiving module, used for receiving GNSS observation files and ephemeris files;

[0067] A data processing module, used for preprocessing the GNSS observation file and the ephemeris file;

[0068] A feature extraction module is used to extract the total electron content change rate index feature, the time integral and total electron content index feature, and the Doppler coefficient feature according to the preprocessed GNSS observation file and ephemeris file;

[0069] The ionospheric irregular body recognition module is used to recognize ionospheric irregular bodies according to the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics, and output the recognition results.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] The present invention proposes an ionospheric irregular body recognition method and system based on machine learning. By preprocessing received GNSS observation files and ephemeris files, the reliability and practicability of data can be improved, and the efficiency of subsequent data processing can be improved. The extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics avoid using a single threshold to judge whether an EPB event occurs, and effectively avoid data loss caused by cycle slip detection, thereby effectively identifying EPB events and, at the same time, improving the accuracy of EPB event and TID event detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 A flowchart of a method for identifying ionospheric irregularities based on machine learning provided in an embodiment of the present application;

[0073] Figure 2 A flowchart of the machine learning model training provided in the embodiment of the present application;

[0074] Figure 3 A schematic diagram of the TEC file in the experiment provided in the embodiment of the present application and the result predicted based on the machine learning model;

[0075] Figure 4 A schematic diagram of a confusion matrix of a test set during the training of a machine learning model in an experiment provided in an embodiment of the present application;

[0076] Figure 5 A schematic diagram of the TEC sequence and ROTI sequence in the experiment provided in the examples of this application;

[0077] Figure 6 A second schematic diagram of the predicted samples and confusion matrix of the test set during the machine learning model training process in the experiment provided in the embodiment of the present application;

[0078] Figure 7 A structural schematic diagram of an ionospheric irregularity recognition system based on machine learning provided in an embodiment of the present application. DETAILED DESCRIPTION

[0079] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Preferred embodiments of the present invention are provided in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present invention more thorough and comprehensive.

[0080] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0081] Embodiment 1:

[0082] This embodiment provides a method for identifying ionospheric irregularities based on machine learning. Figure 1 , the method comprises the following steps:

[0083] Step S1: The receiver receives a GNSS observation file and an ephemeris file, and pre-processes the GNSS observation file and the ephemeris file;

[0084] Step S2: extracting the total electron content change rate index feature, the time integral and total electron content index feature, and the Doppler coefficient feature according to the preprocessed GNSS observation file and ephemeris file;

[0085] Step S3: Identify ionospheric irregularities based on the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics, and Doppler coefficient characteristics, and output the identification results.

[0086] Among them, the total electron content change rate index characteristic is ROTI (Rate of TEC Index), the time integral and total electron content index characteristic is IROTI (the time Integral and Rate Of the Total electron content Index), and the Doppler coefficient characteristic is DI (Doppler Index)

[0087] In step S1, the process of preprocessing the GNSS observation file and the ephemeris file includes:

[0088] Perform gross error detection and abnormal data elimination on the received GNSS observation files;

[0089] Perform data integrity check, data error analysis and data correction processing on the received ephemeris files.

[0090] It can be understood that performing gross error detection and abnormal data elimination on GNSS observation files, performing data integrity check, data error analysis and data correction on ephemeris files, and finally performing noise reduction on GNSS observation files and ephemeris files can improve the reliability and practicality of the data and improve the efficiency of subsequent data processing.

[0091] In step S2, the process of extracting the total electron content change rate index feature according to the preprocessed GNSS observation file includes:

[0092] Extract total electron content data (TEC data, Total Electron Content) based on the preprocessed GNSS observation files;

[0093] Extracting the change rate characteristics of the total electron content according to the extracted total electron content data;

[0094] The total electron content change rate index feature is extracted based on the change rate feature of the total electron content.

[0095] Specifically, the time derivative between two consecutive time periods is calculated based on the extracted total electron content data to extract the change rate characteristics of the total electron content. The expression is:

[0096]

[0097] The expression for extracting the index characteristic of the total electron content change rate based on the change rate characteristic of the total electron content is:

[0098]

[0099] Among them, t i represents the timestamp, k represents the satellite serial number, STEC represents the original calculated total electron content data, and <> represents the average value within a time interval.

[0100] It can be understood that by extracting the total electron content data based on the preprocessed GNSS observation file, the total electron content data can be used to monitor typical indicators of ionospheric irregularities within a short time window, thereby extracting the time derivative between two consecutive time periods to extract the rate of change characteristics of the total electron content, thereby improving the accuracy of TID detection in the process of ionospheric traveling disturbance detection.

[0101] In step S2, the process of extracting the time integral and total electron content index features according to the preprocessed GNSS observation file includes:

[0102] Extract the detrended total electron content based on the preprocessed GNSS observation files;

[0103] Extracting the integral of the detrended total electron content data in a fixed time window based on the detrended total electron content;

[0104] The extracted total electron content change rate index feature is combined with the integral of the detrended total electron content data to obtain the time integral and total electron content index features.

[0105] Specifically, the moving average method is used to extract the total electron content without trend, and the expression is:

[0106]

[0107] SG filtering is used to extract the background value of the total electron content data. The expression is:

[0108] DTEC=MF×(STEC-SGTEC)

[0109] Based on the extracted detrended total electron content and the background value of the total electron content data, the integral of the detrended total electron content data in a fixed time window is extracted, and the expression is:

[0110]

[0111] The total electron content change rate index characteristic is combined with the integral of the total electron content data without trend to obtain the time integral and total electron content index characteristics, which are expressed as:

[0112] IROTI=IDTEC·ROTI

[0113] Where N is a natural number, MF is the satellite altitude angle, STEC is the total electron content data obtained by original calculation, SGTEC is the background total electron content data obtained after SG filtering and fitting, |DTEC(i)| is the absolute value of DTEC data at time point t, Δt is the time interval of DTEC data, n is the number of DTEC data in the time window, the unit of IDTEC value is TECU, and the unit of IROTI is generally expressed as TECU. 2 , where 1TECU = 10 16 Electrons / m2.

[0114] It can be understood that based on the extracted total electron content without trend and the background value of the total electron content data, the integral of the total electron content data without trend in a fixed time window is extracted, and then the total electron content change rate index characteristic is combined with the integral of the total electron content data without trend to obtain the time integral and total electron content index characteristics.

[0115] In step S2, the process of extracting Doppler coefficient features according to the preprocessed ephemeris file includes:

[0116] Calculate the Doppler measurement at two frequencies using the expression:

[0117]

[0118] According to the Doppler measurement values ​​at two frequencies, the differential frequency ΔD of the Doppler measurement value is calculated, and the expression is:

[0119]

[0120] Based on the differential frequency ΔD of the Doppler measurement value, the sliding window method is used to calculate the Doppler coefficient characteristics of ΔD in a specific time period. The expression is:

[0121]

[0122] Where i represents the frequency number, D i represents the Doppler measurement value, λ i represents the wavelength, represents the geometric distance rate between the satellite and the receiver, c represents the speed of light, and are the clock errors of the satellite and receiver respectively, represents the ionospheric delay rate, represents the tropospheric delay rate, ε i Indicates noise.

[0123] In step S3, see Figure 2, avoid using thresholds to judge whether an ionospheric irregularity event occurs, and the process of identifying ionospheric irregularities based on the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics, and Doppler coefficient characteristics includes:

[0124] A machine learning model is set up, and the total electron content change rate index feature, time integral and total electron content index feature and Doppler coefficient feature are used to iteratively optimize the machine learning model. The process is as follows:

[0125] Based on the total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics, a decision tree is iteratively constructed, and the expression is:

[0126]

[0127] The regularization method is used to construct the objective function to iteratively optimize the machine learning model. The expression is:

[0128]

[0129] Set the loss function to optimize the parameters of the machine learning model during the iterative optimization process. When the round of iterative optimization training ends or the loss function converges, the trained machine learning model is obtained.

[0130] The trained machine learning model is used to identify ionospheric irregularities in the GNSS observation files and ephemeris files to be identified. The expression is:

[0131]

[0132] Design a binary classifier to classify the recognition results and convert the output of the logic function into probability. The expression is:

[0133]

[0134] The identification results include equatorial plasma bubble events, ionospheric traveling disturbance events and quiet periods, K t represents the number of leaf nodes in the tth tree, represents the weight of the kth leaf node, represents an indicator function, x represents sample data, Represents the region of the kth leaf node in the tth tree n represents the number of samples, y i represents the true label of the i-th sample, represents the model's predicted value for the i-th sample, Ω(f t ) represents the regularization term.

[0135] As a preferred embodiment, in the process of iteratively optimizing the machine learning model by constructing the objective function using the regularization method, the number of leaf nodes of the tree and the sum of the squares of the leaf node weights are calculated by the regularization term, and the expression is:

[0136]

[0137] Among them, γ and λ are regularization parameters.

[0138] In this embodiment, the machine learning model adopts the XGBoost (eXtreme Gradient Boosting) classification model, and XGBoost improves the prediction performance of the model by constructing multiple decision trees.

[0139] It can be understood that the total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics are used to perform parameter iterative optimization training on the set machine learning model, so that the machine learning model can effectively grasp the data changes in the GNSS observation files and ephemeris files, thereby avoiding the use of a single threshold to determine whether an EPB event occurs in the process of identifying ionospheric irregularities, and effectively avoiding data loss caused by cycle slip detection, thereby effectively identifying EPB events and, at the same time, improving the accuracy of EPB event and TID event detection.

[0140] In this embodiment, by preprocessing the received GNSS observation files and ephemeris files, the reliability and practicality of the data can be improved, and the efficiency of subsequent data processing can be improved; the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics can avoid using a single threshold to determine whether an EPB event has occurred during the process of ionospheric irregularity identification, and effectively avoid data loss caused by cycle slip detection, thereby effectively identifying EPB events and improving the accuracy of EPB event and TID event detection.

[0141] Embodiment 2:

[0142] This example is based on the method described in Example 1, and provides experimental data corresponding to , which are as follows:

[0143] Case 1:

[0144] The data was received on March 3, 2024, and the satellite was R22 (GLONASS).

[0145] from Figure 3From the TEC image, we can see that due to the cycle slip detection, a large amount of data from 12:30 to 13:30 was identified as a cycle slip and removed, so the data for that period was missing and could not be identified when calculating ROTI. Since multiple coefficients were introduced instead of a single coefficient, the XGBoost model will not fail to identify the ROTI data, and can still identify this EPB event.

[0146] from Figure 3 It can be seen that the time period of recognition corresponds to the TEC in the figure above, and the time period of occurrence is around 20:00-22:00 (night) local time, which is consistent with the occurrence time of the EPB event. Before and after the data disappears, it can be clearly seen that it is consuming downwards, which is consistent with the characteristics of the EPB event. In summary, this is an EPB event, which proves the reliability of the XGBoost recognition model.

[0147] from Figure 4 From the confusion matrix in , we can see that a total of 760 epoch data were identified, of which 565 were non-EPB events (quiet periods) and 195 were EPB events. This is also in line with the actual situation.

[0148] Case 2:

[0149] The data was received on July 17, 2014, and the satellite was G05 (GPS).

[0150] from Figure 5 It can be seen from the TEC sequence in that a fluctuation occurred around UTC14:00, which is consistent with the performance of TID in the TEC sequence. From the time period and duration of occurrence, it is consistent with the characteristics of MSTID (mesoscale traveling ionospheric disturbance). However, if only the ROTI sequence is used as a judgment, it is difficult to set a good threshold to distinguish TID, because the ROTI value in a calm period may also reach about 0.2. Therefore, the IROTI index is used in the XGBoost model, so that the recognition model is able to distinguish TID.

[0151] See also Figure 6 ,From the current recognition effect, the XGBoost recognition model of the present application has a strong ,ability to recognize equatorial plasma bubble (EPB) events, and can also ,well recognize traveling ionospheric disturbance (TID) events.

[0152] Embodiment three:

[0153] This embodiment provides an ionospheric irregular body recognition system based on machine learning, see Figure 7 , the system comprising:

[0154] A data receiving module, used for receiving GNSS observation files and ephemeris files;

[0155] A data processing module, used for preprocessing the GNSS observation file and the ephemeris file;

[0156] A feature extraction module is used to extract the total electron content change rate index feature, the time integral and total electron content index feature, and the Doppler coefficient feature according to the preprocessed GNSS observation file and ephemeris file;

[0157] The ionospheric irregular body recognition module is used to recognize ionospheric irregular bodies according to the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics, and output the recognition results.

[0158] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for identifying ionospheric irregularities based on machine learning, characterized in that: The method comprises the following steps: Receiving GNSS observation files and ephemeris files, and preprocessing the GNSS observation files and ephemeris files; Extract the total electron content change rate index characteristics, time integral and total electron content index characteristics, and Doppler coefficient characteristics based on the preprocessed GNSS observation files and ephemeris files; Ionospheric irregularities are identified based on the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics, and Doppler coefficient characteristics, and the identification results are output.

2. The ionospheric irregularity recognition method based on machine learning according to claim 1, characterized in that: The process of preprocessing the GNSS observation file and the ephemeris file includes: Perform gross error detection and abnormal data elimination on the received GNSS observation files; Perform data integrity check, data error analysis and data correction processing on the received ephemeris files.

3. The ionospheric irregularity recognition method based on machine learning according to claim 1, characterized in that: The process of extracting the total electron content change rate index characteristics based on the preprocessed GNSS observation file includes: Extract total electron content data based on the preprocessed GNSS observation files; Extracting the change rate characteristics of the total electron content according to the extracted total electron content data; The total electron content change rate index feature is extracted based on the change rate feature of the total electron content.

4. The ionospheric irregularity recognition method based on machine learning according to claim 3, characterized in that: The time derivative between two consecutive time periods is calculated based on the extracted total electron content data to extract the change rate characteristics of the total electron content. The expression is: The expression for extracting the index characteristic of the total electron content change rate based on the change rate characteristic of the total electron content is: Among them, t i represents the timestamp, k represents the satellite serial number, STEC represents the original calculated total electron content data, and <> represents the average value within a time interval.

5. The ionospheric irregularity recognition method based on machine learning according to claim 3, characterized in that: The process of extracting time integral and total electron content index features from the preprocessed GNSS observation files includes: Extract the detrended total electron content based on the preprocessed GNSS observation files; Extracting the integral of the detrended total electron content data in a fixed time window based on the detrended total electron content; The extracted total electron content change rate index feature is combined with the integral of the detrended total electron content data to obtain the time integral and total electron content index features.

6. The method for identifying ionospheric irregularities based on machine learning according to claim 5, characterized in that: The moving average method is used to extract the total electron content to remove the trend, and the expression is: SG filtering is used to extract the background value of the total electron content data. The expression is: DTEC=MF×(STEC-SGTEC) Based on the extracted detrended total electron content and the background value of the total electron content data, the integral of the detrended total electron content data in a fixed time window is extracted, and the expression is: The total electron content change rate index characteristic is combined with the integral of the total electron content data without trend to obtain the time integral and total electron content index characteristics, which are expressed as: IROTI=IDTEC·ROTI Where N is a natural number, MF is the satellite altitude angle, STEC is the total electron content data obtained by original calculation, SGTEC is the background total electron content data obtained after SG filtering and fitting, |DTEC(i)| is the absolute value of DTEC data at time point t, Δt is the time interval of DTEC data, n is the number of DTEC data in the time window, the unit of IDTEC value is TECU, and the unit of IROTI is generally expressed as TECU. 2 .

7. The ionospheric irregularity recognition method based on machine learning according to claim 1, characterized in that: The process of extracting Doppler coefficient features from the preprocessed ephemeris file includes: Calculate the Doppler measurement at two frequencies using the expression: According to the Doppler measurement values ​​at two frequencies, the differential frequency ΔD of the Doppler measurement value is calculated, and the expression is: Based on the differential frequency ΔD of the Doppler measurement value, the sliding window method is used to calculate the Doppler coefficient characteristics of ΔD in a specific time period. The expression is: Where i represents the frequency number, D i represents the Doppler measurement value, λ i represents the wavelength, represents the geometric distance rate between the satellite and the receiver, c represents the speed of light, and are the clock errors of the satellite and receiver respectively, represents the ionospheric delay rate, represents the tropospheric delay rate, ε i Indicates noise.

8. The method for identifying ionospheric irregularities based on machine learning according to claim 7, characterized in that: The process of identifying ionospheric irregularities based on the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics includes: A machine learning model is set up, and the total electron content change rate index feature, time integral and total electron content index feature and Doppler coefficient feature are used to iteratively optimize the machine learning model. The process is as follows: Based on the total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics, a decision tree is iteratively constructed, and the expression is: The regularization method is used to construct the objective function to iteratively optimize the machine learning model. The expression is: Set the loss function to optimize the parameters of the machine learning model during the iterative optimization process. When the round of iterative optimization training ends or the loss function converges, the trained machine learning model is obtained. The trained machine learning model is used to identify ionospheric irregularities in the GNSS observation files and ephemeris files to be identified. The expression is: Design a binary classifier to classify the recognition results and convert the output of the logic function into probability. The expression is: The identification results include equatorial plasma bubble events, ionospheric traveling disturbance events and quiet periods, K t represents the number of leaf nodes in the tth tree, represents the weight of the kth leaf node, represents an indicator function, x represents sample data, Represents the region of the kth leaf node in the tth tree n represents the number of samples, y i represents the true label of the i-th sample, represents the model's predicted value for the i-th sample, Ω(f t ) represents the regularization term.

9. The method for identifying ionospheric irregularities based on machine learning according to claim 8, characterized in that: In the process of iteratively optimizing the machine learning model by constructing the objective function using the regularization method, the number of leaf nodes in the tree and the sum of the squares of the leaf node weights are calculated through the regularization term. The expression is: Among them, γ and λ are regularization parameters.

10. An ionospheric irregularity recognition system based on machine learning, characterized in that: The system comprises: A data receiving module, used for receiving GNSS observation files and ephemeris files; A data processing module, used for preprocessing the GNSS observation file and the ephemeris file; A feature extraction module is used to extract the total electron content change rate index feature, time integral and total electron content index feature according to the preprocessed GNSS observation file, and to extract the Doppler coefficient feature according to the preprocessed ephemeris file; The ionospheric irregular body recognition module is used to recognize ionospheric irregular bodies according to the extracted total electron content change rate index characteristics, time integral and total electron content index characteristics and Doppler coefficient characteristics, and output the recognition results.

Images