GNSS (Global Navigation Satellite System) slow-varying fault detection method combining observation domain, position domain and time domain

Abstract

The invention provides a GNSS (Global Navigation Satellite System) slow-varying fault detection method combining an observation domain, a position domain and a time domain, which relates to the technical field of satellite positioning and comprises the following steps of: based on the concept of solution separation in the position domain, regarding the difference between a universal set solution and a subset solution as a fault detection amount, and storing the fault detection amount into a smooth window; meanwhile, a consistency check factor is designed by utilizing residual error information in an observation domain, so that the ending moment of the slow-changing fault is accurately obtained, and the false alarm risk caused by abnormal information is avoided; besides, on the basis of the fault detection amount stored in the smooth window in the time domain, a new fault detection amount is obtained through reconstruction, and slow-changing faults are detected rapidly. And finally, by comparing the fault detection quantity with the detection threshold value, completing the detection of the slow-changing fault. The method not only supports rapid and accurate detection of slow-varying faults, but also is suitable for complex dynamic scenes, and provides a new thought for GNSS data quality control.

Description

Technical Field

[0001] The present invention relates to the field of satellite positioning technology, and in particular to a GNSS slow-varying fault detection method combining observation domain, position domain and time domain. Background Art

[0002] With the completion of the networking of my country's Beidou navigation system and the continuous development of the Global Navigation Satellite System (GNSS), its positioning, navigation and timing (PNT) services have been widely used in the fields of smart cities, intelligent transportation and unmanned driving, and are developing in a more ubiquitous, intelligent and reliable direction. Due to the inherent fragility of GNSS, the universality of errors, uncertainty and the susceptibility of signals to shielding, PNT services will inevitably be interfered by various signal anomalies or faults. As an abnormal state, if a fault is not successfully detected or identified, it threatens the navigation service performance of the end user. Therefore, improving fault detection capabilities is the key to promoting the development of GNSS high-precision positioning services, which is of great significance and research value.

[0003] Considering that after error correction, there are still slowly growing errors (SGE) of the same magnitude as the observation noise in the observation value due to receiver aging, satellite orbit and clock anomalies. Given the trend of the slow growing error over time, it takes a long time for the fault detection to trigger the detection condition, resulting in a gradual decrease in the accuracy of the positioning solution and seriously affecting the continuity of the GNSS navigation service. Therefore, slowly growing faults are regarded as the most dangerous type of fault, and it is necessary to study a robust detection method for slow growing faults. Summary of the invention

[0004] In order to overcome the above-mentioned shortcomings of the prior art, the main purpose of the present invention is to provide a GNSS slow-varying fault detection method combining observation domain, position domain and time domain.

[0005] To achieve the above object, the present invention adopts the following technical solution: a GNSS slow-varying fault detection method combining observation domain, position domain and time domain, comprising the following steps:

[0006] Constructing fault detection quantity T based on solution separation in position domain s1 and the fault detection threshold T d1 , and count the count information of each subset;

[0007] Design consistency test factor F based on observed residuals c ; If the consistency test factor F is tested by the 3Sigma criterionc If there is no significant difference in the sequence and the lock-loss duration of the subset exceeds the set threshold, the fault test quantity of the subset within the smoothing window is initialized and the current fault test quantity is recorded; if the consistency test factor F c If there are significant differences in the sequence, the fault detection quantity of all subsets in the smoothing window is initialized and the current fault detection quantity is recorded;

[0008] The count of each subset is compared with the size of the smoothing window. If the count of each subset is equal to the size of the smoothing window, the information stored in the smoothing window is used to reconstruct the fault test amount and the detection threshold, and the relationship between the reconstructed fault test amount and the detection threshold is determined; if the count of each subset is not equal to the size of the smoothing window, the relationship between the current fault test amount and the detection threshold is determined; when the fault test amount is less than the detection threshold, the fault location information is obtained according to the detection threshold and the corresponding relationship between the count information of the subset and the location domain.

[0009] It also includes: if the fault inspection amount is greater than or equal to the detection threshold, the observation value corresponding to the fault subset is eliminated.

[0010] The fault detection quantity T s1 Solve X for the whole set (0) And the subset solution X (k) The distance between (0) Refers to the positioning solution obtained by involving all satellites in the solution; Subset solution X (k) It refers to the positioning solution obtained by removing the kth satellite from the solution. Taking subset k as an example, the fault detection quantity T s1 The calculation formula is as follows:

[0011]

[0012] in, is the fault test quantity of the kth subset; k is the subset identifier; q is the direction identifier, and the values ​​of q are 1, 2, and 3, respectively representing the three components of the coordinates in the positioning solution; is the standard deviation of the solution separation information of subset k in direction q; Solve X for the whole set (0) The value of the middle q direction; Solve X for the subset ( k ) The value of the middle q direction;

[0013] The fault detection threshold T d1 The false alarm rate is evenly distributed to the three components of each subset and determined using the probability density function. The formula is as follows:

[0014]

[0015] Where x is the integral variable identifier; χ 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N set is the number of subsets; is the fault detection threshold for subset k.

[0016] The method of obtaining the fault inspection results of the previous and next epochs comprises the following steps:

[0017] In the observation domain, the global test quantity T g , the formula is as follows:

[0018] T g =r T W

[0019] Among them, r is the vector of observation residuals, W is the observation weight matrix, (·) T Represents vector transpose;

[0020] Consistency test factor F c It is the global test value at the previous moment divided by the global test value at the current moment. The formula is as follows:

[0021]

[0022] in, is the consistency check factor at the jth moment; j is the time index, and its value is greater than 1; is the global test quantity at the jth moment; is the global test quantity at the j-1th moment;

[0023] After a period of continuous tracking, the consistency test factor of each epoch is stored in the cache, and the 3Sigma criterion is used to test whether there is a significant difference in the fault test amount of the previous and next epochs.

[0024] The method of checking the fault inspection quantity results by the 3Sigma criterion to see whether there is a significant difference comprises the following steps:

[0025] Get the mean of the consistency check factor in the cache and its standard deviation σ F , the formula is as follows:

[0026]

[0027] Where j is the time index; m is the consistency check factor F in the cache c The number of

[0028] like This indicates that the consistency check factor F at the current moment cThe sample composed of the cache does not meet the 3Sigma criterion, and the current time is regarded as the fault end time.

[0029] The determining of the relationship between the reconstructed fault detection amount and the detection threshold comprises the following steps:

[0030] For the fault detection quantity T s2 If the count of a subset is equal to the size of each subset in the smoothing window, the fault detection quantity T stored in the smoothing window can be used. s1 , reconstruct the fault detection quantity T s2 , the formula is as follows:

[0031]

[0032] In the formula, k is the subset identifier; j is the count identifier; is the fault detection quantity after reconstruction of subset k; N is the fault detection quantity stored in the jth position of subset k in the smoothing window; w is the size of each subset in the smoothing window;

[0033] Fault detection threshold T d2 Calculation: Assume that the fault detection quantity between epochs is T s1 Independent of each other, taking subset k as an example, the reconstructed fault detection quantity T s2 The threshold T d2 , the formula is as follows:

[0034]

[0035] Where x is the integral variable identifier; X 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N set is the number of subsets; N w is the size of each subset in the smoothing window; is the fault detection threshold of the kth subset.

[0036] A GNSS slow-varying fault detection system combining observation domain, position domain and time domain, comprising:

[0037] Solution separation module: construct fault detection quantity T based on solution separation in location domain s1 and the fault detection threshold T d1 , and count the count information of each subset;

[0038] Consistency test module: Design consistency test factor F based on observed residuals c , reflecting the fault inspection results of the previous and next epochs;

[0039] Fault detection and analysis module: If the consistency test factor F is tested by the 3Sigma criterion c If there is no significant difference in the sequence and the lock-loss duration of the subset exceeds the set threshold, the fault test quantity of the subset within the smoothing window is initialized and the current fault test quantity is recorded; if the consistency test factor F c If there are significant differences in the sequence, the fault detection amount of all subsets in the smoothing window is initialized and the current fault detection amount is recorded; then, the count of each subset is compared with the size of the smoothing window. If the count of each subset is equal to the size of the smoothing window, the information stored in the smoothing window is used to reconstruct the fault detection amount and the detection threshold;

[0040] Fault detection result judgment module: if the fault test quantity is greater than or equal to the detection threshold, the observation value corresponding to the fault subset is eliminated; if the fault test quantity is less than the detection threshold, the positioning solution is output according to the detection threshold and the corresponding relationship between the counting information of the subset and the location domain.

[0041] Compared with the prior art, the beneficial effects of the present invention are as follows: the present application calculates the fault test quantity around the location domain, and reconstructs a new fault test quantity through the information of the smoothing window, which can greatly shorten the detection time of slow-varying faults. In addition, the construction of the consistency check factor can avoid the risk of false alarms caused by the smoothing window. The example results show that compared with the AIME method, the present invention significantly improves the detection performance of slow-varying faults, and effectively alleviates the impact of abnormal observation information on positioning in complex scenes. In short, the present invention not only supports the rapid and accurate detection of slow-varying faults, but also is suitable for complex dynamic scenes, providing new ideas for GNSS data quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings described herein are used to provide further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0043] Figure 1 It is a schematic diagram of the process structure of the present invention;

[0044] Figure 2 This is a diagram showing the detection effect of the present invention and the classic slow-changing fault detection method AIME under a double fault situation;

[0045] Figure 3 It is a positioning error curve diagram of the present invention in a dynamic urban scene. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0047] The present invention aims to solve the influence of slow-varying faults in GNSS observations on the positioning solution. The positioning technology relied on is standard single point positioning (SPP), and its observation equation is as follows:

[0048] P=ρ+c·(dt r -dt s )+I+T+ε (1)

[0049] Where P is the pseudorange observation value; ρ is the receiver position (x r ,y r ,z r ) and satellite position (x s ,y s ,z s ) between the two; c is the speed of light; dt r is the receiver clock error; dt s is the satellite clock error; I is the ionospheric error; T is the tropospheric error; ε is other error terms, mainly including observation noise and multipath error caused by the observation environment.

[0050] Satellite positions and clock errors can be calculated using GNSS observation files and broadcast ephemeris, while tropospheric and ionospheric errors can also be corrected using empirical models. The Taylor series expansion at can obtain the error equation between the receiver and satellite i:

[0051]

[0052] In the formula, r i is the observation residual of satellite i; ρ0 is the satellite position (x si ,y si ,z si ) and the initial position of the receiver The geometric distance between them; dx, dy and dz represent the correction values ​​of the three coordinate components in the positioning solution, and their coefficients are respectively denoted as lpx i , lpy i and lpz i ; is the equivalent clock error; l i is the observation vector of satellite i, satisfying l i =P i -ρ i +c·dt s -I i -T i .

[0053] Formula (2) is the error equation of a single satellite. Similarly, the observation equation is constructed by n satellites, as shown in the following formula:

[0054]

[0055] The observation equation of formula (3) can be simplified as the following matrix expression:

[0056]

[0057] Where r is the vector of observation residuals; A is the observation matrix; is the correction value of the positioning solution; l is the observation vector.

[0058] If the number of visible satellites is greater than the dimension of the correction value of the positioning solution, the correction value of the positioning solution can be obtained by least squares estimation. Calculated by the following formula:

[0059]

[0060] Where W is the observation weight matrix, which can be simplified to the unit matrix; (·) T Represents the transpose of a vector or matrix; (·) -1 Represents the inverse operation of a matrix.

[0061] Finally, superimpose it to the approximate coordinates The final positioning solution X can be obtained, which can be calculated by the following formula:

[0062]

[0063] See also Figure 1 , we can see that the specific steps include:

[0064] A GNSS slow-varying fault detection method combining observation domain, position domain and time domain comprises the following steps:

[0065] Constructing fault detection quantity T based on solution separation in position domain s1 and the fault detection threshold T d1 , and count the count information of each subset;

[0066] Design consistency test factor F based on observed residuals c ; If the consistency test factor F is tested by the 3Sigma criterion c If there is no significant difference in the sequence and the lock-loss duration of the subset exceeds the set threshold, the fault test quantity of the subset within the smoothing window is initialized and the current fault test quantity is recorded; if the consistency test factor F c If there are significant differences in the sequence, the fault detection quantity of all subsets in the smoothing window is initialized and the current fault detection quantity is recorded;

[0067] The count of each subset is compared with the size of the smoothing window. If the count of each subset is equal to the size of the smoothing window, the information stored in the smoothing window is used to reconstruct the fault test quantity and the detection threshold, and the relationship between the reconstructed fault test quantity and the detection threshold is determined; if the count of each subset is not equal to the size of the smoothing window, the relationship between the current fault test quantity and the detection threshold is determined; when the fault test quantity is less than the detection threshold, the fault location information is obtained according to the detection threshold and the corresponding relationship between the count information of the subset and the location domain.

[0068] It also includes: when the fault inspection amount is greater than or equal to the detection threshold, eliminating the observation value corresponding to the fault subset until the fault inspection amount is less than the detection threshold.

[0069] Fault detection quantity T s1 Solve X for the whole set (0) and subset solution X (k) The distance between (0) Refers to the positioning solution obtained by involving all satellites in the solution; Subset solution X (k) It refers to the positioning solution obtained by excluding the kth satellite from the solution. Taking subset k as an example, the fault detection quantity T s1 The calculation formula is as follows:

[0070]

[0071] in, is the fault test quantity of the kth subset; k is the subset identifier; q is the direction identifier, and the values ​​of q are 1, 2, and 3, respectively representing the three components of the coordinates in the positioning solution; is the standard deviation of the solution separation information of subset k in direction q; Solve X for the whole set (0) The value of the middle q direction; Solve X for the subset (k) The value of the middle q direction;

[0072] Fault detection threshold T d1 The false alarm rate is evenly distributed to the three components of each subset and determined using the probability density function. The formula is as follows:

[0073]

[0074] Where x is the integral variable identifier; χ 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N set is the number of subsets; is the fault detection threshold for subset k.

[0075] Obtaining the fault inspection results of the previous and next epochs includes the following steps:

[0076] In the observation domain, the global test quantity T g , the formula is as follows:

[0077] T g =r T W

[0078] Among them, r is the vector of observation residuals, W is the observation weight matrix, (·) T Represents vector transpose;

[0079] Consistency test factor F c It is the global test value at the previous moment divided by the global test value at the current moment. The formula is as follows:

[0080]

[0081] in, is the consistency check factor at the jth moment; j is the time index, and its value is greater than 1; is the global test quantity at the jth moment; is the global test quantity at the j-1th moment;

[0082] After a period of continuous tracking, the consistency test factor of each epoch is stored in the cache, and the 3Sigma criterion is used to test whether there is a significant difference in the fault test amount of the previous and next epochs.

[0083] The results of the fault inspection quantity are significantly different by using the 3Sigma criterion, including the following steps:

[0084] Get the mean of the consistency check factor in the cache and its standard deviation σ F , the formula is as follows:

[0085]

[0086]

[0087] Where j is the time index; m is the consistency check factor F in the cache c The number of

[0088] like This indicates that the consistency check factor F at the current moment c The sample composed of the cache does not meet the 3Sigma criterion, and the current time is regarded as the fault end time.

[0089] Determining the relationship between the reconstructed fault detection amount and the detection threshold includes the following steps:

[0090] For the fault detection quantity Ts2 If the count of a subset is equal to the size of each subset in the smoothing window, the fault detection quantity T stored in the smoothing window can be used. s1 , reconstruct the fault detection quantity T s2 , the formula is as follows:

[0091]

[0092] In the formula, k is the subset identifier; j is the count identifier; is the fault detection quantity after reconstruction of subset k; N is the fault detection quantity stored in the jth position of subset k in the smoothing window; w is the size of each subset in the smoothing window;

[0093] Fault detection threshold T d2 Calculation: Assume that the fault detection quantity between epochs is T s1 Independent of each other, taking subset k as an example, the reconstructed fault detection quantity T s2 The threshold T d2 , the formula is as follows:

[0094]

[0095] Where x is the integral variable identifier; χ 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N set is the number of subsets; N w is the size of each subset in the smoothing window; is the fault detection threshold of the kth subset.

[0096] Example 2

[0097] This is the basis of Example 1. When determining the fault location, the following steps are included:

[0098] Step 1: construct the fault detection quantity and detection threshold around the location domain, and count the count information of each subset. This section introduces the solution separation idea based on the location domain, constructs the fault detection quantity T s1 and its detection threshold T d1 , mainly includes 2 small steps:

[0099] (1) Fault detection quantity T s1 Construction

[0100] The idea of ​​fault detection based on location domain is to regard the distance between the full set solution and the subset solution as the fault detection quantity T s1, which is also called "solution separation". The subset is constructed based on the fault hypothesis. Taking the single fault hypothesis as an example, that is, assuming that only one satellite is abnormal at the current moment, the number of subsets is equivalent to the number of satellites. (0) It refers to the positioning solution obtained by all satellites participating in the calculation, and the subset solution X (k) It refers to the positioning solution obtained by removing some satellites, such as the kth satellite, from the calculation. s1 It can be calculated by the following formula:

[0101]

[0102] In the formula, k is the subset identifier; is the fault detection quantity of the kth subset; q is the direction identifier, and its values ​​1, 2 and 3 represent the three components of the coordinates in the positioning solution respectively; Solve X for the whole set (0) The value of the middle q direction; Solve X for the subset (k) The value of the middle q direction; is the standard deviation of the solution separation in the q direction for subset k.

[0103] The fault detection quantity of each subset can be calculated by formula (7), and then stored in the specified position of the smoothing window, and the counting information of each subset is counted to reflect whether the satellite signal is continuously tracked.

[0104] (2) Fault detection threshold T d1 Calculation

[0105] Assuming that the three coordinate components in the location solution X are independent of each other, the fault detection quantity of each subset is If the error is less than 0, it obeys the chi-square distribution with 3 degrees of freedom, otherwise it obeys the non-central chi-square distribution with 3 degrees of freedom. According to the risk requirements of the application scenario, the false alarm rate is evenly distributed to the three components of each subset, and the probability density function is used to determine the fault detection threshold of each subset. It can be calculated by the following formula:

[0106]

[0107] Where x is the integral variable identifier; 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N set is the number of subsets; is the fault detection threshold for subset k.

[0108] Step 2: Design consistency check factors based on the residuals in the observation domain to check whether there is a significant difference in the fault inspection quantity before and after the epoch. If there is no significant difference, proceed to step 3, otherwise proceed to step 4. This section introduces the calculation of consistency check factors around the residuals in the observation domain, which mainly includes 3 small steps:

[0109] (1) In the observation domain, the global test quantity T g , can be calculated by the following formula:

[0110] T g =r T Wr (9)

[0111] In the formula, r is the vector of observation residuals, W is the observation weight matrix, (·) T Represents vector transpose.

[0112] (2) Consistency test factor F c It is the global test value at the previous moment divided by the global test value at the current moment, which can be calculated by the following formula:

[0113]

[0114] in, is the consistency check factor at the jth moment; j is the time index, and its value is greater than 1; is the global test quantity at the jth moment; is the global test quantity at the j-1th moment.

[0115] (3) After continuous tracking for a period of time, the consistency test factor of each epoch is stored in the cache, and the 3Sigma criterion is used to test whether there is a significant difference in the fault test quantity of the previous and next epochs (there is no specific quantitative indicator for this difference, which depends on the situation). The specific operation is as follows:

[0116] ① Calculate the mean of the consistency check factor in the cache and its standard deviation σ F , can be calculated by the following formula:

[0117]

[0118] Where j is the time index; m is the consistency check factor F in the cache c The number of .

[0119] ②If This indicates that the consistency check factor F at the current moment c If the sample composed of the cache is not satisfied, the current time is regarded as the fault end time, and step 4 is executed.

[0120] Step 3: If the unlocking time of each satellite or subset exceeds the tolerance time, execute step 5; otherwise, execute step 6.

[0121] Step 4: Initialize the fault detection amount of all subsets in the smoothing window and record the current fault detection amount.

[0122] Step 5: Initialize the fault detection amount of the subset within the smoothing window and record the current fault detection amount.

[0123] Step 6: If the count of each subset is equal to the size of the smoothing window, execute step 7; otherwise, execute step 8.

[0124] Step 7: Use the fault detection quantity T stored in the smoothing window of the subset s1 , reconstruct the fault detection quantity T s2 and its fault detection threshold T d2 This part mainly includes 2 small steps:

[0125] (1) Fault detection quantity T s2 Construction

[0126] From step 6, we can know that the count of a subset is equal to the size of each subset in the smoothing window. The fault detection quantity T stored in the smoothing window can be used. s1 , reconstruct the fault detection quantity T s2 , can be calculated by the following formula:

[0127]

[0128] In the formula, k is the subset identifier; j is the count identifier; is the fault detection quantity after reconstruction of subset k; is the fault detection quantity T stored in the jth position of subset k in the smoothing window s1 ; N w is the size of the smoothing window for each subset.

[0129] (2) Fault detection threshold T d2 Calculation

[0130] Assume that the fault detection quantity between epochs is T s1 are independent of each other, then the reconstructed fault detection quantity T s2 The threshold T d2 , can be determined by the following formula:

[0131]

[0132] Where x is the integral variable identifier; 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N setis the number of subsets; N w is the size of each subset in the smoothing window; is the fault detection threshold of the kth subset.

[0133] Step 8, determine the relationship between the current fault test quantity and the detection threshold. If the test quantity is greater than or equal to the detection threshold, the observation value corresponding to the fault subset is eliminated. Otherwise, the positioning solution is output.

[0134] Figure 2 This is a detection effect diagram of the present invention and the classic slow-changing fault detection method, that is, the present invention and the autonomous integrity monitoring extrapolation method AIME, under the double fault situation, and the fault simulation period is 1600 seconds to 1799 seconds; Figure 3 This is a curve diagram of the positioning error in a dynamic urban scene.

[0135] It should be noted that, in the present invention, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0136] The above embodiments are merely examples of the present invention and do not limit the protection scope of the present invention. All designs that are the same or similar to the present invention fall within the protection scope of the present invention.

Claims

1. A GNSS slow-varying fault detection method combining observation domain, position domain and time domain, characterized in that: The following steps are involved: Constructing fault detection quantity T based on solution separation in position domain s1 and the fault detection threshold T d1 , and count the count information of each subset; Design consistency test factor F based on observed residuals c ; If the consistency test factor F is tested by the 3Sigma criterion c If there is no significant difference in the sequence and the lock-out duration of the subset exceeds the set threshold, the fault detection amount of the subset within the smoothing window is initialized and the current fault detection amount is recorded; If the consistency test factor F c If there are significant differences in the sequence, the fault detection quantity of all subsets in the smoothing window is initialized and the current fault detection quantity is recorded; Compare the count of each subset with the size of the smoothing window. If the count of each subset is equal to the size of the smoothing window, use the information stored in the smoothing window to reconstruct the fault detection amount and the detection threshold, and determine the relationship between the reconstructed fault detection amount and the detection threshold. If the counts of each subset are not equal to the size of the smoothing window, the relationship between the current fault inspection amount and the detection threshold is determined; When the fault detection amount is less than a detection threshold, fault location information is obtained according to the detection threshold and in combination with a corresponding relationship between the count information of the subset and the location domain.

2. The GNSS slow-varying fault detection method in the combined observation domain, position domain and time domain as claimed in claim 1, characterized in that: Also includes: If the fault verification value is greater than or equal to the detection threshold, the observation values ​​corresponding to the fault subset are eliminated.

3. The GNSS slow-varying fault detection method in the combined observation domain, position domain and time domain as claimed in claim 1, characterized in that: The fault detection quantity T s1 Solve X for the whole set (0) And the subset solution X (k) The distance between (0) Refers to the positioning solution obtained by involving all satellites in the solution; Subset solution X (k) It refers to the positioning solution obtained by removing the kth satellite from the solution. Taking subset k as an example, the fault detection quantity T s1 The calculation formula is as follows: in, is the fault test quantity of the kth subset; k is the subset identifier; q is the direction identifier, and the values ​​of q are 1, 2, and 3, respectively representing the three components of the coordinates in the positioning solution; is the standard deviation of the solution separation information of subset k in direction q; Solve X for the whole set (0) The value of the middle q direction; Solve X for the subset (k) The value of the middle q direction; The fault detection threshold T d1 The false alarm rate is evenly distributed to the three components of each subset and determined using the probability density function. The formula is as follows: Where x is the integral variable identifier; χ 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N set is the number of subsets; is the fault detection threshold for subset k.

4. The GNSS slow-varying fault detection method in the combined observation domain, position domain and time domain as claimed in claim 1, characterized in that: The method of obtaining the fault inspection results of the previous and next epochs comprises the following steps: In the observation domain, the global test quantity T g , the formula is as follows: T g =r T Wr Among them, r is the vector of observation residuals, W is the observation weight matrix, (·) T Represents vector transpose; Consistency test factor F c It is the global test value at the previous moment divided by the global test value at the current moment. The formula is as follows: in, is the consistency check factor at the jth moment; k is the time index, and its value is greater than 1; is the global test quantity at the kth moment; is the global test quantity at the k-1th moment; After a period of continuous tracking, the consistency test factor of each epoch is stored in the cache, and the 3Sigma criterion is used to test whether there is a significant difference in the fault test amount of the previous and next epochs.

5. The GNSS slow-varying fault detection method in the combined observation domain, position domain and time domain as claimed in claim 1, characterized in that: The method of checking the fault inspection quantity results by the 3Sigma criterion to see whether there is a significant difference comprises the following steps: Get the mean of the consistency check factor in the cache and its standard deviation σ F , the formula is as follows: Where j is the time index; m is the consistency check factor F in the cache c The number of like This indicates that the consistency check factor F at the current moment c The sample composed of the cache does not meet the 3Sigma criterion, and the current time is regarded as the fault end time.

6. The GNSS slow-varying fault detection method in the combined observation domain, position domain and time domain as claimed in claim 1, characterized in that: The determining of the relationship between the reconstructed fault detection amount and the detection threshold comprises the following steps: For the fault detection quantity T s2 If the count of a subset is equal to the size of each subset in the smoothing window, the fault detection quantity T stored in the smoothing window can be used. s1 , reconstruct the fault detection quantity T s2 , the formula is as follows: In the formula, k is the subset identifier; j is the count identifier; is the fault detection quantity after reconstruction of subset k; N is the fault detection quantity stored in the jth position of subset k in the smoothing window; w is the size of each subset in the smoothing window; Fault detection threshold T d2 Calculation: Assume that the fault detection quantity between epochs is T s1 Independent of each other, taking subset k as an example, the reconstructed fault detection quantity T s2 The threshold T d2 , the formula is as follows: Where x is the integral variable identifier; χ 2 (·) is the probability density function of the chi-square distribution; fr is the degree of freedom; P fa is the false alarm rate; N set is the number of subsets; N w is the size of each subset in the smoothing window; is the fault detection threshold of the kth subset.

7. A GNSS slow-varying fault detection system combining observation domain, position domain and time domain, characterized in that: include: Solution separation module: construct fault detection quantity T based on solution separation in location domain s1 and the fault detection threshold T d1 , and count the count information of each subset; Consistency test module: Design consistency test factor F based on observed residuals c , reflecting the fault inspection results of the previous and next epochs; Fault detection and analysis module: If the consistency test factor F is tested by the 3Sigma criterion c If there is no significant difference in the sequence and the lock-out duration of the subset exceeds the set threshold, the fault detection amount of the subset within the smoothing window is initialized and the current fault detection amount is recorded; If the consistency test factor F c If there are significant differences in the sequence, the fault detection amount of all subsets in the smoothing window is initialized and the current fault detection amount is recorded; then, the count of each subset is compared with the size of the smoothing window. If the count of each subset is equal to the size of the smoothing window, the information stored in the smoothing window is used to reconstruct the fault detection amount and the detection threshold; Fault detection result judgment module: if the fault inspection amount is greater than or equal to the detection threshold, the observation value corresponding to the fault subset is eliminated; The fault verification amount is less than a detection threshold, and a positioning solution is output according to the detection threshold and in combination with the corresponding relationship between the counting information of the subset and the location domain.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

9. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method described in any one of claims 1 to 7 is implemented.

Images