A RAIM method based on robust adaptive Kalman filter innovation accumulation
Through the RAIM method of anti-difference adaptive Kalman filtering new information accumulation, the accumulated new information detection statistic TRAKF is constructed and compared with the detection threshold, which solves the problem of insufficient detection capability of pseudorange slow-change fault signal, and improves the fault Detection rate and positioning accuracy.
Patent Information
- Application Number
- CN202410773560.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-17
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-06-17
AI Technical Summary
The traditional Kalman filtered RAIM algorithm has poor detection capabilities for pseudorange slow-change fault signals, and it needs to be improved to improve the detection and elimination capabilities of fault satellites.
The RAIM method of anti-difference adaptive Kalman filtering new information accumulation is adopted to define the detection statistic TRAKF by constructing the accumulated new information, and compared it with the detection threshold TD to troubleshoot satellites. In the positioning and solution stage of the satellite navigation system, the pseudorange observation data is input into the difference-resistant adaptive Kalman filter loop, and state prediction, error covariance matrix update and filter gain calculation are performed, and the estimated value of the state vector is finally obtained. New information data, build accumulated new information for fault detection.
It improves the fault detection rate and detection delay, significantly improves the positioning accuracy, and can more accurately identify and eliminate pseudo-range slow-change fault signals.
Smart Images

Figure CN118584510B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of wireless communications, and in particular to a RAIM method based on robust adaptive Kalman filtering new information accumulation. Background Art
[0002] Since the successful launch of the first artificial satellite in 1957, the Global Navigation Satellite System (GNSS) has been developed for a long time and has been widely used in many fields of modern society. With the continuous enhancement of GNSS navigation positioning and timing functions, people have higher and higher requirements for the integrity of GNSS navigation positioning in special fields related to safety such as aviation and autonomous driving. GNSS integrity monitoring means that when a navigation system fails, the system should be able to detect the fault in time and issue an alarm. Integrity monitoring is generally divided into integrity monitoring based on external augmentation systems and receiver autonomous integrity monitoring (RAIM). Among them, receiver autonomous integrity monitoring refers to the function of detecting and troubleshooting faults by using the redundant observations of the receiver for consistency detection without relying on the external augmentation system. Compared with the external augmentation system, RAIM has been recognized and accepted by the majority of users because it can perform integrity monitoring directly on the user side through redundant observations, is not restricted by geographical location, has low cost, and can directly inform users of integrity information. Therefore, improving the RAIM algorithm's ability to detect and eliminate faulty satellites plays a very important role in improving positioning accuracy.
[0003] Kalman filter technology has been widely used in the field of satellite navigation and positioning. Applying Kalman filter technology to RAIM can achieve better performance than the traditional snapshot method, but the traditional Kalman filter RAIM algorithm has poor detection ability for pseudorange slow-changing faults. Extended Kalman filter is an extended form of standard Kalman filter in nonlinear situations. The extended Kalman filter innovation RAIM algorithm is derived from it. This algorithm essentially relies on the pseudorange residual at the current moment to construct statistics to identify and eliminate faulty satellite signals. This method is very effective for pseudorange mutation fault signals. However, when pseudorange slow-changing fault signals appear, the innovation RAIM requires the pseudorange deviation to be large enough so that the statistical detection amount exceeds the set threshold in order to make an accurate judgment. Therefore, it is necessary to improve the RAIM algorithm.
[0004] The existing technology comparison is as follows;
[0005] Comparative Patent 1: A RAIM algorithm based on Kalman filtering in a satellite navigation system, authorization announcement date, 2017 / 12 / 01; authorization announcement date number: CN105487088B,
[0006] The invention relates to a RAIM algorithm based on Kalman filtering in a satellite navigation system. The method applies the RAIM algorithm to Kalman filtering and combines the least squares method with Kalman filtering to improve the convergence speed. The method comprises the following steps: firstly, obtaining the orbital parameters of the satellite according to the navigation message, calculating the position of the satellite, and calculating the altitude angle and azimuth angle of the satellite from the position of the receiver, and eliminating the low elevation angle satellite; establishing a relationship matrix between the satellite data after eliminating the low elevation angle satellite and the erroneous data satellite and the state vector of the system, and establishing an observation equation; selecting a number of satellites in turn, calculating their geometric dilution of precision (GDOP) values; selecting N satellites from P satellites, selecting a group with the smallest GDOP, and the corresponding satellites are used as solution satellites; estimating the initial position and velocity information of the receiver by using the least square method; estimating the position and velocity information in real time by using the Kalman filter method; initializing the parameters of the Kalman filter, assigning the initial value to the Kalman filter by using the solution of the least square method, and giving the positioning result; starting the RAIM algorithm to detect the global fault; after the detection algorithm reports the fault, starting the local detection algorithm to identify and isolate the fault; judging the availability of the least square method, and resetting the Kalman filter by using the solution of the least square method.
[0007] Although it also involves the use of Kalman filtering for faulty satellite detection, the comparative technology mainly uses Kalman filtering and the least squares method to eliminate faults, and uses the residuals of the observed quantity to define the detection quantity; while the present application mainly uses robust adaptive Kalman filtering to eliminate faults, and combines adaptive filtering and robust filtering on the basis of traditional Kalman filtering; for the definition of the detection quantity, the present application uses the new information data output by the filter loop to construct cumulative new information, and uses the cumulative new information to define the detection quantity, rather than using the residuals of the observed quantity to define it; and the detection threshold definition of the present application is inconsistent with that of the comparative technology. Summary of the invention
[0008] In order to solve the above technical problems, the present invention proposes a RAIM method based on the accumulation of new information of robust adaptive Kalman filtering. The method can comprehensively control the state noise and observation noise, comprehensively utilize the new information data of the current and previous epochs, have a higher detection rate and shorter detection delay for faults, and improve the positioning accuracy to a certain extent.
[0009] To achieve the above object, the technical solution adopted by the present invention is:
[0010] A RAIM method based on robust adaptive Kalman filter innovation accumulation, the specific steps are as follows:
[0011] (1) In the positioning solution stage of the satellite navigation system, the pseudo-range observation data is input into the robust adaptive Kalman filter loop, and the new information at time k is obtained;
[0012] (2) Construct the cumulative new information at time k based on the new information at each time obtained in step (1);
[0013] (3) Based on the k-time cumulative new information obtained in step (2), the detection statistic T of the new information accumulation RAIM algorithm is constructed. RAKF ;
[0014] (4) Constructing the detection threshold T D ;
[0015] (5) Substitute the statistic T obtained in step (3) RAKF The detection threshold T obtained in step (4) D Compare and troubleshoot satellites.
[0016] As a further improvement of the present invention, in step (1), the observation data of the pseudorange is input into the robust adaptive Kalman filter loop, and the state equation of the satellite navigation system is expressed as X k =AX k-1 +w k , where X k is the state vector, w k is the process noise vector, in the robust adaptive Kalman filter positioning solution process X k =[x k ,y k ,z k ,Δt k ] T , where x k ,y k 、z k are the position coordinates of the receiver at time k, Δt k is the receiver clock error, the loop update time is T, A is the state transfer matrix, and the observation equation is expressed as: Z k =f(X k )+v k , where Z k is the pseudo-range observation vector at time k, f(X k ) represents Z k With X k The nonlinear relationship, v k is the observation noise vector, and the state vector X of the observation equation at time k k The predicted value of Taylor expansion is performed at ΔZ to obtain the linearized equation: k =H k ΔXk +v k , where H k is the observation matrix obtained by linearization.
[0017] As a further improvement of the present invention, in step (1), the robust adaptive Kalman filter loop process is as follows:
[0018] (1.1) According to X k =AX k-1 +w k Get the state vector X k Predicted value of:
[0019] (1.2) Update the error covariance matrix: P k,k-1 =AP k-1 A T +Q k ;
[0020] (1.3) Calculate the filter gain matrix
[0021] In the formula, is the observed value Z k The equivalent weight matrix of is obtained by the IGG method;
[0022] a k Constructed by the following formula:
[0023]
[0024] In the formula, k0 and k1 are constants, the value range of k0 is 1.0:1.5, and the value range of k1 is 3.0:8.5. is the robust estimate of the state vector at time k, is the predicted value of the state vector at time k, ||g|| represents the norm operation, and tr represents the trace operation of the matrix;
[0025] (1.4) Update X k The estimated error covariance matrix
[0026] (1.5) Get X k Estimated value:
[0027] (1.6) Calculate the new interest:
[0028] (1.7) Let k-1=k and continue the calculation at the next moment.
[0029] As a further improvement of the present invention, the accumulated new information R in step (2) kDefined as:
[0030]
[0031] Where V k is the new information at the kth moment, D k is its covariance matrix, V k The weight, V k-n+1 ,V k-n+2 ,LV k is the new information from time k-n+1 to time k.
[0032] As a further improvement of the present invention, step (3) specifically comprises:
[0033] (3.1) Calculate R k The mean matrix of :
[0034] (3.2) Calculate R k The covariance matrix of is:
[0035] (3.3) According to steps (3.1) and (3.2), R k The mean and covariance matrices of define the detection statistic of the innovation accumulation RAIM algorithm:
[0036] As a further improvement of the present invention, step (4) specifically comprises:
[0037] (4.1) According to the hypothesis testing method, the fault judgment criterion of the cumulative innovation detection method is assumed to be:
[0038]
[0039] Where, T D is the fault detection threshold;
[0040] (4.2) When there is no fault, T RAKF Subject to the center χ with m degrees of freedom 2 Distribution, its probability density is:
[0041]
[0042] Where Γ is the gamma function;
[0043] (4.3) Given the RAIM false alarm rate: P fa =α;
[0044] (4.4)Construction formula:
[0045] (4.5) Order
[0046] (4.6) Calculate the detection threshold: T D =F -1 (1-α).
[0047] As a further improvement of the present invention, step (5) specifically comprises:
[0048] (5.1) Comparative statistic T RAKF and the detection threshold T D size;
[0049] (5.2) If T RAKF >T D , then there is a faulty satellite and it is necessary to eliminate the faulty satellite;
[0050] (5.3) If T RAKF ≤T D , there is no faulty satellite and the solution is performed normally.
[0051] Beneficial effect: Compared with the prior art, the present invention has the following significant advantages: the present invention proposes a RAIM method based on robust adaptive Kalman filter new information accumulation. In the positioning solution stage of the satellite navigation system, the satellite observation data is input into the robust adaptive Kalman filter loop. In the loop, the state at time k is first predicted in one step according to the state equation of the satellite navigation system, and then the error covariance matrix, the filter gain matrix and the estimated error covariance matrix are calculated and updated, and finally the estimated value of the state vector is obtained and the new information V at time k is calculated. k , and then set k-1=k, and continue the calculation at the next moment. After obtaining the new information at each moment, define the cumulative new information at moment k as R k , and construct the detection statistic T of the new information accumulation RAIM algorithm RAKF , the test statistic T RAKF and the detection threshold T D For comparison, if T RAKF >T D , it means there is a satellite fault. The present invention has a high fault detection rate and a short detection delay, and improves the positioning accuracy to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 is a flow chart of one embodiment of the present invention. DETAILED DESCRIPTION
[0053] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments:
[0054] like Figure 1As shown, the RAIM method based on robust adaptive Kalman filtering innovation accumulation of the present invention includes:
[0055] (1) In the positioning solution stage of the satellite navigation system, the pseudo-range observation data is input into the robust adaptive Kalman filter loop, and the new information at time k is obtained;
[0056] Step (1) specifically includes:
[0057] The pseudo-range observation data is input into the robust adaptive Kalman filter loop, and the state equation of the satellite navigation system can be expressed as X k =AX k-1 +w k , where X k is the state vector, w k is the process noise vector. In the robust adaptive Kalman filter positioning solution process, X k =[x k ,y k ,z k ,Δt k ] T , where x k ,y k 、z k are the position coordinates of the receiver at time k, Δt k is the receiver clock error, the loop update time is T, and A is the state transfer matrix. The observation equation can be expressed as: k =f(X k )+v k , where Z k is the pseudo-range observation vector at time k, f(X k ) represents Z k With X k The nonlinear relationship, v k is the observation noise vector. For the observation equation at time k, the state vector X k The predicted value of Taylor expansion is performed at ΔZ to obtain the linearized equation: k =H k ΔX k +v k , where H k is the observation matrix obtained by linearization.
[0058] The robust adaptive Kalman filter loop process is as follows:
[0059] (1.1) According to X k =AX k-1 +w k Get the state vector X k Predicted value of:
[0060] (1.2) Update the error covariance matrix: P k,k-1 =AP k-1 A T +Q k ;
[0061] (1.3) Calculate the filter gain matrix
[0062] In the formula, is the observed value Z k The equivalent weight matrix of can be obtained by the IGG method.
[0063] a k It can be constructed by the following formula:
[0064]
[0065] In the formula, k0 and k1 are constants, the value range of k0 is 1.0:1.5, and the value range of k1 is 3.0:8.5. is the robust estimate of the state vector at time k, is the predicted value of the state vector at time k, ||g|| represents the norm operation, and tr represents the trace operation of the matrix.
[0066] (1.4) Update X k The estimated error covariance matrix
[0067] (1.5) Get X k Estimated value:
[0068] (1.6) Calculate the new interest:
[0069] (1.7) Let k-1=k and continue the calculation at the next moment.
[0070] (2) Construct the cumulative new information at time k based on the new information at each time obtained in step (1);
[0071] The accumulated new interest R in step (2) k Defined as:
[0072]
[0073] Where V k is the new information at the kth moment, D k is its covariance matrix, V k The weight, V k-n+1 ,V k-n+2 ,LV kis the new information from time k-n+1 to time k.
[0074] (3) Based on the k-time cumulative new information obtained in step (2), the detection statistic T of the new information accumulation RAIM algorithm is constructed. RAKF ;
[0075] Step (3) specifically includes:
[0076] (3.1) Calculate R k The mean matrix of :
[0077] (3.2) Calculate R k The covariance matrix of is:
[0078] (3.3) According to steps (3.1) and (3.2), R k The mean and covariance matrices of define the detection statistic of the innovation accumulation RAIM algorithm:
[0079] (4) Constructing the detection threshold T D ;
[0080] Step (4) specifically includes:
[0081] (4.1) According to the hypothesis testing method, the fault judgment criterion of the cumulative innovation detection method is assumed to be:
[0082]
[0083] Where, T D is the fault detection threshold.
[0084] (4.2) When there is no fault, T RAKF Subject to the center χ with m degrees of freedom 2 Distribution, its probability density is:
[0085]
[0086] Where Γ is the gamma function.
[0087] (4.3) Given the RAIM false alarm rate: P fa =α;
[0088] (4.4)Construction formula:
[0089] (4.5) Order
[0090] (4.6) Calculate the detection threshold: T D =F -1 (1-α);
[0091] (5) Substitute the statistic T obtained in step (3) RAKF The detection threshold T obtained in step (4) D Compare and eliminate faulty satellites;
[0092] Step (5) specifically includes:
[0093] (5.1) Comparative statistic T RAKF and the detection threshold T D size;
[0094] (5.2) If T RAKF >T D , then there is a faulty satellite and it is necessary to eliminate the faulty satellite;
[0095] (5.3) If T RAKF ≤T D , there is no faulty satellite and the solution is performed normally.
[0096] The above description is only a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent change made based on the technical essence of the present invention still falls within the scope of protection required by the present invention.
Claims
1. A RAIM method based on robust adaptive Kalman filtering innovation accumulation, characterized in that: The specific steps are as follows: (1) In the positioning solution stage of the satellite navigation system, the pseudo-range observation data is input into the robust adaptive Kalman filter loop, and the new information at time k is obtained; (2) Construct the cumulative new information at time k based on the new information at each time obtained in step (1); (3) Based on the k-time cumulative new information obtained in step (2), the detection statistic T of the new information accumulation RAIM algorithm is constructed. RAKF ; (4) Constructing the detection threshold T D ; (5) Substitute the statistic T obtained in step (3) RAKF The detection threshold T obtained in step (4) D Compare and troubleshoot satellites.
2. The RAIM method based on robust adaptive Kalman filtering innovation accumulation according to claim 1, characterized in that: In step (1), the observation data of pseudorange is input into the robust adaptive Kalman filter loop, and the state equation of the satellite navigation system is expressed as X k =AX k-1 +w k , where X k is the state vector, w k is the process noise vector, in the robust adaptive Kalman filter positioning solution process X k =[x k ,y k ,z k ,Δt k ] T , where x k ,y k 、z k are the position coordinates of the receiver at time k, Δt k is the receiver clock error, the loop update time is T, A is the state transfer matrix, and the observation equation is expressed as: Z k =f(X k )+v k , where Z k is the pseudo-range observation vector at time k, f(X k ) represents Z k With X k The nonlinear relationship, v k is the observation noise vector, and the state vector X of the observation equation at time k k The predicted value of Taylor expansion is performed at ΔZ to obtain the linearized equation: k =H k ΔX k +v k , where H k is the observation matrix obtained by linearization.
3. The RAIM method based on robust adaptive Kalman filtering innovation accumulation according to claim 2, characterized in that: In step (1), the robust adaptive Kalman filter loop process is as follows: (1.1) According to X k =AX k-1 +w k Get the state vector X k Predicted value of: (1.2) Update the error covariance matrix: P k,k-1 =AP k-1 A T +Q k ; ( 1.3) Calculate the filter gain matrix In the formula, is the observed value Z k The equivalent weight matrix of is obtained by the IGG method; a k Constructed by the following formula: In the formula, k0 and k1 are constants, the value range of k0 is 1.0:1.5, and the value range of k1 is 3.0:8.
5. is the robust estimate of the state vector at time k, is the predicted value of the state vector at time k, ||g|| represents the norm operation, and tr represents the trace operation of the matrix; (1.4) Update X k The estimated error covariance matrix (1.5) Get X k Estimated value: (1.6) Calculate the new interest: (1.7) Let k-1=k and continue the calculation at the next moment.
4. The RAIM method based on robust adaptive Kalman filter innovation accumulation according to claim 1, characterized in that: The accumulated new interest R in step (2) k Defined as: Where V k is the new information at the kth moment, D k is its covariance matrix, V k The weight, V k-n+1 ,V k-n+2 ,LV k is the new information from time k-n+1 to time k.
5. The RAIM method based on robust adaptive Kalman filter innovation accumulation according to claim 1, characterized in that: Step (3) specifically includes: (3.1) Calculate R k The mean matrix of : (3.2) Calculate R k The covariance matrix of is: (3.3) According to steps (3.1) and (3.2), R k The mean and covariance matrix of define the detection statistic of the innovation accumulation RAIM algorithm: RAKF =R k T Q k -1 R k .
6. The RAIM method based on robust adaptive Kalman filter innovation accumulation according to claim 1, characterized in that: Step (4) specifically includes: (4.1) According to the hypothesis testing method, the fault judgment criterion of the cumulative innovation detection method is assumed to be: Where, T D is the fault detection threshold; (4.2) When there is no fault, T RAKF Subject to the center χ with m degrees of freedom 2 Distribution, its probability density is: Where Γ is the gamma function; (4.3) Given the RAIM false alarm rate: P fa =α; (4.4)Construction formula: (4.5) Order (4.6) Calculate the detection threshold: T D =F -1 (1-α).
7. The RAIM method based on robust adaptive Kalman filter innovation accumulation according to claim 1, characterized in that: Step (5) specifically includes: (5.1) Comparative statistic T RAKF and the detection threshold T D size; (5.2) If T RAKF >T D , then there is a faulty satellite and it is necessary to eliminate the faulty satellite; (5.3) If T RAKF ≤T D , there is no faulty satellite and the solution is performed normally.
Citation Information
Patent Citations
A Raim Algorithm Based on Kalman Filter in Satellite Navigation System
CN105487088B
RAIM algorithm in satellite navigation system based on Kalman filtering
CN105487088A
Method for monitoring autonomous integrity of GNSS receiver
CN109490916A
Slow fault detection method for GNSS / INS integrated navigation satellite
CN113670337A
Synchronous multi-fault detection method for GNSS / INS (Global Navigation Satellite System / Inertial Navigation System) integrated navigation satellite
CN115047496A
