Multi-dimensional collaborative anomaly detection method and device after satellite integer clock solution

Abstract

The invention provides a multi-dimensional collaborative anomaly detection method and device after calculation of a satellite integer clock. The method comprises the following steps: acquiring distribution characteristics of observation residual errors after calculation of the satellite integer clock; determining the stability degree of the UPD along with the time change and the whole cycle ambiguity fixed rate change condition; constructing a detection model for detecting the abnormal deviation of the satellite clock error based on the distribution characteristics of the observation residual error, the stability degree and the change condition of the integer ambiguity fixed rate; carrying out PPP floating point solution based on an integer clock product of the satellite, carrying out gross error detection, and counting the number of times that each satellite is marked as gross error; on the basis of the detection model and the number of times that each satellite is marked as gross error, constructing a PPP expansion model considering clock error deviation parameters; and performing inter-satellite single difference and integer ambiguity fixed solution on the UPD product to obtain a PPP fixed solution, and detecting a satellite integer clock with fixed abnormality. According to the invention, the problem of insufficient robustness caused by only dependence on a single observation station in the existing satellite integer clock anomaly detection method can be solved.

Description

Technical Field

[0001] This invention relates to the field of satellite navigation and positioning technology, specifically to a method and apparatus for multi-dimensional collaborative anomaly detection after satellite integer clock calculation. Background Technology

[0002] In high-precision positioning and timing using GNSS (Global Navigation Satellite System), the correct fixation of integer ambiguity parameters is crucial for achieving centimeter-level positioning accuracy. High-precision satellite integer clock products directly determine the reliability of PPP-AR (Precise Point Positioning Ambiguity Resolution) calculations. When satellite integer clocks malfunction, it often leads to a significant decrease in the fixation rate of integer parameters and deviations from Gaussian residual distribution, ultimately causing positioning calculation divergence or significant errors, severely impacting system service performance. However, utilizing multi-dimensional collaborative methods to detect satellite integer clock anomalies has become a research hotspot in the field of precise positioning.

[0003] Existing anomaly detection methods mainly rely on residual analysis or clock error sequence comparison at single stations. While these methods can identify anomalies to some extent, the results are often not robust enough due to the limitations of single-station observations caused by geometric conditions and noise. Furthermore, traditional methods are largely based on static threshold judgments, lacking comprehensive utilization of redundant observation information from multiple stations, the distribution patterns of observation residuals, the stability of UPD (uncalibrated phase delay), and the fixed rate of integer parameters. This results in insufficient statistical basis for diagnosis, easily leading to misjudgments or missed detections.

[0004] Therefore, existing satellite integer clock anomaly detection methods rely solely on a single station, resulting in insufficient robustness: traditional threshold determination lacks statistical basis, easily leading to misjudgments and missed detections, and fails to fully incorporate information such as observation residual characteristics, UPD stability, and ambiguity fixation rate during the solution process. Summary of the Invention

[0005] In view of this, it is necessary to provide a multi-dimensional collaborative anomaly detection method and device after satellite integer clock calculation, so as to solve the technical problem of insufficient robustness caused by relying on a single station in the existing satellite integer clock anomaly detection methods.

[0006] To address the aforementioned problems, in a first aspect, this invention provides a multi-dimensional collaborative anomaly detection method after satellite integer clock calculation, comprising: Obtain the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation; Determine the stability of the satellite's UPD over time, as well as the variation in integer ambiguity fixation rate; In the event of an anomaly in satellite clock error estimation, a detection model for detecting abnormal deviations in satellite clock error is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. PPP floating-point solution is performed on satellite-based integer clock products, and gross error detection is performed based on the residuals of the PPP floating-point solution. The number of times each satellite is marked as a gross error is counted. Based on the detection model and the number of times each satellite was marked as a gross error, a PPP extended model that takes into account the satellite clock bias parameters is constructed. Based on the PPP extended model, the calculated inter-satellite single-difference UPD product is used to determine the fixed value of integer ambiguity and obtain the PPP fixed solution; Satellite integer clocks that detect fixed anomalies are based on PPP fixed solutions and integer ambiguity fixed values.

[0007] In one possible implementation, the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation are obtained, including: Obtain the observation residuals of BDS / GNSS satellite integer clocks; The observed residuals are sorted according to sample size, and a test statistic is constructed based on the sorted observed residuals; Determine the approximate distribution of the test statistic under different sample sizes, and based on the approximate distribution, determine the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation.

[0008] In one possible implementation, a test statistic is constructed based on the sorted observation residuals, including: Determine the order statistic vector and covariance matrix of the expected normal distribution corresponding to the sorted observation residuals; Based on the order statistic vector and covariance matrix of the expected normal distribution, the weighting coefficients of the observation residuals are determined; Based on the weighting coefficients, the sorted observation residuals, and the mean of the sorted observation residuals, a test statistic is constructed.

[0009] In one possible implementation, determining the stationarity of the satellite's UPD over time includes: Calculate the wide lane UPD and narrow lane UPD for each satellite, and determine the standard deviation of the wide lane UPD and narrow lane UPD, as well as the difference between adjacent epochs. Based on the standard deviations of the wide-lane UPD and the narrow-lane UPD, as well as the differences between adjacent epochs, the stationarity of the wide-lane UPD over time and the stationarity of the narrow-lane UPD over time are determined respectively.

[0010] In one possible implementation, determining the integer ambiguity fixation rate variation of the satellite includes: The integer ambiguity fixation rate is determined based on the ratio of the number of epochs in which satellite integer ambiguity was successfully fixed to the total number of epochs. Based on the integer ambiguity fixation rate, cluster analysis is used to extract the spatiotemporal distribution characteristics of fixation failures, so as to quantify the correlation between the integer ambiguity fixation rate and clock anomalies and obtain the changes in the integer ambiguity fixation rate of the satellite.

[0011] In one possible implementation, upon detecting anomalies in satellite clock bias estimation, a detection model for detecting abnormal deviations in satellite clock bias is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. This model includes: In the event of an anomaly detected in the satellite clock error estimation, a multidimensional anomaly feature vector is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. Construct an integer clock anomaly diagnostic index based on multidimensional anomaly feature vectors; Based on the integer clock anomaly diagnostic index, a detection model including satellite clock bias is constructed.

[0012] In one possible implementation, PPP floating-point calculations are performed based on the satellite's integer clock products, and gross errors are detected based on the residuals from the PPP floating-point calculations. The number of times each satellite is marked as a gross error is counted, including: The PPP floating-point solution is performed on the integer clock products of the satellite to obtain the IF ambiguity floating-point solution; Based on the IF ambiguity floating-point solution, determine the post-hoc residual of the PPP floating-point solution; The post-verification residuals based on PPP floating-point calculations are used to perform gross error checks station by station to count the number of times each satellite is marked as a gross error in each station during the current epoch.

[0013] In one possible implementation, based on the detection model and the number of times each satellite is labeled as a gross error, a PPP extended model that takes into account the satellite clock bias parameters is constructed, including: Based on the order of the number of times each satellite is marked as a gross error at each station from largest to smallest, an initial PPP extended model that takes into account the satellite clock bias parameters is constructed. Based on the detection model and the initial PPP extended model, the clock bias deviation parameter is determined; If the rationality verification of the initial PPP extended model based on the clock deviation parameter is passed, the initial PPP extended model is determined to be the final PPP extended model.

[0014] In one possible implementation, based on the PPP extended model, the calculated inter-satellite single-difference UPD product is used to determine the integer ambiguity fixed value, thereby obtaining the PPP fixed solution, including: Perform inter-satellite single difference on the wide lane and narrow lane UPD products respectively to obtain the wide lane inter-satellite single difference UPD product and the narrow lane inter-satellite single difference UPD product. Calculate the wide-lane floating-point ambiguity, and based on the wide-lane floating-point ambiguity and the wide-lane inter-satellite single-difference UPD product, solve the wide-lane integer ambiguity; Based on the wide lane integer ambiguity and the narrow lane inter-satellite single-difference UPD product, the floating-point solution of the IF ambiguity is corrected, and the narrow lane ambiguity is fixed. Then, based on the PPP extended model and the wide lane integer ambiguity, the fixed IF ambiguity is recombined to obtain the fixed ambiguity solution, and the fixed PPP solution is obtained based on the fixed ambiguity solution. The IF ambiguity floating-point solution is obtained by performing PPP floating-point solution calculation on the satellite's integer clock products; based on the PPP fixed solution and the fixed integer ambiguity value, the detection of fixed anomalies in satellite integer clocks includes: Based on the fixed IF ambiguity and integer ambiguity fixed values ​​corresponding to the PPP fixed solution, the carrier phase post-hoc residual is calculated. Based on the carrier phase post-hoc residual, an inter-satellite single-difference carrier phase residual is constructed; Satellite integer clocks with fixed anomalies are detected based on the offset set of inter-satellite single-difference carrier phase residuals.

[0015] Secondly, the present invention also provides a multi-dimensional collaborative anomaly detection device after satellite integer clock calculation, comprising: The acquisition module is used to obtain the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation; The determination module is used to determine the stability of the satellite's UPD over time, as well as the change in integer ambiguity fixation rate; The first model building module is used to build a detection model for detecting abnormal deviations in satellite clock error estimation based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate when an abnormality in satellite clock error estimation is detected. The gross error statistics module is used to perform PPP floating-point calculation on satellite-based integer clock products, and to perform gross error detection based on the residuals of the PPP floating-point calculation, and to count the number of times each satellite is marked as a gross error. The second model building module is used to build a PPP extended model that takes into account the satellite clock bias parameters based on the detection model and the number of times each satellite is labeled as a gross error. The model solution module is used to determine the fixed value of integer ambiguity based on the PPP extended model and the calculated inter-satellite single-difference UPD product, so as to obtain the PPP fixed solution. An anomaly detection module is used to detect fixed anomalies in satellite integer clocks based on PPP fixed solutions and fixed integer ambiguity values.

[0016] The beneficial effects of the above implementation are as follows: By introducing multi-dimensional statistical indicators such as observation residuals, UPD, and carrier integer ambiguity fixation rate, this invention avoids the limitations of a single criterion and significantly improves the robustness and reliability of anomaly detection. Based on satellite-by-satellite UPD stability analysis, and considering the distribution characteristics of the observation residuals, the stability of the UPD over time, and the changes in the integer ambiguity fixation rate, a detection model for detecting satellite clock bias anomalies is constructed, achieving the differentiation of anomaly causes and the precise location of anomaly sources. By introducing gross error checks on the residuals of multiple stations in an independent monitoring network, the limitations of single-station observations being excessively affected by geometric conditions and noise are avoided, significantly improving the robustness and reliability of anomaly detection. Adding satellite clock bias parameter estimation to the PPP extended model allows for the direct quantification of the impact of anomalous satellites on the positioning results, thereby achieving clear identification of anomaly sources, making it more targeted than traditional residual analysis methods.

[0017] Therefore, the present invention can solve the technical problem of insufficient robustness caused by relying on only a single station in existing satellite integer clock anomaly detection methods. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart of an embodiment of the multi-dimensional collaborative anomaly detection method after satellite integer clock calculation provided by the present invention; Figure 2 A flowchart of another embodiment of the multi-dimensional collaborative anomaly detection method after satellite integer clock calculation provided by the present invention; Figure 3 This is a schematic diagram of a principle block diagram of an embodiment of the multi-dimensional collaborative anomaly detection device after satellite integer clock calculation provided by the present invention; Figure 4 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] In the description of the embodiments of this application, unless otherwise stated, "a plurality of" means two or more.

[0022] In this embodiment of the invention, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, apparatus, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product or device.

[0023] The naming or numbering of steps in the embodiments of the present invention does not mean that the steps in the method flow must be executed in the time / logical order indicated by the naming or numbering. The execution order of the named or numbered process steps can be changed according to the technical purpose to be achieved, as long as the same or similar technical effect can be achieved.

[0024] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0025] This invention provides a method and apparatus for multi-dimensional collaborative anomaly detection after satellite integer clock calculation, which will be described below.

[0026] This invention provides a multi-dimensional collaborative anomaly detection method after satellite integer clock calculation. The method can be implemented by executing an application on a terminal or server. The terminal can be a mobile phone or computer, and the server can be an edge server or a cloud server. Figure 1 As shown, the method includes: S101. Obtain the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation.

[0027] It is understandable that the distribution characteristics of the observation residuals, including the bias and standard deviation of the observation residual distribution, are used to diagnose whether the observation residuals of each satellite conform to a Gaussian distribution through normality tests, thereby identifying potentially anomalous observation sequences.

[0028] BDS stands for BeiDou satellite system, and GNSS stands for Global Navigation Satellite System.

[0029] S102. Determine the stability of the satellite's UPD over time, as well as the change in integer ambiguity fixation rate.

[0030] Understandably, satellite UPD (Uncalibrated Phase Delay) refers to a combination of errors, including uncalibrated initial phase deviation and hardware delays at the satellite end. It primarily affects GNSS (Global Navigation Satellite System) carrier phase observations, causing integer ambiguity to lose its integer characteristics, thus interfering with precise positioning (such as Precise Point Positioning, PPP).

[0031] The stability of a satellite's UPD over time can be used to evaluate whether there are any sudden changes or anomalies in UPD.

[0032] The changes in integer ambiguity fixation rate include the percentage of each satellite whose integer ambiguity parameters have been successfully fixed.

[0033] S103. In the event of an anomaly in satellite clock error estimation, a detection model for detecting abnormal deviations in satellite clock error is constructed based on the distribution characteristics of the observation residual, the stability of the UPD over time, and the change in the integer ambiguity fixation rate.

[0034] Understandably, for satellites with detected clock anomalies, a detection model incorporating satellite clock bias is further constructed. Post-test residual significance analysis is then used to further confirm whether the satellite exhibits clock anomalies. Quality markers are added to confirmed abnormal satellite clock products to facilitate identification and removal in subsequent positioning calculations.

[0035] S104. Perform PPP floating-point calculation on satellite-based integer clock products, and perform gross error detection based on the residuals of the PPP floating-point calculation, and count the number of times each satellite is marked as a gross error.

[0036] Understandably, PPP floating-point solution is essentially a positioning calculation: by accurately modeling and correcting all errors, the receiver can calculate a high-precision position using only its own observation data.

[0037] Analysis of the residuals reveals that there may be varying degrees of jumps in the visible satellite residuals for the corresponding time periods. The possible cause of this problem is that the real-time orbit and clock error products used contain gross errors during this time period.

[0038] S105. Based on the detection model and the number of times each satellite is marked as a gross error, construct a PPP extended model that takes into account the satellite clock bias parameters.

[0039] Understandably, an extended PPP model that takes into account satellite clock bias parameters is established sequentially based on the number of times each satellite is marked as a gross error at each station, from largest to smallest. The satellite clock bias is estimated and its rationality is judged by simultaneously solving the normal equation components of all stations and using the least squares adjustment method.

[0040] S106. Based on the PPP extended model, the calculated inter-satellite single-difference UPD product is used to determine the fixed value of integer ambiguity, and the PPP fixed solution is obtained.

[0041] Understandably, based on satellite clock difference anomaly information, the inter-satellite single-difference narrow-lane carrier integer offset quantum set is optimized and selected. The LAMBDA search algorithm is used to first fix some integer parameters, solve and update other integer parameters, until no more integer parameters can be fixed, and then the inter-satellite single-difference UPD product and the fixed value of integer ambiguity are solved.

[0042] S107. Satellite integer clocks that detect fixed anomalies based on PPP fixed solutions and integer ambiguity fixed values.

[0043] It is understandable that, based on the inter-satellite single-difference UPD product and the fixed integer ambiguity value, the spatiotemporal distribution of satellites with PPP-AR (Precise Point Positioning) ambiguity fixation failure and post-verification residuals in the monitoring network are analyzed. Taking into account factors such as elevation angle, floating-point ambiguity accuracy, and ratio value, the fixed anomaly satellite integer clock is detected.

[0044] To address the shortcomings of existing satellite integer clock anomaly detection methods, such as insufficient robustness due to reliance on a single station, the lack of statistical basis for traditional threshold determination leading to misjudgments and missed detections, and the failure to fully integrate information such as observation residual characteristics, UPD stability, and ambiguity fixation rate during the solution process, this invention proposes a multi-dimensional collaborative anomaly detection method after satellite integer clock solution. This method improves the real-time performance, robustness, and accuracy of anomaly detection by employing multi-station residual distribution characteristic diagnosis, wide / narrow lane UPD time stability and cross-receiver consistency verification, spatiotemporal distribution analysis of ambiguity fixation rate, and cross-validation using the PPP extended model and inter-satellite single-difference ambiguity fixation and residual statistical characteristics. This ensures the stability and reliability of BDS / GNSS precise positioning services.

[0045] In some embodiments, obtaining the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation includes: Obtain the observation residuals of BDS / GNSS satellite integer clocks; The observed residuals are sorted according to sample size, and a test statistic is constructed based on the sorted observed residuals; Determine the approximate distribution of the test statistic under different sample sizes, and based on the approximate distribution, determine the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation.

[0046] Understandably, the Shapiro-Wilk statistical hypothesis test can be used to determine whether the residuals conform to a Gaussian distribution. Suppose a satellite... The observation residuals after integer clock estimation of BDS / GNSS satellites at each epoch are: Its null hypothesis The residuals follow a normal distribution. The statistic is calculated as follows. Then, approximate distributions of the sample under different sample sizes are obtained through numerical fitting or table lookup. .like Less than the significance level If the value is 0.05, then the null hypothesis is rejected. If the satellite residual does not conform to a Gaussian distribution, an anomaly marker is triggered.

[0047] In some embodiments, a test statistic is constructed based on the sorted observation residuals, including: Determine the order statistic vector and covariance matrix of the expected normal distribution corresponding to the sorted observation residuals; Based on the order statistic vector and covariance matrix of the expected normal distribution, the weighting coefficients of the observation residuals are determined; Based on the weighting coefficients, the sorted observation residuals, and the mean of the sorted observation residuals, a test statistic is constructed.

[0048] Understandably, the specific steps are as follows: Sort the samples in ascending order. The residuals of the sorted samples are: .

[0049] Construct the test statistic :

[0050] In the formula: The coefficient is the mean of the observed residuals. The order statistic vector of the expected normal distribution and its covariance matrix The calculation yielded:

[0051] Where: Expected order statistic vector With covariance matrix All are predetermined by the standard normal distribution and depend only on the sample size. This is unrelated to the actual observed residual data. Calculated from the quantiles of the standard normal distribution. It is obtained by calculation or numerical simulation of the joint distribution of the corresponding order statistics.

[0052] In some embodiments, determining the stationarity of a satellite's UPD over time includes: Calculate the wide lane UPD and narrow lane UPD for each satellite, and determine the standard deviation of the wide lane UPD and narrow lane UPD, as well as the difference between adjacent epochs. Based on the standard deviations of the wide-lane UPD and the narrow-lane UPD, as well as the differences between adjacent epochs, the stationarity of the wide-lane UPD over time and the stationarity of the narrow-lane UPD over time are determined respectively.

[0053] Understandably, the wide-lane UPD and narrow-lane UPD are calculated for each satellite using the following formula:

[0054]

[0055] In the formula: and These are combined observations of wide and narrow lanes (in meters); and Corresponding to two frequencies; and These are the wide-lane wavelength and the narrow-lane wavelength, respectively. It is the geometric distance; It is a tilted tropospheric delay; and These are the ionospheric delays caused by the combination of wide and narrow alleyways; and These are the wide alley ambiguity and the narrow alley ambiguity, respectively. and It measures noise.

[0056] Assess the stationarity of UPD over time: Suppose a certain satellite In the calendar The calculated UPD for wide and narrow lanes is The time series is generated, and the mean and standard deviation are calculated using the following formulas:

[0057]

[0058] The above shows the mean and standard deviation of UPD for wide lanes; the same applies to narrow lanes. If... A small value indicates consistent UPD across receivers; a significant increase indicates a possible UPD mutation or anomaly in the satellite.

[0059] To directly measure the jump in UPD over time, the difference between adjacent epochs can be calculated:

[0060] If the following conditions are met:

[0061] This is defined as a UPD mutation, where the threshold is... An empirical threshold can be used.

[0062] In some embodiments, determining the integer ambiguity fixation rate variation of the satellite includes: The integer ambiguity fixation rate is determined based on the ratio of the number of epochs in which satellite integer ambiguity was successfully fixed to the total number of epochs. Based on the integer ambiguity fixation rate, cluster analysis is used to extract the spatiotemporal distribution characteristics of fixation failures, so as to quantify the correlation between the integer ambiguity fixation rate and clock anomalies and obtain the changes in the integer ambiguity fixation rate of the satellite.

[0063] Understandably, further statistical analysis can be conducted to determine the successful fixation rate of integer ambiguity parameters for each satellite, and cluster analysis can be used to extract the spatiotemporal distribution characteristics of fixation failures, thereby quantifying the correlation between the integer ambiguity fixation rate and clock anomalies.

[0064] Calculate the integer ambiguity fixation rate:

[0065] In the formula: For satellite The number of epochs in which the ambiguity is successfully fixed throughout the entire cycle. This represents the total epoch number.

[0066] Cluster analysis was used to cluster epochs where integer ambiguity fixation failed. First, vectorized features are constructed, using time, receiver location, and satellite number as multi-dimensional features. For each failure point, the following can be used:

[0067] in, For epoch index, The latitude and longitude coordinates of the station. It is the satellite PRN code.

[0068] The DBSCAN clustering method is used to identify concentrated bursts of fixed failures. If a failure cluster mainly spans multiple receivers and corresponds to the same satellite, it is determined that the satellite has a systemic anomaly; if the failure cluster is limited to a single receiver, it is determined to be local interference, thus achieving effective differentiation of the causes of anomalies.

[0069] In some embodiments, upon detecting anomalies in satellite clock bias estimation, a detection model for detecting abnormal deviations in satellite clock bias is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. This model includes: In the event of an anomaly detected in the satellite clock error estimation, a multidimensional anomaly feature vector is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. Construct an integer clock anomaly diagnostic index based on multidimensional anomaly feature vectors; Based on the integer clock anomaly diagnostic index, a detection model including satellite clock bias is constructed.

[0070] Understandably, when anomalies in satellite clock bias estimation are detected, the distribution characteristics of observation residuals, the time stationarity of UPD, and the changes in ambiguity fixation rate are extracted to construct a multidimensional anomaly feature vector. A comprehensive diagnosis is then performed using methods such as hypothesis testing, cluster analysis, support vector machines, and neural networks to extract anomaly feature patterns. Furthermore, by combining time series abrupt change point detection with spatial correlation analysis, it is determined whether the anomaly is caused by systematic clock bias or by local interference at a single station, thereby effectively eliminating false anomalies and improving the reliability of the diagnostic results.

[0071] By integrating indicators such as the magnitude and distribution characteristics of observation residuals, UPD consistency, and ambiguity fixation rate, a hierarchical weighted scoring model is established, and an integer-clock anomaly diagnostic index is designed. This index can quantify the contribution of different dimensional features to anomaly judgment, and its effectiveness and robustness are verified through experimental data.

[0072] In some embodiments, PPP floating-point solutions are performed based on the satellite's integer clock products, and gross errors are detected based on the residuals from the PPP floating-point solutions. The number of times each satellite is marked as a gross error is counted, including: The PPP floating-point solution is performed on the integer clock products of the satellite to obtain the IF ambiguity floating-point solution; Based on the IF ambiguity floating-point solution, determine the post-hoc residual of the PPP floating-point solution; The post-verification residuals based on PPP floating-point calculations are used to perform gross error checks station by station to count the number of times each satellite is marked as a gross error in each station during the current epoch.

[0073] Understandably, the ambiguity is corrected and fixed based on the wide-lane UPD product. A PPP floating-point solution (positioning solution) is then performed using a real-time integer clock product to obtain the IF ambiguity floating-point solution. After fixing the narrow-lane ambiguity, the IF ambiguity is reorganized and constrained in the normal equations to obtain a fixed ambiguity solution.

[0074] Gross errors are checked station by station using the post-validation residuals of PPP floating-point calculations to identify anomalous observations and reduce their weights to eliminate positioning effects. Then, the number of times each satellite is marked as a gross error in each station in the current epoch is counted. The more times it is marked, the higher the probability that the satellite has clock anomalies.

[0075] In satellite positioning, especially in Precise Point Positioning (PPP), IF ambiguity refers to the carrier phase ambiguity corresponding to the ionosphere-free combination.

[0076] In some embodiments, based on the detection model and the number of times each satellite is labeled as a gross error, a PPP extended model that takes into account satellite clock bias parameters is constructed, including: Based on the order of the number of times each satellite is marked as a gross error at each station from largest to smallest, an initial PPP extended model that takes into account the satellite clock bias parameters is constructed. Based on the detection model and the initial PPP extended model, the clock bias deviation parameter is determined; If the rationality verification of the initial PPP extended model based on the clock deviation parameter is passed, the initial PPP extended model is determined to be the final PPP extended model.

[0077] Understandably, based on the number of times each satellite is marked as a gross error at each station obtained above, a PPP extended model taking into account the satellite clock bias parameter is established in descending order. The satellite clock bias is estimated and its rationality is judged by simultaneously solving the normal equation components of all stations and using the least squares adjustment method.

[0078] Specifically, the weighted sum of squares of the post-test residuals of all stations and the 3D orientation positioning error distribution are recalculated and compared with the results before estimating the satellite clock bias parameter. An empirical threshold is used to judge the rationality of the introduction of the satellite clock bias parameter.

[0079] In some embodiments, based on the PPP extended model, the calculated inter-satellite single-difference UPD product is used to determine the integer ambiguity fixed value to obtain the PPP fixed solution, including: Perform inter-satellite single difference on the wide lane and narrow lane UPD products respectively to obtain the wide lane inter-satellite single difference UPD product and the narrow lane inter-satellite single difference UPD product. Calculate the wide-lane floating-point ambiguity, and based on the wide-lane floating-point ambiguity and the wide-lane inter-satellite single-difference UPD product, solve the wide-lane integer ambiguity; Based on the wide lane integer ambiguity and the narrow lane inter-satellite single-difference UPD product, the floating-point solution of the IF ambiguity is corrected, and the narrow lane ambiguity is fixed. Then, based on the PPP extended model and the wide lane integer ambiguity, the fixed IF ambiguity is recombined to obtain the fixed ambiguity solution, and the fixed PPP solution is obtained based on the fixed ambiguity solution. The IF ambiguity floating-point solution is obtained by performing PPP floating-point solution calculation on the satellite's integer clock products; based on the PPP fixed solution and the fixed integer ambiguity value, the detection of fixed anomalies in satellite integer clocks includes: Based on the fixed IF ambiguity and integer ambiguity fixed values ​​corresponding to the PPP fixed solution, the carrier phase post-hoc residual is calculated. Based on the carrier phase post-hoc residual, an inter-satellite single-difference carrier phase residual is constructed; Satellite integer clocks with fixed anomalies are detected based on the offset set of inter-satellite single-difference carrier phase residuals.

[0080] Understandably, based on satellite clock difference anomaly information, the inter-satellite single-difference narrow lane carrier integer offset quantum set is optimized and selected. The LAMBDA search algorithm is used to first fix some integer ambiguities, and then solve and update other integer ambiguities until no more integer ambiguities can be fixed.

[0081] In some embodiments, such as Figure 2 As shown, the method provided by the present invention includes: The first step is to analyze the statistical characteristics of the post-hoc residuals of BDS / GNSS satellite integer clock estimation.

[0082] Calculate the skewness and standard deviation of the observed residual distribution.

[0083] Suppose a certain satellite is in The observation residuals after integer clock estimation of BDS / GNSS satellites at each epoch are: Calculate the skewness of the observed residual distribution. and standard deviation The formula is as follows:

[0084]

[0085] In the formula: To observe the mean of the residuals, It reflects the symmetry of the residual distribution relative to the normal distribution. It reflects the degree of dispersion of the residual distribution.

[0086] By using the normality test method, we can diagnose whether the observation residuals of each satellite conform to a Gaussian distribution, thereby identifying potentially anomalous observation sequences.

[0087] To determine whether the residuals conform to a Gaussian distribution, the Shapiro-Wilk statistical hypothesis test can be used. Suppose a satellite... The observation residuals after integer clock estimation of BDS / GNSS satellites at each epoch are: Its null hypothesis The residuals are assumed to follow a normal distribution. The specific steps are as follows: Sort the samples in ascending order. The residuals of the sorted samples are: .

[0088] Construct the test statistic :

[0089] In the formula: The coefficient is the mean of the observed residuals. The order statistic vector of the expected normal distribution and its covariance matrix The calculation yielded:

[0090] Where: Expected order statistic vector With covariance matrix All are predetermined by the standard normal distribution and depend only on the sample size. This is unrelated to the actual observed residual data. Calculated from the quantiles of the standard normal distribution. It is obtained by calculation or numerical simulation of the joint distribution of the corresponding order statistics.

[0091] Hypothesis testing: Calculate the statistic Then, approximate distributions of the sample under different sample sizes are obtained through numerical fitting or table lookup. .like Less than the significance level If the value is 0.05, then the null hypothesis is rejected. If the satellite residual does not conform to a Gaussian distribution, an anomaly marker is triggered.

[0092] The second step is to analyze the stability and consistency of UPD.

[0093] The wide-lane UPD and narrow-lane UPD are calculated for each satellite using the following formula:

[0094]

[0095] In the formula: and These are combined observations of wide and narrow lanes (in meters); and Corresponding to two frequencies; and These are the wide-lane wavelength and the narrow-lane wavelength, respectively. It is the geometric distance; It is a tilted tropospheric delay; and These are the ionospheric delays caused by the combination of wide and narrow alleyways; and These are the wide alley ambiguity and the narrow alley ambiguity, respectively. and It measures noise.

[0096] Assess the stationarity of UPD over time: Suppose a certain satellite In the calendar The calculated UPD for wide and narrow lanes is The time series is generated, and the mean and standard deviation are calculated using the following formulas:

[0097]

[0098] The above shows the mean and standard deviation of UPD for wide lanes; the same applies to narrow lanes. If... A small value indicates consistent UPD across receivers; a significant increase indicates a possible UPD mutation or anomaly in the satellite.

[0099] To directly measure the jump in UPD over time, the difference between adjacent epochs can be calculated:

[0100] If the following conditions are met:

[0101] This is defined as a UPD mutation, where the threshold is... An empirical threshold can be used.

[0102] Further statistical analysis was conducted on the successful fixation rate of integer ambiguity parameters for each satellite. Cluster analysis was used to extract the spatiotemporal distribution characteristics of fixation failures, and the correlation between the integer ambiguity fixation rate and clock anomalies was quantified.

[0103] Calculate the integer ambiguity fixation rate:

[0104] In the formula: For satellite The number of epochs in which the ambiguity is successfully fixed throughout the entire cycle. This represents the total epoch number.

[0105] Cluster analysis was used to cluster epochs where integer ambiguity fixation failed.

[0106] First, vectorized features are constructed, using time, receiver location, and satellite number as multi-dimensional features. For each failure point, the following can be used:

[0107] in, For epoch index, The latitude and longitude coordinates of the station. It is a satellite PRN code, which is a pseudo-random noise code.

[0108] The DBSCAN clustering method is used to identify concentrated bursts of fixed failures. If a failure cluster mainly spans multiple receivers and corresponds to the same satellite, it is determined that the satellite has a systemic anomaly; if the failure cluster is limited to a single receiver, it is determined to be local interference, thus achieving effective differentiation of the causes of anomalies.

[0109] The third step involves extracting the distribution characteristics of the observation residuals, the time stationarity of UPD, and the changes in ambiguity fixation rate when anomalies in satellite clock bias estimation are detected, to construct a multidimensional anomaly feature vector. A comprehensive diagnosis is then performed using methods such as hypothesis testing, cluster analysis, support vector machines, and neural networks to extract anomaly feature patterns. Further, by combining time series abrupt change point detection with spatial correlation analysis, it is determined whether the anomaly is caused by systematic clock bias or by local interference at a single station, thereby effectively eliminating false anomalies and improving the reliability of the diagnostic results.

[0110] The fourth step involves integrating indicators such as the magnitude and distribution characteristics of observation residuals, UPD consistency, and ambiguity fixation rate to establish a hierarchical weighted scoring model and design an integer-clock anomaly diagnostic index. This index can quantify the contribution of different dimensional features to anomaly judgment, and its effectiveness and robustness are verified through experimental data.

[0111] The fifth step involves constructing a detection model that incorporates satellite clock bias for satellites with detected clock anomalies. Post-hoc residual significance analysis is then used to further confirm the existence of clock anomalies in the satellites. Quality markers are added to confirmed abnormal satellite clock products to facilitate identification and removal in subsequent positioning calculations.

[0112] Step 6: Correct and fix the wide-lane ambiguity based on the wide-lane UPD product. Perform PPP floating-point solution calculation (positioning solution) using a real-time integer clock product to obtain the IF ambiguity floating-point solution. After fixing the narrow-lane ambiguity, reorganize and fix the IF ambiguity and constrain it in the normal equation to obtain the fixed ambiguity solution.

[0113] Building GNSS on satellite With receiver The original pseudorange observation model and the original carrier phase observation model between:

[0114]

[0115]

[0116]

[0117] In the formula: , and , These are pseudorange and carrier phase observations at two frequencies (in meters). It is the geometric distance; It's the speed of light; It is the receiver clock bias; It is satellite clock bias; It is a tilted tropospheric delay; It is the tilted ionospheric delay on the main frequency; This indicates the ratio of the main frequency to the secondary frequency; and These are wavelengths of two frequencies; and It is the carrier integer offset; , and , These are receiver-side hardware delay and UPD, respectively. , and , These are satellite-side hardware latency and UPD; and It is pseudorange measurement noise; and It is carrier phase measurement noise.

[0118] A dual-frequency ionosphere-free (IF) combined observation model was constructed based on the original observation model:

[0119]

[0120] In the formula: and These are the IF combined pseudorange observations and carrier phase observations, respectively; and There are two frequencies; It is the geometric distance; It's the speed of light; It is the receiver clock bias; It is satellite clock bias; and These are the combined IF hardware delay deviations at the receiver and satellite ends, respectively. and These are the IF combination UPDs at the receiver end and the satellite end, respectively. It is the tropospheric delay along the propagation path; It is the carrier phase IF combined carrier integer offset; It is the observation noise of the pseudorange IF combination; This is the observation noise of the carrier phase IF combination. Specifically, the floating-point ambiguity of the IF combination... Determine using a clock error estimation model for ambiguity floating-point solutions:

[0121]

[0122] In the formula: , , , , The station coordinates and satellite orbits are known values. This represents the OMC value of the IF pseudorange observation; This represents the OMC value observed during IF carrier phase observation; It is a wet mapping function of tropospheric delay; It is the wet delay of the zenith troposphere; It is the carrier phase IF combination integer offset; and These are the IF hardware delay deviations at the receiver and satellite ends, respectively. and These are the IF combination UPDs at the receiver and satellite ends, respectively.

[0123] Based on the UPD product correction and fixed wide lane (WL) ambiguity, the integer WL ambiguity can be estimated using MW combination, and a MW combination model is established:

[0124] In the formula: It is floating-point WL ambiguity; It is the blurriness of WL throughout the week; It is the WL wavelength; and These are WLUPD at the receiver end and the satellite end, respectively.

[0125] After the integer WL ambiguity is recovered, the floating-point narrow lane (NL) ambiguity can be derived using the IF carrier integer offset, and the NL ambiguity can be fixed using the UPD product:

[0126] In the formula: It is floating-point NL ambiguity; It is the NL wavelength; It is the wavelength of the IF combination; and These are the receiver-side and satellite-side NL UPDs, respectively. 4. By recombining the fixed WL fuzziness and NL fuzziness and constraining them in the normal equation, a fixed fuzziness solution is obtained.

[0127] The relationships between IF ambiguity, WL ambiguity, and NL ambiguity are as follows:

[0128] In the formula: For IF wavelength; It is the carrier phase IF combination ambiguity; WL ambiguity; For NL wavelength; For NL ambiguity.

[0129] Based on the relationship between the ambiguities of IF, WL, and NL combinations, the clock error model for IF combinations with fixed ambiguities is as follows:

[0130] In the formula: and It is a clock error parameter with fixed ambiguity; It is the carrier phase IF combination ambiguity; and These are the IF combination UPDs at the receiver and satellite ends, respectively.

[0131] The seventh step involves using the post-verification residuals calculated by the PPP floating-point solution to perform gross error checks station by station to identify abnormal observations and reduce their weights to eliminate positioning effects. Then, the number of times each satellite is marked as a gross error in each station during the current epoch is counted. The more times it is marked, the higher the probability that the satellite has clock anomalies.

[0132] Step 8: Based on the number of times each satellite is marked as a gross error at each station obtained in the above steps, establish an extended PPP model that takes into account the satellite clock bias parameter in descending order. Estimate the satellite clock bias and judge its rationality by simultaneously solving the normal equation components of all stations and using the least squares adjustment method.

[0133] By combining the components of the normal equations from all stations, a system is established that takes into account the satellite. Extended PPP model for clock bias parameters:

[0134]

[0135] In the formula: subscript Indicates the index of the station. This is the residual vector. For observation vectors; It is the design matrix of station parameters; It is a vector of station parameters, namely station coordinates, receiver clock error, tropospheric wet delay, and carrier integer offset; It is a satellite The design matrix for clock error parameters; It is a satellite Estimates of the clock error parameters. Assuming observations are uncorrelated, the individual weight matrix... It is a diagonal matrix, therefore the entire weight matrix It is also a diagonal matrix.

[0136] Satellite estimation using least squares adjustment Clock bias parameters:

[0137] An estimate can be derived by eliminating the station parameters. The normal equation:

[0138] In the formula: .

[0139] The above formula shows that the estimation The components of the normal equation can be calculated separately for each station and then accumulated together for the final estimate.

[0140] The weighted sum of squares of the post-test residuals of all stations and the 3D orientation positioning error distribution were recalculated and compared with the results before estimating the satellite clock bias parameter. An empirical threshold was used to judge the rationality of the introduction of the satellite clock bias parameter.

[0141] The ninth step involves optimizing the selection of inter-satellite single-difference narrow-lane carrier integer offset quantum sets based on satellite clock difference anomaly information. The LAMBDA search algorithm is then used to first fix some integer ambiguities, and other integer ambiguities are solved and updated until no more integer ambiguities can be fixed.

[0142] Constructing an inter-satellite single-difference carrier integer offset quantum set and solving the inter-satellite single-difference UPD product From the above IF observation model, it can be concluded that the IF combination includes hardware delay biases at both the receiver and satellite ends. In this case, the influence of the receiver hardware delay bias can be eliminated using the inter-satellite single-difference method. Therefore, when fixing the PPP integer parameters, only the satellite hardware delay bias needs to be considered. The inter-satellite single-difference observation model is as follows:

[0143]

[0144] In the formula: and Each represents a satellite, among which As a reference satellite, and satisfying .

[0145] Given the IF combined ambiguity and combinations that can be decomposed into WL ambiguity and NL ambiguity:

[0146] In the formula:

[0147]

[0148] and These are the floating-point solutions for the ambiguity of the WL combination and the NL combination, respectively. and These are the actual ambiguities for the WL and NL combinations, respectively. , These are the UPDs for the WL combination and NL combination at the receiver end, respectively; , These are UPDs for the satellite-side WL combination and NL combination, respectively.

[0149] The model after single difference is:

[0150] In the formula:

[0151]

[0152] This shows that the key to fixing the PPP carrier integer offset lies in separating the integer ambiguity from the satellite-end hardware delay deviation. Since the hardware delay has short-term stability, it can be decomposed into a constant integer part and a time-varying fractional part. Because the integer part does not destroy the integer property of the ambiguity, it can be absorbed by the ambiguity. The remaining uncalibrated fractional part of the phase delay, UPD, is estimated at the server end.

[0153] When estimating the inter-satellite single-difference UPD of WL and NL, the equation is rank deficient; therefore, the UPD of the reference satellite needs to be set to 0 during estimation. Assuming there are n stations in the reference network, each station can observe m satellites. One satellite is selected as the reference satellite, and an inter-satellite single-difference is constructed to eliminate the influence of receiver-end UPD. The observation equation for solving the inter-satellite single-difference UPD is as follows:

[0154] In the formula: It is an m-1 order identity matrix; It is an m-1 dimensional column vector. After solving for the single difference UPD of each satellite, the average value of multiple stations is taken to obtain a stable UPD product.

[0155] Obtaining a fixed ambiguity solution: Since the wavelength of WL is relatively long (0.86m), it is relatively easy to fix. Therefore, the WL ambiguity is fixed first. The WL ambiguity can be obtained from the combined MW observations. The combined MW observation model is as follows:

[0156] The WL ambiguity values ​​obtained from the MW combination are smoothed using a multi-epoch averaging method. This smoothing process eliminates the effects of pseudorange noise and multipath error. Let <> be the inter-epoch smoothing operator. The smoothed WL ambiguity is represented by the following relationship:

[0157]

[0158] In the formula: The standard deviation of WL ambiguity; This represents the epoch number for smoothing. After smoothing, the satellite with the highest elevation angle is selected as the reference satellite. The WL ambiguity of each satellite is then single-differentially analyzed, and the resulting single-difference ambiguity is expressed by the following formula:

[0159]

[0160] The WL ambiguity after inter-satellite subtraction eliminates the influence of receiver-side UPD, leaving only integer ambiguity. and satellite-side UPD parameters Since the WL combination retains the integer characteristic of ambiguity, it can be directly rounded to obtain the single-difference integer ambiguity. Then, the single-difference WL ambiguity is corrected using the satellite-side single-difference UPD product provided by the server. After correction, it is rounded to the nearest integer to obtain the integer value of the WL ambiguity. The formula is as follows:

[0161] To verify the correctness of the rounding operation, the continuous deviation fixed method needs to be used. The specific formula is as follows:

[0162]

[0163] In the formula: This indicates the WL ambiguity after UPD correction; This indicates the WL ambiguity after rounding down to the nearest integer. for The mean error. If A value greater than 0.999 indicates successful fixing; otherwise, fixing fails.

[0164] Calculate the fixed value of WL ambiguity Substituting this into the following formula, the floating-point value of the single-difference NL ambiguity can be obtained:

[0165] The obtained NL ambiguity floating-point value is corrected by UPD, and then fixed to an integer by the Least-Squares Ambiguity Decorrelation Adjustment (LAMBDA) algorithm. The threshold for ratio test is set to 3.

[0166] If WL and NL are ambiguous and If all are successfully fixed, then substitute them into the following formula to obtain the fixed solution for the single-difference IF combination ambiguity:

[0167] After obtaining the fixed solution, the UPD of the single difference NL needs to be added and then used as a constraint condition to re-perform Kalman filtering for parameter estimation, so as to realize the solution of the PPP fixed solution.

[0168] Step 10: Analyze the spatiotemporal distribution of satellites that failed to fix PPP-AR ambiguity in the monitoring network and the residuals after verification. Taking into account factors such as elevation angle, floating-point ambiguity accuracy, and ratio value, detect satellite integer clocks that fix anomalies.

[0169] Compared with the prior art, the main advantages of the technical solution of the present invention are as follows: Multidimensional feature fusion enhances diagnostic robustness: By introducing multidimensional statistical indicators such as observation residuals, UPD, and carrier integer offset fixation rate, the limitations of a single criterion are avoided, significantly improving the robustness and reliability of anomaly detection.

[0170] Combining UPD mutations with fixed rates enables more accurate anomaly localization: Based on satellite-by-satellite UPD stability analysis, and combined with the spatiotemporal clustering characteristics of carrier integer offset fixed rates, the causes of anomalies can be distinguished and the sources of anomalies can be accurately located.

[0171] Intelligent diagnostic methods enhance discrimination capabilities: By employing methods such as hypothesis testing, cluster analysis, support vector machines, and neural networks to extract multidimensional abnormal features, high detection accuracy can be maintained even in complex interference environments.

[0172] By introducing gross error verification of residuals from multiple stations in an independent monitoring network, the limitations of single-station observations being excessively affected by geometric conditions and noise are avoided, significantly improving the robustness and reliability of anomaly detection.

[0173] Adding satellite clock bias parameter estimation to the PPP extended model can directly quantify the impact of anomalous satellites on positioning results, thereby enabling clear identification of the source of the anomaly, which is more targeted than traditional residual analysis methods.

[0174] By combining the wide-lane and narrow-lane ambiguity constraints in the PPP-AR ambiguity fixing process, and using inter-satellite single-difference UPD modeling and fixed solution verification, an independent reference is provided for the determination of integer clock anomalies, thus improving the accuracy of anomaly diagnosis.

[0175] like Figure 3 As shown, the present invention also provides a multi-dimensional collaborative anomaly detection device 300 after satellite integer clock calculation, comprising: The acquisition module 301 is used to acquire the distribution characteristics of the observation residuals after the integer clock calculation of BDS / GNSS satellites; The determination module 302 is used to determine the stability of the satellite's UPD over time and the change in integer ambiguity fixation rate; The first model construction module 303 is used to construct a detection model for detecting abnormal deviations in satellite clock error estimation based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate when an abnormality in satellite clock error estimation is detected. The gross error statistics module 304 is used to perform PPP floating-point solution calculation on satellite-based integer clock products, and to perform gross error detection based on the residuals of the PPP floating-point solution calculation, and to count the number of times each satellite is marked as a gross error. The second model construction module 305 is used to construct a PPP extended model that takes into account the satellite clock bias parameters based on the detection model and the number of times each satellite is labeled as a gross error. Model solution module 306 is used to determine the fixed value of integer ambiguity based on the PPP extended model and the calculated inter-satellite single difference UPD product, so as to obtain the PPP fixed solution. Anomaly detection module 307 is used to detect fixed anomalies in satellite integer clocks based on PPP fixed solution and integer ambiguity fixed value.

[0176] The satellite integer clock multi-dimensional collaborative anomaly detection device provided in the above embodiments can realize the technical solutions described in the above embodiments of the satellite integer clock multi-dimensional collaborative anomaly detection method. The specific implementation principles of each module or unit can be found in the corresponding content in the above embodiments of the satellite integer clock multi-dimensional collaborative anomaly detection method, which will not be repeated here.

[0177] like Figure 4 As shown, the present invention also provides an electronic device 400. The electronic device 400 includes a processor 401, a memory 402, and a display 403. Figure 4 Only some components of the electronic device 400 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0178] In some embodiments, memory 402 may be an internal storage unit of electronic device 400, such as a hard disk or memory of electronic device 400. In other embodiments, memory 402 may also be an external storage device of electronic device 400, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 400.

[0179] Furthermore, the memory 402 may include both internal storage units of the electronic device 400 and external storage devices. The memory 402 is used to store application software and various types of data installed on the electronic device 400.

[0180] In some embodiments, processor 401 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 402 or process data, such as the multi-dimensional collaborative anomaly detection method after satellite integer clock calculation in this invention.

[0181] In some embodiments, display 403 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 403 is used to display information from electronic device 400 and to display a visual user interface. Components 401-403 of electronic device 400 communicate with each other via a system bus.

[0182] In some embodiments of the present invention, when the processor 401 executes the multi-dimensional collaborative anomaly detection program after satellite integer clock calculation in the memory 402, the following steps can be implemented: Obtain the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation; Determine the stability of the satellite's UPD over time, as well as the variation in integer ambiguity fixation rate; In the event of an anomaly in satellite clock error estimation, a detection model for detecting abnormal deviations in satellite clock error is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. PPP floating-point solution is performed on satellite-based integer clock products, and gross error detection is performed based on the residuals of the PPP floating-point solution. The number of times each satellite is marked as a gross error is counted. Based on the detection model and the number of times each satellite was marked as a gross error, a PPP extended model that takes into account the satellite clock bias parameters is constructed. Based on the PPP extended model, the calculated inter-satellite single-difference UPD product is used to determine the fixed value of integer ambiguity and obtain the PPP fixed solution; Satellite integer clocks that detect fixed anomalies are based on PPP fixed solutions and integer ambiguity fixed values.

[0183] It should be understood that when the processor 401 executes the multi-dimensional collaborative anomaly detection program after satellite integer clock calculation in the memory 402, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0184] Furthermore, the embodiments of the present invention do not specifically limit the type of electronic device 400 mentioned. Electronic device 400 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, electronic device 400 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0185] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a multi-dimensional collaborative anomaly detection method after satellite integer clock calculation provided by the methods described above, the method comprising: Obtain the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation; Determine the stability of the satellite's UPD over time, as well as the variation in integer ambiguity fixation rate; In the event of an anomaly in satellite clock error estimation, a detection model for detecting abnormal deviations in satellite clock error is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. PPP floating-point solution is performed on satellite-based integer clock products, and gross error detection is performed based on the residuals of the PPP floating-point solution. The number of times each satellite is marked as a gross error is counted. Based on the detection model and the number of times each satellite was marked as a gross error, a PPP extended model that takes into account the satellite clock bias parameters is constructed. Based on the PPP extended model, the calculated inter-satellite single-difference UPD product is used to determine the fixed value of integer ambiguity and obtain the PPP fixed solution; Satellite integer clocks that detect fixed anomalies are based on PPP fixed solutions and integer ambiguity fixed values.

[0186] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0187] The above provides a detailed description of the multi-dimensional collaborative anomaly detection method and device after satellite integer clock calculation provided by the present invention. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A satellite integer clock solution post-multiple dimension collaborative anomaly detection method, characterized in that, The method comprises the following steps: obtaining the distribution characteristics of the observation residuals after the integer clock solution of the BDS / GNSS satellites; determining the smoothness of the UPD of the satellites over time and the change of the integer ambiguity fixing rate; in the case of detecting abnormal satellite clock error estimation, constructing a detection model for detecting abnormal deviation of the satellite clock error based on the distribution characteristics of the observation residuals, the smoothness of the UPD over time and the change of the integer ambiguity fixing rate; performing PPP float solution based on the integer clock product of the satellites, performing rough error detection based on the PPP float solution residual, and counting the number of times each satellite is marked as a rough error; constructing a PPP extension model considering the satellite clock error deviation parameter based on the detection model and the number of times each satellite is marked as a rough error; determining the integer ambiguity fixing value based on the PPP extension model and the calculated inter-satellite single-difference UPD product, and obtaining the PPP fixed solution; detecting the fixed abnormal satellite integer clock based on the PPP fixed solution and the integer ambiguity fixing value.

2. The method according to claim 1, wherein, The method comprises the following steps: obtaining the distribution characteristics of the observation residuals after the integer clock solution of the BDS / GNSS satellites; obtaining the observation residuals of the BDS / GNSS satellite integer clock; sorting the observation residuals according to the sample size, and constructing a test statistic based on the sorted observation residuals; 3. The method according to claim 2, wherein, determining the approximate distribution of the test statistic under different sample sizes, and determining the distribution characteristics of the observation residuals after the integer clock solution of the BDS / GNSS satellites based on the approximate distribution. The method comprises the following steps: determining the order statistics vector and the covariance matrix of the expected normal distribution corresponding to the sorted observation residuals; determining the weighting coefficient of the observation residuals based on the order statistics vector and the covariance matrix of the expected normal distribution; 4. The method of claim 1, wherein the method further comprises: constructing a test statistic based on the weighting coefficient, the sorted observation residuals, and the mean of the sorted observation residuals. The method comprises the following steps: calculating the wide-lane UPD and the narrow-lane UPD of each satellite, and determining the standard deviation of the wide-lane UPD and the narrow-lane UPD and the difference between adjacent epochs respectively; 5. The method of claim 1, wherein the method further comprises: determining the smoothness of the wide-lane UPD over time and the smoothness of the narrow-lane UPD over time based on the standard deviation of the wide-lane UPD and the narrow-lane UPD and the difference between adjacent epochs respectively. The method comprises the following steps: determining the integer ambiguity fixing rate based on the ratio of the number of epochs in which the satellite integer ambiguity is successfully fixed to the total number of epochs; 6. The method of claim 1, wherein the method further comprises: extracting the spatio-temporal distribution characteristics of the fixing failure by cluster analysis based on the integer ambiguity fixing rate, quantifying the correlation degree between the integer ambiguity fixing rate and the clock error anomaly, and obtaining the change of the integer ambiguity fixing rate of the satellite. In the case of detecting abnormal satellite clock error estimation, constructing a detection model for detecting abnormal deviation of the satellite clock error based on the distribution characteristics of the observation residuals, the smoothness of the UPD over time and the change of the integer ambiguity fixing rate, comprising: In the event of an anomaly detected in the satellite clock error estimation, a multidimensional anomaly feature vector is constructed based on the distribution characteristics of the observation residuals, the stability of the UPD over time, and the change in the integer ambiguity fixation rate. Construct an integer clock anomaly diagnostic index based on multidimensional anomaly feature vectors; Based on the integer clock anomaly diagnostic index, a detection model including satellite clock bias is constructed.

7. The method of claim 1, wherein the method further comprises: PPP floating-point solution is performed on satellite-based integer clock products, and gross errors are detected based on the residuals from the PPP floating-point solution. The number of times each satellite is marked as a gross error is counted, including: The PPP floating-point solution is performed on the integer clock products of the satellite to obtain the IF ambiguity floating-point solution; Based on the IF ambiguity floating-point solution, determine the post-hoc residual of the PPP floating-point solution; The post-verification residuals based on PPP floating-point calculations are used to perform gross error checks station by station to count the number of times each satellite is marked as a gross error in each station during the current epoch.

8. The method of claim 1, wherein the method further comprises: Based on the aforementioned detection model and the number of times each satellite is labeled as a gross error, an extended PPP model that takes into account satellite clock bias parameters is constructed, including: Based on the order of the number of times each satellite is marked as a gross error at each station from largest to smallest, an initial PPP extended model that takes into account the satellite clock bias parameters is constructed. Based on the detection model and the initial PPP extended model, the clock bias deviation parameter is determined; If the rationality verification of the initial PPP extended model based on the clock deviation parameter is passed, the initial PPP extended model is determined to be the final PPP extended model.

9. The method according to any one of claims 1-8, wherein, Based on the aforementioned PPP extended model, using the calculated inter-satellite single-difference UPD product, the fixed integer ambiguity value is determined, yielding the PPP fixed solution, including: Perform inter-satellite single difference on the wide lane and narrow lane UPD products respectively to obtain the wide lane inter-satellite single difference UPD product and the narrow lane inter-satellite single difference UPD product. Calculate the wide-lane floating-point ambiguity, and based on the wide-lane floating-point ambiguity and the wide-lane inter-satellite single-difference UPD product, solve the wide-lane integer ambiguity; Based on the wide lane integer ambiguity and the narrow lane inter-satellite single-difference UPD product, the floating-point solution of the IF ambiguity is corrected, and the narrow lane ambiguity is fixed. Then, based on the PPP extended model and the wide lane integer ambiguity, the fixed IF ambiguity is recombined to obtain the fixed ambiguity solution, and the fixed PPP solution is obtained based on the fixed ambiguity solution. The IF ambiguity floating-point solution is obtained by performing PPP floating-point solution calculation on the satellite's integer clock products; based on the PPP fixed solution and the fixed integer ambiguity value, the detection of fixed anomalies in satellite integer clocks includes: Based on the fixed IF ambiguity and integer ambiguity fixed values ​​corresponding to the PPP fixed solution, the carrier phase post-hoc residual is calculated. Based on the carrier phase post-hoc residual, an inter-satellite single-difference carrier phase residual is constructed; Satellite integer clocks with fixed anomalies are detected based on the offset set of inter-satellite single-difference carrier phase residuals.

10. A satellite integer clock solution post-multiple dimension collaborative anomaly detection device, characterized in that, include: The acquisition module is used to obtain the distribution characteristics of the observation residuals after BDS / GNSS satellite integer clock calculation; The determination module is used to determine the stability of the satellite's UPD over time, as well as the change in integer ambiguity fixation rate; The first model construction module is configured to, in the case of detecting satellite clock error estimation anomaly, construct a detection model for detecting satellite clock error abnormal deviation based on distribution characteristics of the observation residual, a stationary degree of the UPD changing over time, and a change condition of the integer ambiguity fixing rate. The gross error statistics module is configured to perform PPP float solution calculation based on the integer clock product of the satellite, and perform gross error detection based on the PPP float solution calculation residual, and count the number of times each satellite is marked as a gross error. The second model construction module is configured to construct a PPP extension model considering the satellite clock error deviation parameter based on the detection model and the number of times each satellite is marked as a gross error. The model calculation module is configured to determine an integer ambiguity fixed value based on the PPP extension model and the calculated inter-satellite single-difference UPD product, and obtain a PPP fixed solution. The anomaly detection module is configured to detect a satellite integer clock with a fixed abnormality based on the PPP fixed solution and the integer ambiguity fixed value.

Images