A Beidou positioning method and system considering ionospheric errors

Abstract

The present invention discloses a Beidou positioning method and system considering ionospheric errors, which relates to the technical field of satellite positioning. A Beidou positioning system considering ionospheric errors includes: a data acquisition module, an atmospheric inference large model module, an ionospheric residual calculation module, a real-time feedback module, a positioning solution calculation module, and a user interface module. Through the ionospheric residual calculation model based on the prediction results of the perturbation state, the present invention can accurately obtain the ionospheric residual correction, statistically analyze the distribution of the ionospheric residuals, further identify and eliminate outliers, or suppress them with weights, optimize the satellite selection strategy, and thus reduce the errors caused by ionospheric perturbations.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite positioning, and in particular to a Beidou positioning method and system considering ionospheric errors. Background Art

[0002] Satellite positioning systems have been widely applied in multiple fields such as navigation, surveying, transportation, and geographic information systems. However, when satellite signals pass through the Earth's atmosphere, especially the ionosphere, they will be affected by the uneven distribution of electron density, resulting in changes in the propagation speed, and further causing ranging errors. In severe cases, it may lead to the failure of positioning solution or a significant decrease in accuracy.

[0003] Ionospheric error is one of the largest error sources in satellite positioning. Especially during periods of enhanced solar activity and intense electromagnetic disturbances, such as magnetic storm events and high-latitude aurora activities, ionospheric disturbances will significantly affect the stability and accuracy of satellite signals. To address this problem, existing technologies mainly adopt methods such as dual-frequency differential, ionospheric model correction, or differential positioning based on observational data to offset or correct ionospheric errors. However, these methods generally suffer from problems such as insufficient model accuracy, inability to dynamically adapt to the disturbance environment, and lag in error estimation. Especially in the case of drastic changes in the ionosphere, traditional methods are difficult to provide stable and reliable positioning solutions.

[0004] In recent years, with the enrichment of data resources such as remote sensing, meteorology, and space environment observations, as well as the development of artificial intelligence technology, using multi-modal data fusion and deep learning methods to model and predict ionospheric errors has become a new research trend. Multi-frequency satellite signals provide more abundant observational information, while meteorological data, solar activity parameters, and ionospheric disturbance data help to more comprehensively characterize the ionospheric state. However, there is still a lack of mature and implementable solutions on how to effectively fuse these multi-source information and make it play a role in real-time satellite positioning. Summary of the Invention

[0005] A Beidou positioning method considering ionospheric errors includes:

[0006] Obtain multi-modal data, including multi-frequency satellite signals, ionospheric residual data, meteorological data, solar activity data, and ionospheric disturbance data;

[0007] Process the multi-modal data through an atmospheric inference large model to generate a prediction of the disturbance state and activity of the ionosphere at the current epoch;

[0008] Based on the prediction result of the disturbance state, use an ionospheric residual calculation model to obtain ionospheric residual correction and residual statistical information; and identify and suppress or eliminate outlier satellites with weighted according to the activity prediction and residual statistical information to obtain a satellite selection strategy;

[0009] Combined with real-time positioning error feedback, adaptively adjust ionospheric residual correction and satellite selection strategy;

[0010] When ionospheric activity is detected, the system switches to an ionospheric-free combined filtering strategy, correcting ionospheric errors by combining multi-frequency signals to provide accurate positioning solutions. The ionospheric activity is determined by the ionospheric residual calculation model.

[0011] It provides real-time query and short-term forecast capabilities for ionospheric disturbance conditions, generates positioning accuracy assessment reports based on disturbance state predictions and activity predictions, and provides data support for high-precision operations.

[0012] As a preferred technical solution of the present invention, the atmospheric reasoning large model includes:

[0013] Input layer, receives and preprocesses multimodal data;

[0014] The disturbance state prediction layer learns the spatiotemporal features of multimodal data based on a multi-scale convolutional neural network and an autoregressive model to generate a disturbance score.

[0015] The ionospheric activity prediction layer performs a weighted fusion of the disturbance score, ROTI, and VTEC through a multi-head self-attention mechanism to output an activity prediction. ROTI is a statistic of the temporal rate of change of the total electron content of the ionosphere, and VTEC is the total electron content in the vertical direction of the ionosphere.

[0016] Adaptive correlation adjustment layer, which performs correlation verification on key parameters, disturbance states and disturbance scores in multimodal data according to the ionospheric correlation table, and obtains optimized disturbance state prediction and activity prediction;

[0017] The output layer generates disturbance state prediction and activity prediction.

[0018] As a preferred technical solution of the present invention, the atmospheric reasoning large model is trained and acquired based on historical ionospheric data.

[0019] As a preferred technical solution of the present invention, the ionospheric activity is calculated jointly based on ROTI and VTEC and normalized and divided into four levels: calm, moderate, strong, and super strong.

[0020] As a preferred technical solution of the present invention, the ionospheric residual calculation model includes:

[0021] The ionospheric residual correction layer models and estimates the ionospheric delay based on the disturbance state prediction results output by the atmospheric inference large model and combines the multi-frequency satellite signals of the current epoch to generate the ionospheric residual correction value for each satellite;

[0022] Ionospheric residual statistics layer, which statistically analyzes the ionospheric residuals of all Beidou satellites within the current epoch, calculates their mean and standard deviation, and identifies the satellites whose ionospheric residuals exceed the outlier determination threshold;

[0023] Satellite selection layer, which obtains the satellite selection strategy, performs weighted suppression on the satellites whose ionospheric residuals are greater than the outlier determination threshold and less than or equal to the rejection threshold; rejects the satellites whose residuals exceed the rejection threshold;

[0024] Ionospheric activity determination layer, which determines whether the ionosphere is in an active state based on the real-time perturbation state prediction result through the ratio of Ion75 and Tro90. When Ion75 > 0.006 and Ion75 > Tro90, it is marked as ionospheric active and triggers the ionosphere-free combined filtering strategy.

[0025] As a preferred technical solution of the present invention, the ionospheric residual calculation model adopts the XGBoost model based on the gradient boosting tree algorithm, which is trained using historical ionospheric residual data and multi-frequency satellite signal data to achieve fine modeling and prediction of ionospheric errors; the model is iteratively updated by continuously introducing real-time feedback data to improve the accuracy and robustness of residual correction under different ionospheric active conditions.

[0026] As a preferred technical solution of the present invention, the real-time positioning error feedback includes the fixation rate, error distribution, and prediction deviation.

[0027] As a preferred technical solution of the present invention, the ionosphere-free combined filtering strategy includes:

[0028] Cancel the ionospheric residual correction and instead perform positioning solution based on geometric and differential information;

[0029] When the ionospheric activity returns to calm for K consecutive epochs, automatically resume the normal positioning mode, where K is the dynamic monitoring index.

[0030] As a preferred technical solution of the present invention, the real-time query and short-term forecast capabilities of the ionospheric disturbance situation include:

[0031] Generate ionospheric disturbance forecasts through real-time monitoring and analysis of multi-modal data; optimize the outlier determination threshold and rejection threshold in the satellite selection strategy based on the ionospheric disturbance forecasts; provide user-defined forecast periods and disturbance prediction threshold settings, and provide historical data backtracking functions.

[0032] A Beidou positioning system considering ionospheric errors, including:

[0033] Data acquisition module: Real-time collect multi-modal data and transmit it to the atmospheric inference large model for processing;

[0034] Atmospheric Inference Large Model Module: Outputs disturbance state prediction and activity prediction based on multi-modal data;

[0035] Ionospheric Residual Calculation Module: Obtains ionospheric residual correction and residual statistical information, and obtains satellite selection strategies;

[0036] Real-time Feedback Module: Combines real-time positioning error feedback to adaptively adjust ionospheric residual correction and satellite selection strategies;

[0037] Positioning Solution Module: Performs positioning solution based on ionospheric residual correction and satellite selection strategies. When the ionospheric disturbance exceeds the activity threshold, it switches to a non-ionospheric combined filtering strategy for solution;

[0038] User Interface Module: Provides positioning results and positioning accuracy evaluation reports to users, and supports users' monitoring and adjustment of positioning services.

[0039] The present invention has the following advantages:

[0040] By acquiring multi-frequency satellite signals, ionospheric residual data, meteorological data, solar activity data, and ionospheric disturbance data, and processing these data using an atmospheric inference large model, the present invention can comprehensively evaluate the ionospheric disturbance state and activity. By fusing multi-source information, the accuracy of ionospheric disturbance prediction is significantly improved, thereby enhancing the positioning accuracy.

[0041] Through a real-time positioning error feedback mechanism, the present invention adaptively adjusts ionospheric residual correction and satellite selection strategies. In the case of changes in ionospheric activity, it can quickly respond and adjust the positioning strategy to ensure the stability and accuracy of the positioning solution. This adaptive ability enables the positioning system to cope with the dynamic changes of the ionosphere and provide more reliable positioning services.

[0042] Through an ionospheric residual calculation model based on the disturbance state prediction result, the present invention can accurately obtain ionospheric residual correction, statistically analyze the distribution of ionospheric residuals, and then identify and eliminate outlier satellites, or suppress them by weighting, optimizing the satellite selection strategy, thereby reducing errors caused by ionospheric disturbances.

[0043] When the ionosphere is detected to be active, the present invention can automatically switch to a non-ionospheric combined filtering strategy, using multi-frequency signal combinations to correct ionospheric errors, thereby effectively avoiding the negative impact on the positioning solution during ionospheric activity and improving the robustness of the positioning system under extreme ionospheric conditions.

[0044] The present invention provides the ability to query the ionospheric disturbance status in real time and make short-term forecasts, can predict the change trend of ionospheric activity in advance, provides data support for the positioning system, helps optimize the satellite selection strategy, and provides an accurate positioning accuracy evaluation report for high-precision operations.

[0045] Through ionospheric disturbance prediction, activity prediction and real-time feedback adjustment mechanism, the present invention can provide more accurate and stable positioning services in high-precision measurement and navigation operations, meet various high-precision application requirements, and is particularly suitable for measurement, navigation and positioning tasks that require high precision and high reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only schematic diagrams of the present invention. Those skilled in the art can also derive other drawings based on the provided drawings without inventive effort.

[0047] Figure 1 A schematic diagram of the structure of a Beidou positioning system that takes ionospheric errors into account, as adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] To make the objectives, technical solutions, and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0049] Embodiment 1, a Beidou positioning method considering ionospheric error, comprises the following steps:

[0050] Step S1: Acquire multimodal data, including multi-frequency satellite signals, ionospheric residual data, meteorological data, solar activity data, and ionospheric disturbance data. The following is a further explanation of the technical details and implementation methods for acquiring and processing each data type:

[0051] S101. Acquisition of multi-frequency satellite signal data:

[0052] Multi-frequency satellite signal data is the core data for satellite positioning. It includes satellite signals from at least two frequencies (L1 and L2). Received by a GNSS receiver with high temporal accuracy and spatial resolution, the receiver receives and decodes satellite signals from different frequencies, obtaining corresponding pseudoranges, carrier phases, and signal strength information. The received satellite signals are decoded to extract navigation information (satellite orbit information and timestamps).

[0053] S102. Acquisition of ionospheric residual data:

[0054] Ionospheric residual data refers to the errors caused by the inhomogeneity of the ionosphere when satellite signals pass through it. These residuals are important error sources in satellite positioning. Ionospheric residual data is obtained through differential positioning technology or based on ionospheric models. The specific process includes:

[0055] Obtaining ionospheric residual data through GNSS reference stations deployed at different geographical locations; or estimating ionospheric residuals through an ionospheric model (Klobuchar model) based on parameters such as satellite visibility, geographical location, and total electron content (TEC) of the ionosphere.

[0056] S103. Acquisition of meteorological data:

[0057] Meteorological data has an indirect impact on ionospheric activity and positioning errors, especially parameters such as atmospheric pressure, temperature, and humidity. These meteorological data are mainly obtained through meteorological stations and satellite meteorological observation systems, including air temperature, air pressure, humidity, wind speed, etc.; or through remote sensing satellites (meteorological satellites) to obtain meteorological data over a larger range, including temperature distribution and humidity distribution in the atmosphere. This data is obtained through a public data platform and is used in real-time to assist in ionospheric modeling and analysis.

[0058] Fuse the data obtained from ground meteorological stations and satellite systems to improve the accuracy of meteorological data.

[0059] S104. Acquisition of solar activity data:

[0060] Solar activity has a direct impact on ionospheric disturbances, especially during solar activity periods. It is obtained through specialized solar observatories or through the International Space Weather Forecast System, including sunspot numbers, solar wind speed, solar radiation intensity, etc.

[0061] S105. Acquisition of ionospheric disturbance data:

[0062] Ionospheric disturbance data refers to the influence of disturbances or inhomogeneous distributions in the ionosphere during certain time periods or geographical regions, including:

[0063] ROTI (Rate of change of total electron content of the ionosphere), obtained by monitoring the rate of change of ionospheric TEC; VTEC (Vertical total electron content of the ionosphere), VTEC data reflects the electron density distribution of the ionosphere. Especially during strong solar activity, VTEC data is used to evaluate the activity of the ionosphere. Ionospheric disturbance data is obtained in real-time through satellites and ground ionospheric monitoring stations.

[0064] Step S2: Process the multi-modal data through an atmospheric inference large model to generate predictions of the disturbance state and activity of the ionosphere at the current epoch;

[0065] S201. The atmospheric inference model is a neural network model built using deep learning technology. Its primary task is to learn and analyze the spatiotemporal characteristics of multimodal data and generate predictions of the ionospheric disturbance state and activity. The model consists of multiple layers, including an input layer, a disturbance state prediction layer, an ionospheric activity prediction layer, an adaptive correlation adjustment layer, and an output layer.

[0066] S202. Input layer, responsible for receiving and preprocessing the multimodal data from step S1. In order to ensure the comparability of data from different sources, all input data need to be standardized and normalized. The specific method includes converting satellite signal data into set standard units, quantifying ionospheric residual data and ionospheric disturbance data, and making the scale of each data source consistent. Perform integrity check on each data. If there is data with integrity less than the set value, Lagrange interpolation is used to fill the missing values to ensure the continuity and integrity of the data. Sort the multimodal data according to timestamps and divide the time window to ensure the accurate extraction of spatiotemporal features.

[0067] S203. Disturbance State Prediction Layer: This layer uses a multiscale convolutional neural network (CNN) and an autoregressive model to learn spatiotemporal features from multimodal data and generate a disturbance score. The multiscale CNN extracts spatial features from the ionospheric data. By applying convolutional filters to the input ionospheric residual data and ionospheric disturbance data, it identifies local patterns and features (such as ionospheric density changes and disturbance phenomena). The AR analyzes the time series characteristics of the ionospheric data, particularly its dynamics.

[0068] The combination of multi-scale CNN and AR model enables the model to simultaneously capture the spatial distribution characteristics and temporal variation patterns of ionospheric data, and more accurately predict the state of ionospheric disturbances.

[0069] S204, the ionospheric activity prediction layer: This layer uses a multi-head self-attention mechanism to perform a weighted fusion of the disturbance score, ROTI (rate of change of total electron content in the ionosphere), and VTEC (vertical electron content in the ionosphere), to output an ionospheric activity forecast. ROTI and VTEC are two important parameters that reflect ionospheric disturbances. ROTI indicates the rate of change of the total electron content in the ionosphere over time, while VTEC measures the electron density in the vertical direction of the ionosphere. These two parameters provide a quantitative assessment of the intensity of ionospheric activity.

[0070] Automatically identify the important relationships between disturbance scores, ROTI, and VTEC through the self-attention mechanism, and assign different weights to different parameters to more accurately evaluate the activity of the ionosphere. The multi-head self-attention mechanism analyzes the relationship between the disturbance state and activity from multiple perspectives by processing multiple different attention heads in parallel, enhancing the model's adaptability to complex ionospheric phenomena.

[0071] S205, Adaptive Association Adjustment Layer: According to the ionospheric association table, perform association verification on the key parameters in the multi-modal data (initially set as the sunspot number, air temperature, and air pressure in meteorological data), disturbance state, and disturbance scores, and optimize the disturbance state prediction and activity prediction.

[0072] The ionospheric association table contains the historical data of the ionosphere, the historical association relationships between the disturbance state and activity, and the change patterns of the ionosphere under different meteorological conditions. By referring to the association table, automatically correct the current prediction results to ensure the accuracy of the prediction.

[0073] S206. Output Layer: Integrate the processing results of the foregoing layers to generate the final disturbance state prediction and activity prediction.

[0074] The activity of the ionosphere includes four levels: quiet, medium, strong, and super strong, as shown in the following table:

[0075]

[0076] S207. Model Training; The large atmospheric inference model is trained through historical ionospheric data. During the training process, the model will adjust the parameters according to the historical data to optimize the prediction results of the disturbance state and activity. Use supervised learning and reinforcement learning methods to adapt to changes under different ionospheric conditions; and use cross-validation and early stopping mechanisms to ensure generalization ability and avoid overfitting.

[0077] Step S3: Based on the disturbance state prediction results, use the ionospheric residual calculation model to obtain the ionospheric residual correction and residual statistical information; and according to the activity prediction and residual statistical information, identify the outlier satellites for weighted suppression or elimination to obtain the satellite selection strategy.

[0078] Ionospheric Residual Correction Layer: According to the disturbance state prediction results output by the large atmospheric inference model, combined with the multi-frequency satellite signals of the current epoch, perform ionospheric delay modeling and residual estimation to generate the ionospheric residual correction value for each satellite.

[0079] Ionospheric Delay Modeling: Combine the data of the current multi-frequency satellite signals with the ionospheric disturbance prediction results, and use the ionospheric model to estimate the ionospheric delay; the ionospheric delay is the propagation delay caused by the change of electron density when the satellite signal passes through the ionosphere.

[0080] Ionospheric residual correction: Calculate the ionospheric residuals based on the difference between the actual propagation delay and the theoretical value of the satellite signal. The ionospheric residual correction value for each satellite is corrected through perturbed state prediction. During the ionospheric residual correction process, we use the following formula to calculate the weight of the ionospheric residuals:

[0081] ;

[0082] where is the current value of the ionospheric residual, representing the error amount of the ionospheric delay; is a threshold set according to experience, defining the range of the ionospheric residuals; is a normalization factor, representing the relative magnitude between the current residual value and the threshold, used to adjust the weight of the ionospheric residuals; the weight is used to control the influence of the ionospheric residuals; this formula dynamically adjusts the correction weight according to the magnitude of the ionospheric residuals, ensuring that smaller residuals do not affect the positioning result, while larger residuals are reduced by weighted suppression to avoid excessive interference with the positioning accuracy.

[0083] S302. Ionospheric residual statistics layer: Statistically analyze the ionospheric residuals of all Beidou satellites within the current epoch, calculate their mean and standard deviation, and identify the satellites whose ionospheric residuals exceed the outlier determination threshold.

[0084] Calculation of mean and standard deviation: Statistically analyze the ionospheric residuals of all satellites within the current epoch, and calculate the mean and standard deviation , that is: where is the ionospheric residual of the i-th satellite, is the total number of satellites;

[0085] S303. Satellite selection layer: According to the output of the ionospheric residual statistics layer, formulate a satellite selection strategy, and perform weighted suppression or rejection on the satellites.

[0086] Weighted suppression and rejection: For satellites whose ionospheric residuals are greater than the outlier determination threshold but less than or equal to the rejection threshold, perform weighted suppression. For satellites whose ionospheric residuals exceed the rejection threshold, execute the rejection strategy to completely remove the observation data of these satellites to reduce their influence on the positioning result.

[0087] Weighted suppression algorithm: Weighted suppression adjusts the contribution of a satellite according to the severity of its residual by applying a weighting factor.

[0088] S304. Ionospheric activity determination layer. Based on the real-time prediction results of the disturbance state, it determines whether the ionosphere is in an active state by judging the ratio of Ion75 to Tro90. By sorting the residual data of the ionosphere and the troposphere, the values of Ion75 and Tro90 are calculated. When Ion75 > 0.006 and Ion75 > Tro90, it is marked as ionospheric activity, and the ionosphere-free combined filtering strategy is triggered; otherwise, it is considered that the ionosphere is calm, and the conventional positioning mode (RTK mode) is continued.

[0089] S305. The ionospheric residual calculation model is trained using the XGBoost model based on the gradient boosting tree algorithm. It uses historical ionospheric residual data and multi-frequency satellite signal data for modeling and prediction. The model is trained through the XGBoost algorithm to learn the rules of ionospheric residuals and perform residual prediction. The model is iteratively updated by continuously introducing real-time feedback data to improve the accuracy and robustness of residual correction under different ionospheric activity conditions.

[0090] Step S4: Combine the real-time positioning error feedback, adaptively adjust the ionospheric residual correction and satellite selection strategy, and perform positioning solution;

[0091] S401. Obtaining real-time positioning error feedback. Real-time positioning error feedback refers to the positioning accuracy information obtained in each positioning epoch, mainly including the fixation rate, error distribution, and prediction deviation.

[0092] Fixation rate: The fixation rate refers to the proportion of successfully obtaining a fixed solution (i.e., an accurate positioning result). In high-precision positioning, the fixed solution is an ideal state, indicating that the positioning result is relatively stable and the error is small. Real-time monitoring of the fixation rate is used to judge whether the current positioning is reliable. A lower fixation rate usually indicates a larger ionospheric error or unstable satellite signals.

[0093] Error distribution: The error distribution describes the distribution of the positioning error in space and time. By analyzing the error distribution, it is judged whether the error is concentrated in a certain specific direction (such as the vertical direction) or evenly distributed globally. The error distribution information is used to identify the source of the error and provide directional guidance.

[0094] Prediction deviation: The prediction deviation refers to the difference between the predicted error and the actual error in real-time positioning. By calculating the prediction deviation, the prediction ability of the large atmospheric inference model under ionospheric disturbance conditions is evaluated. If the prediction deviation is too large, it means that the large atmospheric inference model fails to accurately capture the current characteristics of the ionosphere, and the hyperparameters of the model need to be adjusted.

[0095] S402. Real-time positioning error analysis. The core purpose of real-time positioning error feedback is to analyze the sources of positioning errors and adaptively adjust the ionospheric residual correction and satellite selection strategies according to different error types. The error analysis process includes the following steps:

[0096] According to the spatial distribution of the positioning error, determine whether the error is concentrated in a specific area. If the error is mainly concentrated in areas with large ionospheric influence (such as high-latitude or solar activity active areas), strengthen the ionospheric error correction or adjust the satellite selection strategy. By analyzing the changing trend of the error over time, determine whether the ionospheric error is stable. If the error changes violently in a short period of time, it indicates that the ionospheric activity has changed, and quickly respond and adjust the weight W of the ionospheric residual correction model. Combine the prediction results of the ionospheric disturbance state and the real-time positioning error feedback, and analyze the relationship between the two. If the ionospheric disturbance state is strong and the positioning error is large, it means that the current ionospheric error correction strategy is not precise enough.

[0097] S403. Adaptive adjustment of ionospheric residual correction and satellite selection strategies. Based on the real-time positioning error feedback and the error analysis results, adaptively adjust the ionospheric residual correction and satellite selection strategies, including:

[0098] According to the fixation rate and error distribution in the real-time error feedback, determine whether the current ionospheric residual correction strategy is effective. If the fixation rate is low and the error distribution is uneven, it indicates that the current ionospheric correction is insufficient, and automatically enhance the correction of the ionospheric residual. The adjustment methods include dynamically updating the correction value of the ionospheric residual; enhancing the correction intensity of the ionospheric residual during the active period of the ionosphere; reducing the correction intensity when the ionosphere is quiet to avoid errors caused by overcorrection.

[0099] According to the real-time error feedback and satellite observation data, dynamically adjust the satellite selection strategy. Specifically, regulate the values of the outlier determination threshold and the rejection threshold, and adjust the intensity of weighted suppression;

[0100] S404. Positioning solution. After completing the error analysis and strategy adjustment, perform positioning solution according to the new ionospheric residual correction and satellite selection strategies. The positioning solution process includes the following steps:

[0101] Use the weighted least squares method to perform positioning solution on the adjusted satellite signals. The signals of each satellite are weighted according to their weights (weights after weighted suppression), so as to optimize the final positioning result.

[0102] During the real-time positioning process, use the Kalman filtering algorithm to dynamically adjust the positioning results of each epoch. The Kalman filter optimizes the positioning solution according to the real-time error feedback and prediction deviation, and improves the solution accuracy.

[0103] Step S5: When the ionosphere is detected to be active, switch to the ionosphere-free combined filtering strategy to correct the ionospheric error through multi-frequency signal combination and provide accurate positioning solution. The ionospheric activity is determined by the ionospheric residual calculation model;

[0104] S501. Ionospheric activity determination. In the ionospheric activity determination layer (Step S3), the system determines whether the ionosphere is active by analyzing the ratio of Ion75 and Tro90: Ion75 and Tro90 are values used to measure the ionospheric and tropospheric residuals. Ion75 represents the quantile of the ionospheric residual, and Tro90 represents the quantile of the tropospheric residual. When Ion75 > 0.006 and Ion75 > Tro90, the ionosphere is determined to be active, triggering the ionosphere-free combined filtering strategy.

[0105] S502. Switching of the ionosphere-free combined filtering strategy. When the ionosphere is determined to be active, automatically switch to the ionosphere-free combined filtering strategy. This strategy corrects the ionospheric error through multi-frequency signal combination and improves the accuracy of the positioning solution. In the ionosphere-free combined filtering mode, the ionospheric residual correction is cancelled; when the ionosphere is active, it is more stable to directly perform positioning solution through geometric and differential information.

[0106] The positioning solution is changed to be based on the geometric information and differential information of the multi-frequency signal combination. By combining multi-frequency signals (L1 and L2 frequencies), the influence of the ionospheric error is eliminated. The multi-frequency signal combination is calculated by the following formula:

[0107] ; where are the pseudorange observations of the L1 and L2 frequencies respectively, are the carrier phases of the L1 and L2 frequencies respectively, are the frequencies of the L1 and L2 frequencies respectively, and are the combined pseudorange and carrier phase respectively.

[0108] The above formula reduces the influence brought by the ionospheric error through the weighted combination of multi-frequency signals, thereby improving the accuracy of the positioning solution.

[0109] S503. Restore the RTK mode. After the ionospheric activity returns to calm, automatically restore to the RTK mode. This means that when the ionospheric perturbation decreases and returns to calm, the influence of the ionospheric error weakens, and the RTK mode can provide a more accurate positioning solution through higher-precision differential technology.

[0110] Continuous K Epochs Return to Calm: The system sets a dynamic monitoring index K. When the ionospheric activity returns to a calm state within continuous K epochs (i.e., the values of Ion75 and Tro90 are lower than the set thresholds), it automatically switches back to the RTK mode. The value of K is a dynamically adjustable parameter, which is set according to the actual situation to balance the flexibility and accuracy of the system.

[0111] RTK Mode: The RTK mode is based on differential technology and calculates the positioning error through the signal difference between the reference station and the mobile station. This mode is relatively sensitive to ionospheric errors, but can provide the highest positioning accuracy when the ionosphere is calm.

[0112] Step S6: Provide the ability to query the real-time ionospheric disturbance status and short-term forecast. Generate a positioning accuracy evaluation report based on the disturbance state prediction and activity prediction to provide data support for high-precision operations.

[0113] S601. Real-time query and monitoring of the ionospheric disturbance status. The real-time query ability of the ionospheric disturbance is based on the real-time monitoring and analysis of multi-modal data. By continuously obtaining information such as multi-frequency satellite signals, ionospheric residual data, meteorological data, solar activity data, and ionospheric disturbance data, the ionospheric disturbance status is evaluated in real time.

[0114] Collect multi-modal data from different sources and transmit it to the data processing platform for synchronous processing in real time to generate real-time monitoring data of the ionospheric disturbance status.

[0115] Through the analysis of the ionospheric data collected in real time, combined with the atmospheric inference large model (as described in step S2), a predictive evaluation of the ionospheric disturbance state is carried out, and the disturbance forecast is updated in real time.

[0116] Provide an interface that can be queried by users, allowing users to view the detailed information of the current ionospheric disturbance in real time, including the activity of the ionosphere, the severity of the disturbance, and the impact on positioning accuracy. To help users optimize their operation strategies under different ionospheric conditions.

[0117] S602. The short-term forecast ability is to predict the change trend of the ionospheric disturbance in the next few hours to several days, especially the change of the ionospheric activity, through the historical analysis and real-time monitoring of the ionospheric disturbance data.

[0118] Prediction Model: Based on the historical ionospheric disturbance data and solar activity data, combined with the support vector machine, the future ionospheric activity is predicted. By learning the patterns of past data, the periodic and trend changes of the ionospheric disturbance are identified, so as to generate an accurate short-term forecast.

[0119] By combining ROTI and VTEC, predict the changes in the ionospheric activity in the next few hours to days, provide early warnings for high-precision operations, enable users to take corresponding preventive measures when the ionospheric interference is strong, and optimize the satellite selection and positioning solution strategies.

[0120] According to specific application requirements, support users to customize the forecast period (such as 1 hour, 6 hours, 12 hours, etc.) and prediction accuracy. The settings of the forecast period and accuracy determine the output time range and the detailed degree of ionospheric disturbances.

[0121] S603. Optimize the satellite selection strategy based on disturbance forecasts. According to the ionospheric disturbance forecast results, the system can adjust the satellite selection strategy in real time, especially optimize the decision thresholds and rejection thresholds for outlier satellites.

[0122] By analyzing the ionospheric disturbance forecast, adjust the decision threshold for outlier satellites in the satellite selection strategy. When the ionospheric activity is high, satellite signals may be more strongly affected by the ionosphere, resulting in larger positioning errors. At this time, the system will lower the decision threshold for outlier satellites.

[0123] Dynamically adjust the rejection threshold according to the accuracy of the disturbance forecast and the prediction of ionospheric activity. For example, when the ionospheric forecast shows strong activity, the system will increase the rejection threshold.

[0124] The satellite selection strategy is not only based on ionospheric disturbance forecasts, but also combined with real-time positioning error feedback. For example, when the fixation rate is low, the system will increase the weights of the remaining satellites through weighted suppression to reduce errors and enhance positioning accuracy.

[0125] S604. User-defined forecast period. To meet different application requirements, support the user-defined function, allowing users to set the time range and accuracy of the ionospheric disturbance forecast according to specific needs.

[0126] Users set the forecast period according to the operation scenario. For example, when highly accurate positioning tasks are required in a short time, users can select a shorter forecast period (such as 1 hour) to obtain ionospheric disturbance information in real time. For long-term operation tasks, users can also select a longer forecast period (such as 1 day or 2 days) to plan operations in advance. Flexibly adapt to different operation requirements, especially for high-precision positioning operations in complex environments, and provide the best data support and warning capabilities.

[0127] S605. Historical data backtracking function and positioning error assessment, which help users review past ionospheric disturbance events and conduct relevant positioning error analysis and assessment.

[0128] Users can view past ionospheric disturbance events and their impact on positioning errors through the historical data retrieval function. By reviewing historical data, users can identify the patterns of ionospheric disturbances and prepare countermeasures in advance.

[0129] Based on historical ionospheric disturbance data and actual positioning results, a positioning error assessment report is generated to help users understand the impact of past ionospheric disturbances on positioning accuracy and provide a reference for future operations.

[0130] Example 2, a Beidou positioning system considering ionospheric errors, see Figure 1 shown in the figure, includes the following modules:

[0131] Data acquisition module: Real-time acquisition of multi-modal data and transmission to the atmospheric inference large model for processing;

[0132] Atmospheric inference large model module: Output disturbance state prediction and activity prediction based on multi-modal data;

[0133] Ionospheric residual calculation module: Obtain ionospheric residual correction and residual statistical information, and obtain satellite selection strategies;

[0134] Real-time feedback module: Combine real-time positioning error feedback to adaptively adjust ionospheric residual correction and satellite selection strategies;

[0135] Positioning solution module: Perform positioning solution based on ionospheric residual correction and satellite selection strategies. When the ionospheric disturbance exceeds the activity threshold, switch to the non-ionospheric combined filtering strategy for solution;

[0136] User interface module: Provide users with positioning results and positioning accuracy assessment reports, and support users' monitoring and adjustment of positioning services.

[0137] The specific implementation manners described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific implementation manners of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A Beidou positioning method considering ionospheric errors, characterized in that, Including: Obtain multi-modal data, including multi-frequency satellite signals, ionospheric residual data, meteorological data, solar activity data, and ionospheric disturbance data; Process the multi-modal data through an atmospheric inference large model to generate a prediction of the disturbance state and activity level of the ionosphere at the current epoch; Based on the prediction result of the disturbance state, use the ionospheric residual calculation model to obtain ionospheric residual correction and residual statistical information; And according to the activity prediction and residual statistical information, identify outlier satellites for weighted suppression or elimination to obtain a satellite selection strategy; The ionospheric residual calculation model includes: Ionospheric residual correction layer, according to the prediction result of the disturbance state output by the atmospheric inference large model, combined with the multi-frequency satellite signals at the current epoch, model and estimate the ionospheric delay, and generate the ionospheric residual correction value for each satellite; Ionospheric residual statistics layer, statistically calculate the ionospheric residuals of all Beidou satellites within the current epoch, calculate their mean and standard deviation, and identify satellites with ionospheric residuals exceeding the outlier determination threshold; Satellite selection layer, obtain the satellite selection strategy, perform weighted suppression on satellites with ionospheric residuals greater than the outlier determination threshold and less than or equal to the elimination threshold; eliminate satellites exceeding the elimination threshold; Ionospheric activity determination layer, according to the real-time disturbance state prediction result, judge whether the ionosphere is in an active state through the ratio of Ion75 and Tro90. When Ion75 > 0.006 and Ion75 > Tro90, mark it as ionospheric active and trigger the no-ionosphere combined filtering strategy. The Ion75 and Tro90 are values used to measure the ionospheric and tropospheric residuals. Ion75 represents the quantile of the ionospheric residual, and Tro90 represents the quantile of the tropospheric residual; Combined with real-time positioning error feedback, adaptively adjust the ionospheric residual correction and satellite selection strategy; When ionospheric activity is detected, switch to the no-ionosphere combined filtering strategy, correct the ionospheric error through multi-frequency signal combination, and provide accurate positioning solution. The ionospheric activity is determined by the ionospheric residual calculation model; Provide the ability to query the real-time ionospheric disturbance status and short-term forecast, generate a positioning accuracy assessment report according to the disturbance state prediction and activity prediction, and provide data support for high-precision operations.

2. The Beidou positioning method considering ionospheric error according to claim 1, wherein: The atmospheric inference large model includes: Input layer, receive and preprocess multi-modal data; Disturbance state prediction layer, perform spatio-temporal feature learning on multi-modal data based on a multi-scale convolutional neural network and an autoregressive model to generate a disturbance score; Ionospheric activity prediction layer, perform weighted fusion on the disturbance score, ROTI, and VTEC through a multi-head self-attention mechanism, and output the activity prediction; the ROTI is a statistic of the rate of change of the total electron content in the ionosphere over time, and the VTEC is the total electron content in the vertical direction of the ionosphere; Adaptive correlation adjustment layer, perform correlation verification on the key parameters, disturbance state, and disturbance score in the multi-modal data according to the ionospheric correlation table, and obtain optimized disturbance state prediction and activity prediction; Output layer, generate disturbance state prediction and activity prediction.

3. The Beidou positioning method considering ionospheric error according to claim 2, characterized in that, The atmospheric inference large model is obtained by training based on historical ionospheric data.

4. The Beidou positioning method considering ionospheric error according to claim 2, characterized in that: The ionospheric activity is classified into four levels: quiet, medium, strong, and super-strong according to the combined calculation and normalization of ROTI and VTEC.

5. The Beidou positioning method considering ionospheric error according to claim 1, characterized in that: The ionospheric residual calculation model adopts the XGBoost model based on the gradient boosting tree algorithm, which is trained using historical ionospheric residual data and multi-frequency satellite signal data to achieve fine modeling and prediction of ionospheric errors; the model is iteratively updated by continuously introducing real-time feedback data to improve the accuracy and robustness of residual correction under different ionospheric activity conditions.

6. The Beidou positioning method considering ionospheric error according to claim 1, characterized in that, The real-time positioning error feedback includes the fixation rate, error distribution, and prediction deviation.

7. A Beidou positioning method considering ionospheric error according to claim 1, characterized in that, The ionosphere-free combination filtering strategy includes: Canceling the ionospheric residual correction and instead performing positioning solution based on geometric and differential information; When the ionospheric activity returns to quiet for K consecutive epochs, the normal positioning mode is automatically restored, where K is a dynamic monitoring index.

8. A Beidou positioning method considering ionospheric error according to claim 1, characterized in that The real-time query and short-term forecasting capabilities of the ionospheric disturbance status include: Generating ionospheric disturbance forecasts through real-time monitoring and analysis of multi-modal data; optimizing the outlier determination threshold and rejection threshold in the satellite selection strategy based on the ionospheric disturbance forecasts; providing user-defined forecast periods and disturbance prediction threshold settings, and providing a historical data backtracking function.

9. A BeiDou positioning system taking ionospheric errors into account, characterized in that: The system applies a Beidou positioning method considering ionospheric errors as described in any one of claims 1 to 8 above, including: Data acquisition module: Real-time acquisition of multi-modal data and transmission thereof to the atmospheric inference large model for processing; Atmospheric inference large model module: Outputting disturbance state prediction and activity prediction based on multi-modal data; Ionospheric residual calculation module: Obtaining ionospheric residual correction and residual statistical information, and obtaining the satellite selection strategy; Real-time feedback module: Adaptive adjustment of ionospheric residual correction and satellite selection strategy in combination with real-time positioning error feedback; Positioning solution module: Performing positioning solution based on ionospheric residual correction and satellite selection strategy, and switching to the ionosphere-free combination filtering strategy for solution when the ionospheric disturbance exceeds the activity threshold; User interface module: Providing positioning results and positioning accuracy evaluation reports to the user, and supporting the user's monitoring and adjustment of the positioning service.

Images