Ionized layer disturbance level prediction method and device, equipment and medium

Abstract

The embodiment of the invention discloses an ionosphere disturbance level prediction method and device, equipment and a medium. The method comprises the steps that a first ionosphere observation value observed by high-frequency equipment and a second ionosphere observation value observed by a common GNSS receiver within a preset time period are acquired; calculating the ROTI value of each region of the ionized layer at the target moment by using an ROTI index judgment method; grading the ROTI value of each region of the ionosphere at the target moment, marking a second ionosphere observation value at the previous moment according to the disturbance grade of each region of the ionosphere at the target moment, generating a training sample set for model training, and obtaining a disturbance grade prediction model. The disturbance level prediction model is trained by acquiring high-frequency observation data and common observation data of a long-time sequence, so that the disturbance level prediction model can predict the disturbance level of the ionosphere at the next moment in real time according to low-frequency observation data; and the disturbance level prediction model can still output a high-precision prediction result when low-frequency observation data is input.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ionosphere monitoring, and in particular to an ionosphere disturbance level prediction method and device, equipment and a medium. BACKGROUND

[0002] In modern communication and navigation technology, the ionosphere plays a key role, and its electron density changes directly affect the propagation of radio waves, and have a significant effect on the signal quality and positioning accuracy of satellite navigation systems such as the Global Positioning System (GPS).

[0003] Ionospheric disturbances, such as traveling ionospheric disturbances and equatorial ionospheric bubbles, can cause abnormal electron density distribution and interfere with radio communication and navigation systems. Existing ionosphere monitoring technology has obvious limitations: it relies on high-cost ground stations and high-frequency receivers, making it difficult to achieve continuous coverage worldwide, especially in remote areas such as the ocean; the early warning system is slow to respond, making it difficult to meet real-time requirements, and there is a long time delay. SUMMARY

[0004] Therefore, it is necessary to propose an ionosphere disturbance level prediction method and device, equipment and a medium to realize real-time ionosphere disturbance level prediction and reduce ionosphere monitoring costs.

[0005] To achieve the above purpose, the first aspect of the present application provides an ionosphere disturbance level prediction method, which comprises:

[0006] Obtaining first observation data observed by a high-frequency device and second observation data observed by a general GNSS receiver at each time in a preset time period;

[0007] Extracting first ionosphere observation values observed by the high-frequency device and second ionosphere observation values observed by the general GNSS receiver at a target time from the first observation data and the second observation data at the target time, respectively, wherein the target time is any time in the preset time period, and the ionosphere observation values include observation values of each region of the ionosphere at a certain time;

[0008] Using the ROTI index determination method to calculate the ROTI values of each region of the ionosphere at the target time according to the first ionosphere observation values observed by the high-frequency device;

[0009] According to a preset level division standard, the ROTI values of each region of the ionosphere at the target time are divided into levels to obtain the disturbance levels of each region of the ionosphere at the target time;

[0010] The disturbance level of each region of the ionosphere at the target moment is marked on the second ionosphere observation value at the previous moment, to generate a training sample set, wherein the previous moment is the previous moment of the target moment;

[0011] The training sample set is used for model training to obtain a disturbance level prediction model, so as to output the disturbance level of the ionosphere at the next moment based on the ionosphere observation data observed by the ordinary GNSS receiver at the current moment.

[0012] Further, the first ionosphere observation value observed by the high-frequency device and the second ionosphere observation value observed by the ordinary GNSS receiver at the target moment are extracted from the first observation data and the second observation data at the target moment, specifically including:

[0013] Obtaining a first observation equation of a preset phase and pseudo-range, and a target epoch corresponding to the target moment;

[0014] Determining an initial epoch range according to the target epoch and a preset arc segment size, and the target epoch is in the initial epoch range;

[0015] Performing ambiguity mean value calculation according to the observation data at each epoch in the initial epoch range to obtain an optimized combined ambiguity;

[0016] Updating the combined ambiguity to the first observation equation to obtain an updated second observation equation;

[0017] Substituting the first observation data and the second observation data at the target moment into the second observation equation to solve, to obtain the first ionosphere observation value observed by the high-frequency device and the second ionosphere observation value observed by the ordinary GNSS receiver at the target moment.

[0018] Further, the ionosphere observation value at least includes a TEC value, and the ambiguity mean value calculation according to the observation data at each epoch in the initial epoch range to obtain the optimized combined ambiguity specifically includes:

[0019] Obtaining observation data of a previous epoch, and the previous epoch is the previous epoch of the target epoch;

[0020] Calculating the initial TEC value at the target epoch and the initial TEC value of the previous epoch according to the observation data of the target epoch and the previous epoch;

[0021] Calculating a change rate according to the initial TEC value at the target epoch and the initial TEC value of the previous epoch to obtain the TEC change rate at the target epoch.

[0022] analyzing the TEC rate at the target epoch and a preset arc length adjustment rule to obtain an updated target epoch range, wherein the arc length adjustment rule comprises a corresponding relationship between a rate size and an epoch range size, and the target epoch is in the target epoch range;

[0023] performing ambiguity mean calculation according to observation data at each epoch in the target epoch range to obtain an optimized combined ambiguity.

[0024] Further, the analyzing the TEC rate at the target epoch and a preset arc length adjustment rule to obtain an updated target epoch range specifically comprises:

[0025] analyzing the TEC rate at the target epoch and a preset arc length adjustment rule to obtain an adjusted epoch range at the target epoch;

[0026] obtaining an adjusted epoch range at the previous epoch;

[0027] performing exponential weighted average calculation according to the adjusted epoch range at the target epoch and the adjusted epoch range at the previous epoch to obtain a target epoch range at the target epoch;

[0028] wherein the arc length adjustment rule is expressed as follows:

[0029]

[0030] In the formula, M(k) is the adjusted epoch range at the target epoch, M min and M max are preset minimum and maximum epoch ranges respectively, ΔI(k) is the TEC rate at the target epoch, and θ I is a preset rate threshold.

[0031] Further, the optimized combined ambiguity is calculated by the following formula:

[0032]

[0033] In the formula, N is the number of satellites, M is the optimized combined ambiguity at the target epoch, M smoothed (k) is the target epoch range at the target epoch, is a double-frequency pseudorange geometry-independent combined observation value of a receiver r to a satellite s at a jth epoch, is a double-frequency carrier phase geometry-independent combined observation value of the receiver r to the satellite s at the jth epoch, A pseudorange hardware delay term for the receiver r and the satellite s.

[0034] Further, the ionospheric observation value at least includes a TEC value, a ROTI, receiver information, satellite information and a piercing point information;

[0035] The disturbance level of each region of the ionosphere at the target moment is marked on the second ionospheric observation value at the previous moment to generate a training sample set, specifically including:

[0036] The TEC value, the ROTI, the receiver information, the satellite information and the piercing point information of the target region of the ionosphere at the previous moment are sequentially arranged to obtain a feature sequence of the target region at the previous moment, wherein the target region is any region in all regions of the ionosphere;

[0037] The feature sequence of the target region at the previous moment is marked according to the disturbance level of the target region at the target moment to generate a training sample corresponding to the target region at the previous moment;

[0038] The training samples corresponding to all regions of the ionosphere at all moments in the preset time period constitute a training sample set.

[0039] Further, the training sample set is used for model training to obtain a disturbance level prediction model, specifically including:

[0040] The training sample set is input into a preset multi-scale spatio-temporal Transformer model for training to generate a disturbance level prediction model.

[0041] To achieve the above object, the second aspect of the present application provides a device for predicting the disturbance level of the ionosphere, which comprises a data acquisition module, a data processing module and a model training module:

[0042] The data acquisition module is used for acquiring first observation data observed by a high-frequency device and second observation data observed by a general GNSS receiver at each moment in a preset time period;

[0043] The data processing module is used for extracting a first ionospheric observation value observed by the high-frequency device and a second ionospheric observation value observed by the general GNSS receiver at a target moment from the first observation data and the second observation data at the target moment, respectively, wherein the target moment is any moment in the preset time period, and the ionospheric observation value contains observation values of each region of the ionosphere at a certain moment;

[0044] According to the first ionosphere observation value observed by the high-frequency equipment, the ROTI values of each region of the ionosphere at the target moment are calculated by using the ROTI index determination method;

[0045] According to the preset grade division standard, the ROTI values of each region of the ionosphere at the target moment are graded to obtain the disturbance grade of each region of the ionosphere at the target moment.

[0046] The model training module is configured to label the disturbance grade of each region of the ionosphere at the target moment on the second ionosphere observation value at the previous moment to generate a training sample set, wherein the previous moment is the previous moment of the target moment.

[0047] The model training module is configured to label the disturbance grade of each region of the ionosphere at the target moment on the second ionosphere observation value at the previous moment to generate a training sample set, wherein the previous moment is the previous moment of the target moment.

[0048] To achieve the above object, the third aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to make the processor execute the steps of the method according to the first aspect.

[0049] To achieve the above object, the fourth aspect of the present application provides a computer device, which comprises a memory and a processor, and the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the method according to the first aspect.

[0050] By adopting the embodiment of the present application, the following beneficial effects are achieved:

[0051] The embodiment of the present application provides a method for predicting the ionospheric disturbance level, which comprises the following steps: obtaining first observation data observed by a high-frequency device and second observation data observed by a general GNSS receiver at each moment in a preset time period; extracting first ionospheric observation values observed by the high-frequency device and second ionospheric observation values observed by the general GNSS receiver at a target moment from the first observation data and the second observation data at the target moment, wherein the target moment is any moment in the preset time period, and the ionospheric observation values comprise observation values of each region of the ionosphere at a certain moment; calculating ROTI values of each region of the ionosphere at the target moment according to the first ionospheric observation values observed by the high-frequency device by using a ROTI index determination method; performing level division on the ROTI values of each region of the ionosphere at the target moment according to a preset level division standard to obtain disturbance levels of each region of the ionosphere at the target moment; labeling the disturbance levels of each region of the ionosphere at the target moment on the second ionospheric observation values at a previous moment to generate a training sample set, wherein the previous moment is a moment before the target moment; performing model training by using the training sample set to obtain a disturbance level prediction model, so as to output disturbance levels of the ionosphere at a next moment according to current ionospheric observation data observed by the general GNSS receiver based on the disturbance level prediction model. The disturbance level prediction model is trained by using long-time sequence high-frequency observation data and general observation data, so that the disturbance level prediction model can perform real-time prediction on the ionospheric disturbance levels at the next moment according to low-frequency observation data, and the disturbance level prediction model can still ensure high-precision prediction results when inputting low-frequency observation data, thereby reducing the monitoring cost of the ionosphere. BRIEF DESCRIPTION OF DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, brief introductions will be given to the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0053] In the formula, the following applies:

[0054] Figure 1 The flowchart of the method for predicting the ionospheric disturbance level of the embodiment of the present application is shown in the figure.

[0055] Figure 2 The structure block diagram of the device for predicting the ionospheric disturbance level of the embodiment of the present application is shown in the figure.

[0056] Figure 3 The internal structure diagram of the computer device in the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0058] The conventional ionospheric disturbance observation method usually uses observation data of a single receiver and satellite pair for observation, resulting in low data utilization, difficulty in comprehensively describing regional ionospheric state, and inability to make real-time prediction on ionospheric disturbance.

[0059] Based on this, an embodiment of the present application provides a method for predicting ionospheric disturbance level, please refer to Figure 1 , Figure 1 The flowchart of the method for predicting ionospheric disturbance level in the embodiment of the present application is shown in the figure, and the method comprises the following steps:

[0060] Step 100, obtaining first observation data observed by a high-frequency device and second observation data observed by a general GNSS receiver at each time in a preset time period.

[0061] In the embodiment of the present application, the preset time period can be a long time sequence period, such as obtaining the first observation data observed by the high-frequency device and the second observation data observed by the general GNSS receiver at each time from 2010 to 2023. The model trained based on the long time sequence observation data can effectively reflect the difference and connection between the observation data of the high-frequency device and the general GNSS receiver.

[0062] The observation data observed by the high-frequency device and the general GNSS receiver are both formed after receiving the signals sent by the satellite, and the difference lies in that the time resolution of the high-frequency device is higher. The working frequency band of the general GNSS receiver is mostly in the L band (1-2 GHz), while the high-frequency device covers a wider high-frequency range, such as HF (3-30 MHz), VHF / UHF (more than 30 MHz) and even microwave. Therefore, the data observed by the high-frequency device has stronger anti-interference ability and higher data propagation rate.

[0063] Step 200, extracting first ionospheric observation values observed by the high-frequency device and second ionospheric observation values observed by the general GNSS receiver at the target time from the first observation data and the second observation data at the target time, respectively, wherein the target time is any time in the preset time period, and the ionospheric observation values contain observation values of each region of the ionosphere at a certain time.

[0064] In the embodiment of the present application, the ionospheric observation value can be extracted from the observation data by using a carrier phase smoothing pseudo-range method or a non-difference non-combination PPP method.

[0065] Specifically, the ionospheric observation value observed by the high-frequency equipment and the ionospheric observation value observed by the general GNSS receiver at each time in a preset time period need to be extracted. Taking a target time in the preset time period as an example, the first ionospheric observation value at the target time is extracted from the first observation data at the target time by using a preset extraction method, and the second ionospheric observation value at the target time is extracted from the second observation data at the target time. In the embodiment, the ionosphere includes a plurality of regions, and each region includes one or more observation points. Therefore, the ionospheric observation value includes the observation value of each region of the ionosphere, and the observation value of each region includes the observation value observed by each observation point.

[0066] In step 300, the ROTI index determination method is used to calculate the ROTI value of each region of the ionosphere at the target time according to the first ionospheric observation value observed by the high-frequency equipment.

[0067] ROTI is a discrimination method for monitoring the activity of the irregular structure of the ionosphere, and the core of the method is to evaluate the dynamic characteristics of the ionosphere by calculating the rate of change of the total electron content (TEC).

[0068] In the embodiment of the present application, the ionospheric observation value at least includes the TEC value, and the ROTI index determination method can be used to determine the dynamic characteristics of each region of the ionosphere according to the ionospheric observation value.

[0069] Specifically, for each observation point, the TEC rate of change (ROT) in the adjacent time interval is calculated according to the ionospheric observation value. Optionally, the ROT can be calculated by the following formula:

[0070]

[0071] wherein TEC t and TEC t-1 are the TEC values of the current time and the previous time of the observation point, respectively, and Δt is the time interval.

[0072] In a certain time interval, the standard deviation of the ROT at each time in the time interval is calculated to obtain the ROTI value, and the ROTI value of each region of the ionosphere is determined according to the ROTI value of each observation point.

[0073] The ROTI value reflects the fluctuation degree of the TEC rate of change, and a higher ROTI value usually indicates the activity degree of the irregular structure of each region of the ionosphere.

[0074] Step 400: according to a preset classification standard, the ROTI values of each region of the ionosphere at the target moment are classified, and the disturbance levels of each region of the ionosphere at the target moment are obtained.

[0075] In the embodiment of the present application, the disturbance level can be divided into n levels, and each level represents a different degree of ionospheric disturbance, and each level corresponds to a ROTI value range. For example, the degree of ionospheric disturbance can be divided into four levels: no disturbance, slight disturbance, moderate disturbance and severe disturbance, wherein no disturbance: ROTI < 0.1 TECU / min; slight disturbance: 0.1 < ROTI < 0.25 TECU / min; moderate disturbance: 0.25 < ROTI < 0.5 TECU / min; severe disturbance: ROTI ≥ 0.25 TECU / min.

[0076] Specifically, the ROTI values of each region of the ionosphere at each moment calculated in step 300 are compared with the ROTI value range corresponding to each disturbance level, and the disturbance level of each region of the ionosphere at each moment is determined.

[0077] Step 500: the disturbance level of each region of the ionosphere at the target moment is labeled on the second ionosphere observation value at the previous moment, and a training sample set is generated, wherein the previous moment is the previous moment of the target moment.

[0078] In order to achieve the accuracy of the disturbance level obtained by inputting the low-frequency observation data collected by the ordinary GNSS receiver, the accuracy of the disturbance result obtained based on the high-frequency observation data collected by the high-frequency device can be the same, and the data of the high-frequency device is labeled to generate the training sample of the ordinary GNSS receiver, so that the model generated based on the training sample can output accurate disturbance level judgment result based on the low-frequency observation data, and the monitoring cost of the ionosphere is reduced.

[0079] Further, in order to realize the prediction of the disturbance level, the disturbance level at t+1 moment is labeled on the observation data at t moment, and the generated training sample is the second ionosphere observation value at the target moment containing the label, and the label is the disturbance level at the next moment of the target moment.

[0080] Step 600: training the model using the training sample set to obtain a disturbance level prediction model, so as to output the disturbance level of the ionosphere at the next moment based on the current ionosphere observation data observed by the ordinary GNSS receiver inputted into the disturbance level prediction model.

[0081] In the embodiment of the present application, the training sample set obtained in step 500 is inputted into the preset model for training, and a trained disturbance level prediction model is obtained.

[0082] When the disturbance level prediction model is applied, the ionospheric observation values ​​of any area of ​​the ionosphere collected by the ordinary GNSS receiver at the current moment are input into the disturbance level prediction model. The disturbance level prediction model outputs the prediction result of the disturbance level of the ionosphere area at the next moment based on the input ionospheric observation values ​​to realize the prediction of the disturbance level.

[0083] The ionospheric disturbance level prediction method proposed in an embodiment of the present invention obtains a long-term series of high-frequency observation data and ordinary observation data to train a disturbance level prediction model. This allows the disturbance level prediction model to make a real-time prediction of the ionospheric disturbance level at the next moment based on the low-frequency observation data. Furthermore, the disturbance level prediction model can still ensure the output of high-precision prediction results when inputting low-frequency observation data, thereby reducing the monitoring cost of the ionosphere.

[0084] In one embodiment of the present invention, a carrier phase smoothed pseudorange method is used to combine pseudorange and phase observations, utilizing the high-precision smoothed pseudorange noise of the carrier phase to improve the accuracy of ionospheric observation extraction. The core concept is to utilize the fact that phase and pseudorange are affected by the ionosphere in equal magnitude and opposite sign, and to reduce the noise effect by averaging ambiguities over consecutive arc segments, thereby achieving improved ionospheric observation extraction accuracy. Based on this, step 200, respectively extracting the first ionospheric observation value observed by the high-frequency device at the target time and the second ionospheric observation value observed by the ordinary GNSS receiver from the first observation data and the second observation data at the target time, specifically includes:

[0085] Step 210: Obtain the first observation equation of the preset phase and pseudorange, and the target epoch corresponding to the target time.

[0086] In an embodiment of the present invention, the original observation equations of phase and pseudorange of the carrier phase smoothing pseudorange method can be simplified as follows:

[0087]

[0088] in: is the geometrically independent combined observation value of the dual-frequency pseudorange from receiver r to satellite s; is the geometrically independent combined observation value of the dual-frequency carrier phase from the receiver r to the satellite s; is the frequency dependence coefficient of ionospheric delay; I is the TEC along the signal path (unit: TECU); c is the speed of light; are the pseudorange hardware delays of receiver r and satellite s, respectively; are the phase hardware delays of receiver r and satellite s respectively; Frequency The wavelength of time, Frequency The ambiguity of the whole week.

[0089] In an embodiment of the present application, the frequency-dependent coefficient of ionospheric delay is calculated by the following formula:

[0090]

[0091] wherein, is the size of i k frequency.

[0092] To simplify the original observation equation, the following definitions are made for the relevant parameters in the original observation equation:

[0093]

[0094]

[0095] Based on the above definitions, the original observation equation can be further simplified into a first observation equation, which is described as follows:

[0096]

[0097] wherein, respectively represent the observation noise of the pseudo-range and the carrier phase, including the multipath effect.

[0098] The noise level of the pseudo-range observation value is high, while the noise of the carrier phase observation value is low (usually millimeter level), but contains the ambiguity term N GF . Within a continuous observation arc segment without cycle slip, N GF is constant. Based on this, the pseudo-range noise can be smoothed by constructing a combination of the pseudo-range and the carrier phase to extract the TEC in the ionospheric observation value.

[0099] Specifically, the pseudo-range and the carrier phase observation values are added:

[0100]

[0101] The above formula is simplified as:

[0102]

[0103] The general hardware bias and is stable in a short time, the product can be used for correction, and the mean value of the noise term is close to zero.

[0104] Step 220, determining an initial epoch range according to a target epoch and a preset arc segment size, the target epoch being within the initial epoch range.

[0105] In the embodiment of the present application, the arc segment size is set in advance, the arc segment is a continuous arc segment, and the arc segment size can determine the length of the year, so that an initial epoch range is determined based on the target epoch, that is, the target epoch is in the initial epoch range.

[0106] Step 230, the ambiguity mean value is calculated according to the observation data at each epoch in the initial epoch range to obtain the optimized combined ambiguity.

[0107] In an embodiment of the present application, the optimized combined ambiguity can be calculated according to the following formula:

[0108]

[0109] In another embodiment of the present application, in order to ensure the reliability of TEC extraction, a method of dynamically adjusting the arc segment length is proposed to solve the problems of signal loss, multipath effect, rapid ionospheric change and abnormal observation. Since the fixed arc segment length cannot adapt to the rapid ionospheric change (such as geomagnetic storm) or stable scene, the method of dynamically adjusting the arc segment length based on the embodiment can dynamically adjust the arc segment size by real-time monitoring the TEC change rate to optimize the smoothing effect. Based on this, Step 230, the ionospheric observation value at least includes the TEC value, and the ambiguity mean value is calculated according to the observation data at each epoch in the initial epoch range to obtain the optimized combined ambiguity, which specifically includes:

[0110] Step 231, the observation data of the previous epoch is obtained, and the previous epoch is the previous epoch of the target epoch.

[0111] Step 232, the initial TEC value at the target epoch and the initial TEC value at the previous epoch are calculated according to the observation data at the target epoch and the previous epoch.

[0112] Specifically, the initial TEC values at the target epoch and the previous epoch are estimated according to the observation data at the target epoch and the previous epoch, and the initial TEC value can be calculated by the following formula:

[0113]

[0114] Wherein, I rough (k) is the initial TEC value at the target epoch. The initial TEC value I rough (k-1) at the previous epoch can also be calculated by the above formula.

[0115] Step 233, the change rate is calculated according to the initial TEC value at the target epoch and the initial TEC value at the previous epoch to obtain the TEC change rate at the target epoch.

[0116] Specifically, the TEC change rate at the target epoch can be calculated by the following formula:

[0117]

[0118] In the formula, ΔI(k) is the TEC change rate at the target epoch, and Δt is the epoch interval between the target epoch and the previous epoch.

[0119] Step 234, according to the TEC change rate at the target epoch and the preset arc segment length adjustment rule, analysis is performed to obtain an updated target epoch range, wherein the arc segment length adjustment rule contains a corresponding relationship between the change rate and the epoch range size, and the target epoch is within the target epoch range.

[0120] Specifically, the epoch range is updated by the following steps:

[0121] A, according to the TEC change rate at the target epoch and the preset arc segment length adjustment rule, analysis is performed to obtain an adjusted epoch range at the target epoch.

[0122] Wherein, the arc segment length adjustment rule is expressed as follows:

[0123]

[0124] In the formula, M(k) is the adjusted epoch range at the target epoch, M min , and M max are the preset minimum epoch range and maximum epoch range respectively, ΔI(k) is the TEC change rate at the target epoch, and θ I is the preset change rate threshold.

[0125] In an embodiment, M min = 10 epochs, which can be applicable to fast changes; M max = 100 epochs, which can be applicable to stable scenarios; θ I can be adjusted according to the actual application scenario environment, for example, θ I = 0.5 TECU / s.

[0126] B, obtaining the adjusted epoch range at the previous epoch; performing exponential weighted average calculation according to the adjusted epoch range at the target epoch and the adjusted epoch range at the previous epoch to obtain the target epoch range at the target epoch.

[0127] To avoid frequent jumping of the epoch range, an embodiment of the present application also uses the exponential weighted average method to smooth the epoch range to obtain the final smoothed target epoch range.

[0128] Specifically, the target epoch range at the target epoch is calculated by the following formula:

[0129] M smoothed (k) = β·M smoothed (k-1)+(1-β)·M(k)

[0130] Where M smoothed (k) is the target epoch range under the target epoch, M smoothed (k-1) is the target epoch range under the previous epoch, β is a preset weight, and optional β=0.9.

[0131] It can be understood that the target epoch range M under the previous epoch in the above formula is smoothed (k-1) can be obtained by solving the target epoch range under the target epoch. In another case, the target epoch range M under the previous epoch smoothed (k-1) can also be M(k-1).

[0132] Step 235: Calculate the ambiguity mean based on the observation data at each epoch within the target epoch range to obtain the optimized combined ambiguity.

[0133] Specifically, the optimized combined ambiguity is calculated by the following formula:

[0134]

[0135] Where, is the optimized combined ambiguity under the target epoch, M smoothed (k) is the target epoch range under the target epoch, is the geometrically independent combined observation value of the dual-frequency pseudorange from receiver r to satellite s at the jth epoch, is the geometrically independent combined observation value of the dual-frequency carrier phase from receiver r to satellite s at the jth epoch, is the pseudorange hardware delay term between receiver r and satellite s.

[0136] In order to cope with the different needs of ionospheric rapid changes and stable scenarios, the embodiment of the present invention proposes a carrier phase smoothing pseudorange algorithm with adaptive adjustment of the smoothing window. By real-time monitoring of the ionospheric TEC change rate, the smoothing window length is dynamically adjusted. This method combines responsiveness to rapid disturbances with accuracy in stable scenarios, significantly improving the TEC calculation accuracy of low-cost GNSS receivers in widely distributed networks.

[0137] Step 240: Update the first observation equation using the combined ambiguity to obtain an updated second observation equation.

[0138] Specifically, after the combined ambiguity is calculated, the combined ambiguity can be substituted into the carrier phase equation in the first observation equation: In other words, the optimized combined ambiguity is used to replace the original ambiguity.

[0139] Step 250, the first observation data and the second observation data of the target time are substituted into the second observation equation respectively to solve, and the first ionospheric observation value observed by the high-frequency device and the second ionospheric observation value observed by the general GNSS receiver at the target time are obtained.

[0140] Specifically, after obtaining the updated second observation equation, the first observation data and the second observation data of the target time are substituted into the second observation equation respectively, and the TEC values observed by the high-frequency device and the general GNSS receiver (that is, I in the second observation equation) are solved.

[0141] Then, on the basis of the ionospheric thin layer assumption, the ionospheric mapping function is calculated using the MSLM (Modified Single Layer Model) projection function, and then the slant path ionospheric delay is converted to the vertical direction ionospheric delay, and the calculation formula is as follows:

[0142]

[0143] In the formula, h is the height at which the ionosphere is concentrated, R is the radius of the earth, and z is the satellite elevation angle.

[0144] In an embodiment of the present application, the ionospheric observation value at least includes TEC value, ROTI, receiver information, satellite information and piercing point information, wherein the receiver information can include receiver ID, receiver longitude and latitude, solar / magnetic index, time characteristics and other information, the satellite information can include satellite ID, satellite elevation angle, azimuth angle and other information; the piercing point information can include IPP grid characteristics, IPP longitude and latitude.

[0145] Therefore, step 500, the disturbance level of each region of the ionosphere at the target time is marked on the second ionospheric observation value at the previous time, and a training sample set is generated, specifically including:

[0146] Step 510, the TEC value, ROTI, receiver information, satellite information and piercing point information of the target region of the ionosphere at the previous time are sequentially arranged to obtain the feature sequence of the target region at the previous time, wherein the target region is any region in all regions of the ionosphere.

[0147] In the embodiment of the present application, the ionospheric observation values of each region of the ionosphere at each moment can be sequentially arranged to obtain the feature sequence of each target region at each moment. For example, the feature sequence of the target region at the previous moment can be (receiver information, satellite information, puncture point region, TEC value, ROTI). It should be noted that the sequence order can be determined according to actual conditions, and is not limited herein.

[0148] Step 520, according to the disturbance level of the target region at the target moment, the feature sequence of the target region at the previous moment is labeled to generate the training sample corresponding to the target region at the previous moment.

[0149] Specifically, according to the disturbance level of the target region at the target moment, the feature sequence of the target region at the previous moment is labeled to update the feature sequence of the target region at the previous moment. The updated feature sequence is the training sample corresponding to the target region at the previous moment. For example, the feature sequence of the target region at the previous moment is originally (receiver information, satellite information, puncture point region, TEC value, ROTI), and the updated feature sequence of the target region at the previous moment is (receiver information, satellite information, puncture point region, TEC value, ROTI, disturbance level).

[0150] Step 530, the training samples corresponding to all regions of the ionosphere at all moments within the preset time period form a training sample set.

[0151] In an embodiment of the present application, step 600, the training sample set is used for model training to obtain a disturbance level prediction model, specifically including: inputting the training sample set into a preset multi-scale spatio-temporal Transformer model for training to generate the disturbance level prediction model.

[0152] To capture the rapid disturbance characteristics reflected by ROTI and enhance the prediction ability of the model, the multi-scale spatio-temporal Transformer (MST-Transformer) model is proposed in this embodiment, which combines multi-scale feature extraction, dynamic attention mechanism and spatial embedding to optimize the modeling of ROTI disturbance mode.

[0153] The specific model architecture and implementation process are as follows:

[0154] 1. Time series encoder:

[0155] For the TEC value and ROTI of the feature sequence, the multi-scale ID convolution in the time series encoder is used to extract features of different time scales, and the extraction process is as follows:

[0156] The multi-scale 1D convolution is used to extract features of different time scales:

[0157]

[0158] where k l ∈{3,5,7} corresponding to short-term (1-3 minutes), medium-term (3-5 minutes), long-term (5-7 minutes) mode respectively.

[0159] 2. Multi-scale feature fusion module:

[0160] The TEC time features and ROTI time features extracted in the previous step are fused to obtain time fusion features, and the fusion process is as follows:

[0161]

[0162] 3. Spatial feature embedding module:

[0163] For discrete features: receiver ID, satellite ID and IPP grid features, feature extraction is performed through the spatial feature embedding module to obtain receiver ID, satellite ID and IPP grid features, and the extraction process is as follows:

[0164] e r = Embedding(Receiver ID)

[0165] e s = Embedding(Satellite PRN)

[0166] e p = Embedding(Grid ID)

[0167] 4. Encode continuous features:

[0168] Including encoding receiver latitude and longitude (φ r ,λ r ), satellite elevation angle ε and azimuth angle α, IPP latitude and longitude (φ IPP ,λ IPP ), and the encoding process is as follows:

[0169] e c = FC(Concat(φ r ,λ r ,ε,α,φ IPP ,λ IPP ))

[0170] 5. Spatial feature fusion module:

[0171] The features extracted in steps 3-4 are fused to obtain spatial fusion features:

[0172] h space= FC (Concat (e r ,e s ,e p ,e c ))

[0173] 6. Contextual feature fusion module:

[0174] The mean of ROTI (ROTI neighbor ) and the mean of TEC (TEC neighbor ) for the surrounding receiver, the solar / magnetic index (F / Kp) and the time feature (h) can be fused according to the contextual feature fusion module to obtain the contextual fusion feature:

[0175] h context = FC (Concat (ROTI neighbor , TEC neighbor , F10.7, Kp, sin (2πh / 24), cos (2πh / 24)))

[0176] 7. Multi-head fusion and classification module:

[0177] (1) Spliced features:

[0178] Splice the spatial fusion feature and the contextual feature fusion module to obtain the final fusion feature:

[0179] h final = FC (Concat (h space , h context ))

[0180] Output disturbance level probability:

[0181] p = Softmax (FC (h final ))

[0182] Predicted label:

[0183]

[0184] Through the multi-scale spatio-temporal Transformer model described above, multi-scale convolution can capture the fluctuations of ROTI in short, medium and long term, adapt to the spatial heterogeneity of ROTI, and make the output disturbance level more accurate.

[0185] To realize real-time monitoring and early warning of ionospheric disturbances, the method uses a multi-scale spatio-temporal Transformer model to process and analyze TEC and ROTI data streams from low-cost GNSS monitoring stations in real time. The model efficiently captures short-term mutations and long-term trends of ionospheric disturbances through multi-scale feature extraction and dynamic attention mechanisms, providing disturbance level predictions with second-level response. Compared with traditional methods, the high timeliness of the multi-scale spatio-temporal Transformer model significantly reduces the potential impact on communication and navigation systems, especially in scenarios such as marine navigation that require high real-time performance.

[0186] In an embodiment of the present application, a prediction device for ionospheric disturbance level is also proposed, which can be referred to Figure 2 , Figure 2 The structure block diagram of the prediction device for ionospheric disturbance level in an embodiment of the present application is shown in the figure. The device includes a data acquisition module 201, a data processing module 202, and a model training module 203.

[0187] The data acquisition module 201 is used to acquire the first observation data observed by the high-frequency device and the second observation data observed by the ordinary GNSS receiver at each time in a preset time period.

[0188] The data processing module 202 is used to extract the first ionospheric observation value observed by the high-frequency device and the second ionospheric observation value observed by the ordinary GNSS receiver at the target time from the first observation data and the second observation data at the target time, respectively. The target time is any time in the preset time period, and the ionospheric observation value contains the observation value of each region of the ionosphere at a certain time. The ROTI value of each region of the ionosphere at the target time is calculated according to the first ionospheric observation value observed by the high-frequency device using the ROTI index determination method. The ROTI values of each region of the ionosphere at the target time are classified according to the preset classification standard to obtain the disturbance level of each region of the ionosphere at the target time.

[0189] The model training module 203 is used to label the disturbance level of each region of the ionosphere at the target time to the second ionospheric observation value at the previous time to generate a training sample set, wherein the previous time is the previous time of the target time. The disturbance level prediction model is obtained by training the training sample set, so as to output the disturbance level of the ionosphere at the next time based on the disturbance level prediction model according to the input ionospheric observation data observed by the ordinary GNSS receiver at the current time.

[0190] The ionospheric disturbance level prediction device provided in the embodiment can train a disturbance level prediction model by acquiring long-time sequence high-frequency observation data and common observation data, so that the disturbance level prediction model can predict the ionospheric disturbance level at the next moment in real time according to low-frequency observation data, and the disturbance level prediction model can still ensure the output of high-precision prediction results when the low-frequency observation data is input, thereby reducing the monitoring cost of the ionosphere.

[0191] Figure 3 The internal structure diagram of the computer device in one embodiment of the application is shown. Figure 3 As shown in the figure, the computer device includes a processor, a memory and a network interface connected through a system bus. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system, and can also store a computer program, which, when executed by the processor, can enable the processor to implement each step in the above method embodiment. The internal memory can also store a computer program, which, when executed by the processor, can enable the processor to execute each step in the above method embodiment. Those skilled in the art can understand that the structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement. Figure 3

[0192] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the computer program being executed by the processor to enable the processor to execute each step in the above method embodiment.

[0193] In one embodiment, a computer readable storage medium is provided, storing a computer program, the computer program being executed by the processor to enable the processor to execute each step in the above method embodiment.

[0194] ​Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer readable storage medium, and when the program is executed, the processes of the above-mentioned embodiment methods can be included. Any reference to memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0195] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0196] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for predicting ionospheric disturbance level, characterized in that: The method comprises: Obtaining first observation data observed by the high-frequency device and second observation data observed by the ordinary GNSS receiver at each moment within a preset time period; Extracting, from the first observation data and the second observation data at a target time, a first ionospheric observation value observed by the high-frequency device and a second ionospheric observation value observed by the ordinary GNSS receiver at the target time, respectively, wherein the target time is any time in the preset time period, and the ionospheric observation values ​​include observation values ​​of various regions of the ionosphere at a certain time; Calculating the ROTI value of each region of the ionosphere at the target time based on the first ionospheric observation value observed by the high-frequency device using a ROTI index determination method; Classifying the ROTI values ​​of each region of the ionosphere at the target time according to a preset classification standard to obtain the disturbance level of each region of the ionosphere at the target time; Annotating the second ionospheric observation value at a previous moment with the disturbance level of each region of the lower ionosphere at the target moment to generate a training sample set, wherein the previous moment is the moment before the target moment; The training sample set is used to perform model training to obtain a disturbance level prediction model, so as to output the disturbance level of the ionosphere at the next moment based on the disturbance level prediction model according to the ionosphere observation data at the current moment observed by the input ordinary GNSS receiver.

2. The method according to claim 1, wherein Extracting the first ionospheric observation value observed by the high-frequency device and the second ionospheric observation value observed by the ordinary GNSS receiver at the target time from the first observation data and the second observation data at the target time, specifically includes: Obtaining a first observation equation of a preset phase and pseudorange, and a target epoch corresponding to the target time; Determine an initial epoch range according to the target epoch and a preset arc segment size, wherein the target epoch is within the initial epoch range; Calculating the ambiguity mean based on the observation data at each epoch within the initial epoch range to obtain an optimized combined ambiguity; updating the first observation equation using the combined ambiguity to obtain an updated second observation equation; The first observation data and the second observation data at the target time are respectively substituted into the second observation equation for solving, to obtain the first ionospheric observation value observed by the high-frequency device at the target time and the second ionospheric observation value observed by the ordinary GNSS receiver.

3. The method according to claim 2, wherein The ionospheric observation value includes at least a TEC value, and performing ambiguity mean calculation based on the observation data at each epoch within the initial epoch range to obtain an optimized combined ambiguity specifically includes: Obtaining observation data of a previous epoch, where the previous epoch is the epoch before the target epoch; Calculating an initial TEC value at the target epoch and an initial TEC value at the previous epoch based on the observation data of the target epoch and the previous epoch; Calculating a change rate based on the initial TEC value at the target epoch and the initial TEC value at the previous epoch to obtain a TEC change rate at the target epoch; Analyzing the TEC change rate at the target epoch and a preset arc length adjustment rule to obtain an updated target epoch range, wherein the arc length adjustment rule includes a corresponding relationship between the change rate and the epoch range, and the target epoch is within the target epoch range; The ambiguity mean is calculated based on the observation data at each epoch within the target epoch range to obtain an optimized combined ambiguity.

4. The method according to claim 3, wherein The updated target epoch range is obtained by analyzing the TEC change rate under the target epoch and the preset arc length adjustment rule, specifically including: Analyzing the TEC change rate at the target epoch and a preset arc length adjustment rule to obtain an adjusted epoch range at the target epoch; Obtaining an adjusted epoch range below the previous epoch; Performing an exponentially weighted average calculation based on the adjusted epoch range under the target epoch and the adjusted epoch range under the previous epoch to obtain a target epoch range under the target epoch; The arc length adjustment rule is as follows: Where M(k) is the adjusted epoch range under the target epoch, M min 、M max are the preset minimum epoch range and maximum epoch range, ΔI(k) is the TEC change rate under the target epoch, θ I is the preset rate of change threshold.

5. The method according to claim 4, wherein The optimized combined ambiguity is calculated by the following formula: Where, is the optimized combined ambiguity under the target epoch, M smoothed (k) is the target epoch range under the target epoch, is the geometrically independent combined observation value of the dual-frequency pseudorange from receiver r to satellite s at the jth epoch, is the geometrically independent combined observation value of the dual-frequency carrier phase from receiver r to satellite s at the jth epoch, is the pseudorange hardware delay term between the receiver r and the satellite s.

6. The method according to claim 1, wherein The ionospheric observation values ​​include at least TEC values, ROTI, receiver information, satellite information and puncture point information; The disturbance level of each region of the ionosphere at the target moment is marked on the second ionosphere observation value at the previous moment to generate a training sample set, which specifically includes: The TEC value, ROTI, receiver information, satellite information, and puncture point information of the ionospheric target area at the previous moment are arranged in order to obtain a characteristic sequence of the target area at the previous moment, wherein the target area is any area in all ionospheric areas; Annotating a feature sequence of the target area at the previous moment according to the disturbance level of the target area at the target moment, and generating a training sample corresponding to the target area at the previous moment; The training samples corresponding to each region of the ionosphere at all times within the preset time period constitute a training sample set.

7. The method according to claim 6, wherein The method of using the training sample set to perform model training to obtain a disturbance level prediction model specifically includes: The training sample set is input into a preset multi-scale spatiotemporal Transformer model for training to generate a disturbance level prediction model.

8. A device for predicting ionospheric disturbance levels, characterized in that: The device includes a data acquisition module, a data processing module and a model training module: The data acquisition module is used to acquire first observation data observed by the high-frequency device and second observation data observed by the ordinary GNSS receiver at each moment within a preset time period; The data processing module is configured to extract, from the first observation data and the second observation data at a target time, a first ionospheric observation value observed by the high-frequency device and a second ionospheric observation value observed by the ordinary GNSS receiver at the target time, respectively, wherein the target time is any time in the preset time period, and the ionospheric observation value includes an observation value of each region of the ionosphere at a certain time; Calculating the ROTI value of each region of the ionosphere at the target time based on the first ionospheric observation value observed by the high-frequency device using a ROTI index determination method; Classifying the ROTI values ​​of each region of the ionosphere at the target time according to a preset classification standard to obtain the disturbance level of each region of the ionosphere at the target time; The model training module is configured to label the disturbance level of each region of the lower ionosphere at the target moment with the second ionosphere observation value at the previous moment to generate a training sample set, wherein the previous moment is the moment before the target moment; The training sample set is used to perform model training to obtain a disturbance level prediction model, so as to output the disturbance level of the ionosphere at the next moment based on the disturbance level prediction model according to the ionosphere observation data at the current moment observed by the input ordinary GNSS receiver.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 7.

10. A computer device comprising a memory and a processor, characterized in that: The memory stores a computer program, and when the computer program is executed by the processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 7.

Images