BDS / INS navigation method and device based on innovation difference fault detection

Abstract

The application provides a BDS / INS navigation method and device based on innovation difference fault detection, and belongs to the technical field of navigation and positioning. A BDS pseudorange measurement prediction value of a visible satellite is determined based on BDS ephemeris; pseudorange innovation is calculated according to the BDS pseudorange measurement prediction value and the pseudorange measurement, and a current innovation set is formed; a dynamic innovation data set is formed by the current innovation set and a historical optimal pseudorange innovation set; the current innovation set and the pseudorange innovation in the dynamic innovation data set are pairwise differentiated, innovation difference statistics of N visible satellites are calculated based on the difference values; then, fault detection is iteratively performed; in each iteration, the innovation difference statistics are compared with a fault detection threshold, and unusable pseudorange innovation is removed. The weight is calculated based on the available pseudorange innovation, Kalman filter measurement update is performed, and navigation error state estimation is obtained. The application can reduce the failure rate of fault detection under the condition of complex environmental interference and the condition of system error accumulation.

Description

Technical Field

[0001] This invention relates to the field of navigation and positioning technology, and specifically to a BDS / INS navigation method and apparatus based on information differential fault detection. Background Technology

[0002] The BeiDou Navigation Satellite System (BDS) signal is susceptible to multipath and non-line-of-sight (NLOS) errors due to environmental obstruction and reflection of electromagnetic waves, affecting positioning accuracy. Therefore, accurately and efficiently detecting and eliminating BDS measurement anomalies during navigation calculations, while ensuring the system's robustness in fault detection under harsh observation conditions, is of paramount practical value.

[0003] Traditional fault detection methods for satellite pseudorange measurements include the Receiver Autonomous Integrity Monitoring (RAIM) algorithm, as well as improved algorithms such as Multiple Hypothesis Separation (MHSS) and Advanced RAIM (ARAIM). However, fault detection methods for standalone BDS systems require high satellite redundancy, and the increased number of subsets in multi-fault hypothesis methods leads to a decrease in computational efficiency. Information fusion in BDS / INS integrated navigation systems is primarily based on Kalman filtering. Prior information obtained from INS calculations is used to predict BDS measurement information, and efficient fault detection is achieved by comparing BDS measurements with prior information.

[0004] Using a filter-based prior method for BDS fault detection, it is first necessary to construct a novelty test statistic, and then perform fault detection based on chi-square distribution and hypothesis testing. The robust estimation algorithm, after checking and eliminating outliers, re-weights the remaining measurements according to the value of the novelty test statistic, thereby increasing the availability of the system.

[0005] Existing BDS fault detection schemes based on INS prior information can efficiently detect BDS outliers, but they have the following two shortcomings:

[0006] (1) The fault detection threshold is usually set to a fixed value and lacks consideration for INS cumulative error, the severity of the observation environment, and the pollution of system navigation state variables.

[0007] (2) In harsh observation scenarios over a long period of time, a high proportion of BDS measurements with anomalies may lead to the accumulation of errors in INS or contamination of the system. In this case, the reduced accuracy of prior information will lead to a decline in the performance verification of the fault detection method. Summary of the Invention

[0008] In view of this, the present invention provides a BDS / INS navigation method and apparatus based on novel information differential fault detection. Its novel information differential fault detection technology can reduce the problem of fault detection failure under complex environmental interference conditions and under conditions of system error accumulation.

[0009] To solve the above-mentioned technical problems, the present invention is implemented as follows.

[0010] A BDS / INS navigation method based on novel information differential fault detection includes:

[0011] Step 1: Based on inertial measurement information, a Kalman filter prediction step is used to predict the navigation error state; based on the navigation error state prediction results, the prior-corrected navigation state is obtained.

[0012] Step 2: Based on the BeiDou Navigation Satellite System (BDS) ephemeris and combined with the prior-corrected navigation status, determine the predicted BDS pseudorange measurement values ​​for each visible satellite. ;

[0013] Step 3: Predict values ​​based on BDS pseudorange measurements from N visible satellites. and pseudorange measurement Calculate N pseudo-distance information , forming the current information set The fault detection threshold Th is calculated by combining the cumulative error of the inertial navigation system (INS), the degree of environmental influence, and the background noise of the BDS receiver.

[0014] Step 4: Set the current information set and historical preferred pseudo-distance new information set To form a dynamic information dataset ; Set the current new information With dynamic information dataset The pseudorange information is pairwise differencing, and the information difference statistics of N visible satellites are calculated based on the difference values. ;

[0015] Step 5: Iterative Fault Detection: In each iteration, determine the current information set. New information differential statistics corresponding to each visible satellite in China Is the maximum value greater than the fault detection threshold Th?

[0016] If the value is greater than the fault detection threshold Th, then the pseudorange information of the satellite corresponding to the maximum value of the information difference statistic will be removed from the current information set. and dynamic information dataset Remove from the middle; based on the new current information set Recalculate the differences and update the new difference statistics. Then based on the new current information set Perform fault detection;

[0017] Otherwise, complete the iteration, current information set The remaining r available pseudorange information are added to the set of historical preferred pseudorange information. And based on time, from oldest to newest, select the best pseudo-distance new information set from history. Delete the r pseudo-distance information from the old time;

[0018] Step 6: Reweight the r available pseudorange information using a robust function to obtain the weight vector in the Kalman filter measurement update formula. ;

[0019] Step 7: Based on the navigation error state prediction results in Step 1, perform Kalman filter measurement updates to obtain the navigation error state estimation results of BDS / INS fusion, and perform feedback correction on the navigation state calculated by INS to realize BDS / INS integrated navigation.

[0020] Preferably, in step 3, the fault detection threshold Th is calculated by combining the accumulated error of the inertial navigation system (INS), the degree of environmental influence, and the background noise of the BDS receiver:

[0021] Using the state error covariance matrix predicted in the first step of Kalman filtering Calculate the position error information entropy EP, which is used to characterize the impact of INS cumulative error on position estimation;

[0022] Environmental factors calculated using BDS-based signal-to-noise ratio measurement It characterizes the impact of the environment on BDS measurement information and the degree of contamination of navigation state variables, reflecting the accuracy of BDS measurement;

[0023] The fault detection threshold Th is then constructed as follows:

[0024] ;

[0025] ;

[0026] in, For correction factor, The baseline coefficient characterizing the noise floor of the BDS receiver, where N is the number of currently visible satellites. Let be the signal-to-noise ratio of the BDS measurement for the i-th visible satellite.

[0027] Preferably, the location error information entropy EP is calculated as follows:

[0028] ;

[0029] in, for The size of the top left corner is submatrix, For determinant operations, It is a natural constant.

[0030] Preferably, the benchmark coefficient Calculated based on the average pseudorange error of BDS fault-free measurements collected in an open environment:

[0031] ;

[0032] in, To calculate the expected value, N is the number of currently visible satellites. This is the pseudorange error vector in an open scene.

[0033] Preferably, in step 6, the r available pseudorange information are reweighted using a robust function to obtain the weight vector in the Kalman filter measurement update formula. for:

[0034] r available pseudorange innovations are represented as follows: The r available pseudorange information are centered and corrected to eliminate the common-mode error effect during the weighting process.

[0035] ;

[0036] in, Let r be the mean of the available pseudo-distance information; This is the available pseudorange information at time k after correction;

[0037] Normalize the modified usable pseudorange innovation to obtain the normalized usable pseudorange innovation. :

[0038] ;

[0039] in, Let be the BDS measurement matrix at time k. The BDS measurement covariance matrix, This is the state error covariance matrix predicted in the first step of the Kalman filter.

[0040] Use robust functions The normalization can be weighted using pseudorange innovation to obtain a weight vector for normal measurements after fault removal. , Weight vector The i-th element.

[0041] Preferably, the robust function is the Tukey function or the IGG-III robust weight function.

[0042] The present invention also provides a BDS / INS navigation device based on innovation differential fault detection, including a Kalman filter module, a pseudorange innovation calculation module, a fault detection threshold calculation module, an innovation differential statistics calculation module, a fault detection module, and a weighting module;

[0043] The Kalman filter module employs the Kalman filter algorithm for BDS / INS integrated navigation. During the filtering prediction stage, it predicts the navigation error state. Based on the navigation error state prediction results, it obtains the prior-corrected navigation state and provides it to the pseudorange information calculation module. During the Kalman filter measurement update stage, it obtains the navigation error state estimation results fused from BDS / INS and provides feedback correction to the navigation state calculated by INS, thereby realizing BDS / INS integrated navigation.

[0044] The pseudorange information calculation module is used to determine the predicted BDS pseudorange measurements for each visible satellite based on the BDS ephemeris and the prior correction navigation state provided by the Kalman filter module. ;Predicted values ​​based on BDS pseudorange measurements and pseudorange measurement Calculate N pseudo-distance information , forming the current information set Provided to the new interest difference statistics calculation module;

[0045] The new interest difference statistics calculation module is used to calculate the current new interest set. and historical preferred pseudo-distance new information set To form a dynamic information dataset ; Set the current new information With dynamic information dataset The pseudorange information is pairwise differencing, and the information difference statistics of N visible satellites are calculated based on the difference values. Provided to the fault detection module;

[0046] The fault detection threshold calculation module is used to calculate the fault detection threshold Th by combining the INS cumulative error, the degree of environmental influence, and the background noise of the BDS receiver, and provide it to the fault detection module.

[0047] The fault detection module is used for iterative fault detection: in each iteration, it determines the current set of information. New information differential statistics corresponding to each visible satellite in China Does the maximum value of Th exceed the fault detection threshold Th? If so, remove the pseudorange information of the satellite corresponding to the maximum value of the information difference statistic from the current information set. and dynamic information dataset Removed from the set of new information. Recalculate the differences and update the new difference statistics. Then based on the new current information set Perform fault detection; otherwise, complete the iteration and set the current information set. The remaining r available pseudorange information are added to the set of historical preferred pseudorange information. And according to time from oldest to newest Delete the r pseudo-distance information from the old time;

[0048] The weighting module is used to reweight the r available pseudorange information using a robust function to obtain the weight vector in the Kalman filter measurement update formula. Update the Kalman filter algorithm in the Kalman filter module.

[0049] Preferably, the fault detection threshold Th used by the fault detection threshold calculation module is constructed as follows:

[0050] Using the state error covariance matrix predicted in the first step of Kalman filtering Calculate the position error information entropy EP, which is used to characterize the impact of INS cumulative error on position estimation;

[0051] Environmental factors calculated using BDS-based signal-to-noise ratio measurement This characterizes the degree of influence of the environment on BDS measurement information and the degree of contamination of navigation state variables;

[0052] The fault detection threshold Th is then constructed as follows:

[0053] ;

[0054] ;

[0055] in, For correction factor, The baseline coefficient characterizing the noise floor of the BDS receiver, where N is the number of currently visible satellites. Let be the signal-to-noise ratio of the BDS measurement for the i-th visible satellite.

[0056] Preferably, in the fault detection threshold calculation module, the location error information entropy EP is calculated as follows:

[0057] ;

[0058] in, for The size of the top left corner is submatrix, For determinant operations, It is a natural constant;

[0059] benchmark coefficient Calculated based on the average pseudorange error of BDS fault-free measurements collected in an open environment:

[0060] ;

[0061] in, To obtain the expected value, This is the pseudorange error vector in an open scene.

[0062] Beneficial effects:

[0063] (1) Based on traditional Kalman filtering, this invention uses the prediction of BDS measurement by INS to construct pseudorange information. By the mutual difference of pseudorange information, the bias effect of clock error, clock drift and system common mode error on measurement fault detection is eliminated. No reference satellite is required, so there is no need to consider the influence of the reliability of the reference satellite.

[0064] (2) Considering that in long-term, harsh observation scenarios, a high proportion of anomalous BDS measurements may lead to error accumulation in INS or system contamination, historically optimized pseudorange information is incorporated into the information difference calculation. This addresses the problem of unstable information difference results caused by an excessively high proportion of anomalous measurements when the number of satellite measurements is insufficient. Furthermore, the historically optimized pseudorange information is continuously updated, adding new information and removing old information, thus avoiding the problem of reduced accuracy of prior information leading to a decline in the performance verification of the fault detection method.

[0065] (3) The fault detection threshold is not a fixed value, but is dynamically calculated using the INS cumulative error based on the state error covariance matrix and the environmental factor based on the measurement signal-to-noise ratio. The INS cumulative error calculated based on the state error covariance matrix characterizes the unreliability of the INS system; the environmental factor based on the measurement signal-to-noise ratio characterizes the influence of the environment on BDS measurement information and the degree of contamination of navigation state variables. It is evident that this threshold construction increases the consideration of INS cumulative error and system navigation state variable contamination, ensuring the accuracy of using the threshold to remove fault data. Furthermore, the threshold calculation also considers the severity of the observation environment, incorporating the BDS receiver's background noise, making the threshold setting more environmentally adaptable. This improved environmental adaptability allows for a tighter threshold in environments with good observation and high redundancy, improving navigation accuracy; while in harsh observation environments, the threshold is relaxed to ensure system availability.

[0066] (4) Traditional fault detection methods may introduce contamination or over-rejection when using binary classification of satellite faults. This invention combines fault detection with robust weighted addition, using fault detection as a prerequisite for robust weighting. Typical faults are eliminated using fault detection, and the remaining usable measurements are corrected using the mean of available BDS measurements, then reweighted using a robust function before being introduced into the system measurements. Compared to traditional fault detection methods, this improves the availability and robustness of the system in complex interference environments. Attached Figure Description

[0067] Figure 1 This is a flowchart of the BDS / INS navigation method based on information differential fault detection in an embodiment of the present invention.

[0068] Figure 2 This is a structural diagram of the BDS / INS navigation device based on information differential fault detection in an embodiment of the present invention. Detailed Implementation

[0069] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0070] This invention provides a BDS / INS navigation method based on information differential fault detection, such as... Figure 1 As shown, the method includes the following steps:

[0071] Step 1: Based on inertial measurement information, a Kalman filter prediction step is used to predict the navigation error state. Based on the navigation error state prediction results, a priori corrected navigation state is obtained, including the predicted navigation position value.

[0072] Step 101: Navigation error state prediction.

[0073] This invention relates to a BDS / INS integrated navigation scheme based on Kalman filtering. Since pure inertial calculations contain errors, it is necessary to estimate the navigation error to correct the inertial measurements. Therefore, the state variables of the Kalman filter... Constructed as a navigation error state variable:

[0074] ;

[0075] in These are the state quantities for longitude, latitude, and altitude errors. and These are the velocity and attitude error state quantities in the navigation coordinate system, respectively. These are the zero-bias acceleration of the IMU and the zero-bias state of the gyroscope, respectively. and These are the clock bias and clock drift state quantities related to BDS, respectively.

[0076] The formula for constructing the Kalman filter prediction steps is as follows:

[0077] ;

[0078] ;

[0079] in Let the state transition matrix be one step. This represents the navigation error state prediction at time k, obtained based on the state prediction at time k-1. Let be the state error covariance matrix for one-step prediction. and Let be the navigation error state matrix and error covariance matrix at time k-1. Let be the system noise covariance matrix at time k-1.

[0080] The Kalman filter measurement update formula is constructed as follows:

[0081] ;

[0082] ;

[0083] ;

[0084] In the formula, and Let K and K represent the state estimate and its mean square error matrix for the current epoch k, respectively. For filter gain, It is the identity matrix. Let be the measurement covariance matrix at time k. Let be the BDS measurement matrix at time k. This is the weight vector. This refers to the measurement and prediction error, also known as the new information.

[0085] The navigation error state prediction is performed using a Kalman filter prediction step to obtain the navigation error state prediction value. This is used for the correction of the inertial navigation state in step 102 and the Kalman filter measurement update in step 6.

[0086] Step 102: Predicting the state based on navigation error results This allows us to obtain a priori corrected navigation state.

[0087] Navigation state prediction is performed based on inertial navigation mechanics arrangement and state-space model.

[0088] The correction calculation for the navigation position in navigation mode is as follows: using the specific force (acceleration) and angular velocity measurement data output by the inertial measurement unit (IMU), the longitude at the current moment is calculated. ,latitude ,high And the speeds in the east, north, and sky directions. and posture Furthermore, the predicted position value in the corrected Earth-centered Earth-fixed coordinate system (ECEF coordinate system) can be obtained as follows:

[0089] ;

[0090] in Let be the radius of curvature of the ramusoidal circle, and e² be the square of the first eccentricity (approximately 0.00669438). This corrected position prediction is used for pseudorange measurement prediction in step 2.

[0091] Step 2: Based on the BeiDou Navigation Satellite System (BDS) ephemeris and combined with the prior-corrected navigation position prediction values, determine the BDS pseudorange measurement prediction values ​​for each visible satellite. , .

[0092] By combining the BDS ephemeris, the orbital position of the i-th visible satellite in the geocentric Earth-fixed coordinate system (ECEF coordinate system) can be calculated. , , By combining the prior-corrected navigation position predictions (x, y, z) in the ECEF coordinate system with the error estimates from the BDS, the predicted BDS pseudorange measurement values ​​are obtained. .

[0093] The prediction formula is:

[0094] ;

[0095] Where c is the speed of light. To utilize the navigation error state variables The receiver clock error obtained through prior correction.

[0096] Step 3: Predict values ​​based on BDS pseudorange measurements from N visible satellites. And the pseudorange measurement of actual satellites, to calculate the pseudorange information of N visible satellites. , forming the current information set The fault detection threshold Th is calculated by combining the cumulative error of the inertial navigation system (INS), the degree of environmental influence, and the background noise of the BDS receiver.

[0097] In this step, the pseudorange information of the i-th satellite is updated. The calculation is as follows:

[0098] ;

[0099] in For the actual pseudorange measurement of the i-th satellite, This is the pseudorange information for the i-th satellite.

[0100] The fault detection threshold Th is constructed as follows:

[0101] Considering the one-step prediction state error covariance matrix It can characterize the impact of IMU error accumulation on position estimates, while the signal-to-noise ratio (SNR) of BDS measurements can reflect the presence of non-line-of-sight or multipath signals to some extent, characterizing the influence of the environment on BDS measurement information and the degree of contamination of navigation state variables. Therefore, based on the one-step prediction state error covariance matrix... In addition, the SNR of the observed satellites is used to calculate the fault detection threshold, ensuring that the threshold can be adaptively adjusted according to the observation scenario.

[0102] The fault detection threshold Th is constructed as follows:

[0103] ;

[0104] in, The entropy is the information entropy of the position error. Environmental factors; The value of the correction factor is set based on the actual data. The reference coefficient is used to characterize the noise floor of a BDS receiver.

[0105] In the Th calculation formula, the position error information entropy EP is calculated as follows:

[0106] ;

[0107] in, for The size of the top left corner is submatrix, This refers to determinant operations.

[0108] In the Th calculation formula, environmental factors The calculation method is as follows:

[0109] ;

[0110] Where N is the current number of visible satellites. Let be the signal-to-noise ratio of the BDS measurement for the i-th visible satellite.

[0111] In the formula for calculating Th, The error can be calculated based on the average pseudorange error of BDS fault-free measurements collected in an open environment. The calculation formula is as follows:

[0112] ;

[0113] in, To calculate the expected value, N represents the number of satellites at the current time. This represents the pseudorange error vector in open scenes, with values ​​ranging from approximately 0.5 meters to 2.5 meters, varying depending on receiver and antenna performance. Furthermore, the coefficients... The performance varies depending on the different INS devices, BDS receivers, and antennas.

[0114] As can be seen, in this step, the fault detection threshold is not a fixed value, but rather takes into account INS cumulative error, the severity of the observation environment, and contamination of system navigation state variables, ensuring the accuracy of threshold-based fault data removal. This fault detection threshold calculation scheme also ensures environmental adaptability, thus tightening the threshold and improving navigation accuracy in environments with good observation and high redundancy, while relaxing the threshold in harsh observation environments to ensure system availability.

[0115] Step 4: Combine historical data with the optimal set of pseudo-distance information Calculate the new interest difference statistic .

[0116] In this step, the current information set will be... and historical preferred pseudo-distance new information set To form a dynamic information dataset ; Set the current new information With dynamic information dataset The pseudorange information is pairwise differencing, and the information difference statistics of N visible satellites are calculated based on the difference values. .

[0117] Among them, the historical preferred pseudo-distance new information set Essentially, it's a sliding window with L (L>20) of new information inside; therefore, the total length of the sliding window is L. This set stores the historical optimal pseudorange information (i.e., usable pseudorange information) up to time k. After fault detection and removal of faulty information, the remaining usable pseudorange information is added to this set. At the same time from this set Remove outdated pseudorange information and preserve the set. The amount of data in the data is L.

[0118] When performing difference calculations, it is not only for the current set of innovations. The present invention incorporates historical preferred pseudo-distance information by performing pairwise difference operations on the information in the middle. and Merging to construct a new dynamic information dataset .

[0119] Suppose that signals are received from N satellites at time k, and the current information set is... It includes N pseudo-distance information, denoted as , Dynamic Information Dataset Including N+L pseudo-distance information, denoted as , Then, the current set of new information... With dynamic information dataset The pseudo-distance information is divided into pairwise differences:

[0120] ;

[0121] This step constructs a dynamic information dataset. Differential calculations are performed, taking into account that the number of visible satellites may be relatively small at certain times. The new information in the sample is divided pairwise, and the randomness of screening outliers is relatively large. Therefore, the difference calculation adds historical optimal information and uses historical data to make up for the lack of satellite search. This solves the problem of unstable information difference results caused by the high proportion of measurement outliers.

[0122] Based on pseudorange new information difference Calculate the novelty differential statistics of pseudorange measurements for each satellite i:

[0123] ;

[0124] In the process of solving for the new interest rate components, a dataset is introduced. In the measurement of innovation, the denominator of the innovation difference statistic formula increases from N to... When the value of N is small, the new information difference component will not cause a significant increase in the difference information result of normal measurement due to individual outliers. Therefore, the stability of the difference result at the current time can be guaranteed.

[0125] Step 5: Iteratively perform fault detection:

[0126] In this step, at each iteration, the current set of new information is determined. New information differential statistics corresponding to each visible satellite in China maximum value Is it greater than the fault detection threshold Th? If so, then set the maximum value of the new information difference statistic. Pseudorange information corresponding to satellites From the current news collection and dynamic information dataset Remove from the middle; and reuse the current set of new information. The residual pseudorange information is used to calculate the information difference statistics of the current time before satellites are removed; then, based on the new current information set... Perform fault detection; if If the value is not greater than the threshold Th, the iteration is complete, and the BDS measurement outlier removal is finished. Assume the current information set... There are r remaining available pseudorange innovations. At this point, according to the dynamic innovation dataset... The data is ordered chronologically from oldest to newest, with r older data points removed and r usable pseudorange information points added to the historical preferred pseudorange information point set. This makes the historical preferred pseudo-distance new information set Updated.

[0127] Step 6: Reweight the r available pseudorange information using a robust function to obtain the weight vector in the Kalman filter measurement update formula. .

[0128] In this step, r available pseudo-distance information are represented as follows: The r available pseudorange information are centered and corrected to eliminate the common-mode error effect during the weighting process.

[0129] ;

[0130] in, Let r be the mean of the available pseudo-distance information; This is the available pseudorange information at time k after correction.

[0131] Obtaining normalized pseudo-distance information for:

[0132] ;

[0133] in, Let be the BDS measurement matrix at time k. This is the BDS measurement covariance matrix.

[0134] Use robust functions The normalization can be weighted using pseudo-range innovation to obtain the weights of normal measurements after removing faults. .

[0135] Here, robust weighting can utilize various truncated robust functions, such as the Tukey function and the IGG-III robust weighting function. Weight vector Only the available information measurement weights are included; the unavailable parts have been removed using fault detection methods.

[0136] Step 7: Transfer the weight vector obtained in Step 6 Substituting the filter gain into the Kalman filter measurement update formula Based on the navigation error state prediction results from step 1, Kalman filtering measurement updates are performed to obtain the navigation error state estimation results of BDS / INS fusion. The position calculated by INS is then corrected by feedback to obtain the navigation state, thus realizing BDS / INS integrated navigation.

[0137] This concludes the process.

[0138] Based on the above method, the present invention also provides a BDS / INS navigation device based on information differential fault detection, such as... Figure 2 As shown, it includes a Kalman filter module, a pseudorange innovation calculation module, a fault detection threshold calculation module, an innovation difference statistics calculation module, a fault detection module, and a weighting module.

[0139] The Kalman filter module employs the Kalman filter algorithm for BDS / INS integrated navigation. In the filtering prediction stage, it obtains navigation error state predictions. Based on the navigation error state prediction results, it obtains the corrected navigation state and provides it to the pseudorange information calculation module. In the Kalman filter measurement update stage, it obtains the navigation error state estimation results fused from BDS / INS and performs feedback correction on the navigation state calculated by INS, thereby realizing BDS / INS integrated navigation.

[0140] The pseudorange information calculation module is used to determine the BDS pseudorange measurement prediction values ​​for each visible satellite based on the BDS ephemeris and the prior corrected navigation position prediction values ​​and clock error prediction values ​​provided by the Kalman filter module. Predicted values ​​based on BDS pseudorange measurements and pseudorange measurement Calculate N pseudo-distance information , forming the current information set It is provided to the new interest difference statistics calculation module.

[0141] The new interest difference statistics calculation module is used to calculate the current new interest set. and historical preferred pseudo-distance new information set To form a dynamic information dataset ; Set the current new information With dynamic information dataset The pseudorange information is pairwise differencing, and the information difference statistics of N visible satellites are calculated based on the difference values. This is provided to the fault detection module.

[0142] The fault detection threshold calculation module is used to calculate the fault detection threshold Th by combining the INS cumulative error, the degree of environmental influence, and the background noise of the BDS receiver, and then provide it to the fault detection module.

[0143] The fault detection module is used for iterative fault detection: in each iteration, it determines the current set of information. New information differential statistics corresponding to each visible satellite in China Does the maximum value of Th exceed the fault detection threshold Th? If so, remove the pseudorange information of the satellite corresponding to the maximum value of the information difference statistic from the current information set. and dynamic information dataset Removed from the set of new information. Recalculate the differences and update the new interest difference statistics, then base the results on the new current interest set. Perform fault detection; otherwise, complete the iteration and set the current information set. The remaining r available pseudorange information are added to the set of historical preferred pseudorange information. And according to time from oldest to newest Delete r pseudo-distances from the old time.

[0144] The weighting module is used to reweight the r available pseudorange information using a robust function to obtain the weight vector in the Kalman filter measurement update formula. Update the Kalman filter algorithm in the Kalman filter module.

[0145] In a preferred embodiment, the fault detection threshold Th is constructed by the fault detection threshold calculation module using a one-step prediction covariance matrix of Kalman filtering. The location covariance information entropy EP is calculated to characterize the impact of INS cumulative error on location estimation; environmental factors are calculated based on BDS measurement signal-to-noise ratio. It characterizes the degree of influence of the environment on BDS measurement information and the degree of contamination of navigation state variables.

[0146] The fault detection threshold Th is then constructed as follows:

[0147] ;

[0148] in, For correction factor, The reference coefficient characterizing the noise floor of a BDS receiver. The calculation formula for the position error index EP, and environmental factors. Calculation formula and benchmark coefficient The method for determining has been described in detail above and will not be repeated here.

[0149] The specific embodiments described above only illustrate the design principles of the present invention. The shapes and names of the components in this description may differ and are not limited. Therefore, those skilled in the art can modify or make equivalent substitutions to the technical solutions described in the foregoing embodiments; and these modifications and substitutions do not depart from the inventive spirit and technical solutions of the present invention, and should all fall within the protection scope of the present invention.

Claims

1. A BDS / INS navigation method based on novel information differential fault detection, characterized in that, include: Step 1: Based on inertial measurement information, a Kalman filter prediction step is used to predict the navigation error state; Based on the navigation error state prediction results, the prior corrected navigation state is obtained; Step 2: Based on the BeiDou Navigation Satellite System (BDS) ephemeris and combined with the prior-corrected navigation status, determine the predicted BDS pseudorange measurement values ​​for each visible satellite. ; Step 3: Predict values ​​based on BDS pseudorange measurements from N visible satellites. and pseudorange measurement Calculate N pseudo-distance information , forming the current information set The fault detection threshold Th is calculated by combining the cumulative error of the inertial navigation system (INS), the degree of environmental influence, and the background noise of the BDS receiver. Step 4: Set the current information set and historical preferred pseudo-distance new information set To form a dynamic information dataset ; Set the current new information With dynamic information dataset The pseudorange information is pairwise differencing, and the information difference statistics of N visible satellites are calculated based on the difference values. ; Step 5: Iterative Fault Detection: In each iteration, determine the current information set. New information differential statistics corresponding to each visible satellite in China Is the maximum value greater than the fault detection threshold Th? If the value is greater than the fault detection threshold Th, then the pseudorange information of the satellite corresponding to the maximum value of the information difference statistic will be removed from the current information set. and dynamic information dataset Remove from the middle; based on the new current information set Recalculate the differences and update the new difference statistics. Then based on the new current information set Perform fault detection; Otherwise, complete the iteration, current information set The remaining r available pseudorange information are added to the set of historical preferred pseudorange information. And based on time, from oldest to newest, select the best pseudo-distance new information set from history. Delete the r pseudo-distance information from the old time; Step 6: Reweight the r available pseudorange information using a robust function to obtain the weight vector in the Kalman filter measurement update formula. ; Step 7: Based on the navigation error state prediction results in Step 1, perform Kalman filter measurement updates to obtain the navigation error state estimation results of BDS / INS fusion, and perform feedback correction on the navigation state calculated by INS to realize BDS / INS integrated navigation.

2. The BDS / INS navigation method based on differential information fault detection as described in claim 1, characterized in that, In step 3, the fault detection threshold Th is calculated by combining the accumulated error of the inertial navigation system (INS), the degree of environmental influence, and the background noise of the BDS receiver: Using the state error covariance matrix predicted in the first step of Kalman filtering Calculate the position error information entropy EP, which is used to characterize the impact of INS cumulative error on position estimation; Environmental factors calculated using BDS-based signal-to-noise ratio measurement It characterizes the impact of the environment on BDS measurement information and the degree of contamination of navigation state variables, reflecting the accuracy of BDS measurement; The fault detection threshold Th is then constructed as follows: ； ； in, For correction factor, The baseline coefficient characterizing the noise floor of the BDS receiver, where N is the number of currently visible satellites. Let be the signal-to-noise ratio of the BDS measurement for the i-th visible satellite.

3. The BDS / INS navigation method based on differential information fault detection as described in claim 2, characterized in that, The location error information entropy EP is calculated as follows: ； in, for The size of the top left corner is submatrix, For determinant operations, It is a natural constant.

4. The BDS / INS navigation method based on differential information fault detection as described in claim 2 or 3, characterized in that, benchmark coefficient Calculated based on the average pseudorange error of BDS fault-free measurements collected in an open environment: ； in, To calculate the expected value, N is the number of currently visible satellites. This is the pseudorange error vector in an open scene.

5. The BDS / INS navigation method based on differential information fault detection as described in claim 1, characterized in that, In step 6, the r available pseudorange information are reweighted using a robust function to obtain the weight vector in the Kalman filter measurement update formula. for: r available pseudorange innovations are represented as follows: The r available pseudorange information are centered and corrected to eliminate the common-mode error effect during the weighting process. ； in, Let r be the mean of the available pseudo-distance information; This is the available pseudorange information at time k after correction; Normalize the modified usable pseudorange innovation to obtain the normalized usable pseudorange innovation. : ； in, Let be the BDS measurement matrix at time k. The BDS measurement covariance matrix, This is the state error covariance matrix predicted in the first step of the Kalman filter. Use robust functions The normalization can be weighted using pseudorange innovation to obtain a weight vector for normal measurements after fault removal. , Weight vector The i-th element.

6. The BDS / INS navigation method based on differential information fault detection as described in claim 5, characterized in that, The robust function is either the Tukey function or the IGG-III robust weight function.

7. A BDS / INS navigation device based on novel information differential fault detection, characterized in that, It includes a Kalman filter module, a pseudorange innovation calculation module, a fault detection threshold calculation module, an innovation difference statistics calculation module, a fault detection module, and a weighting module; The Kalman filter module uses the Kalman filter algorithm of BDS / INS integrated navigation to predict navigation error state during the filtering prediction stage. Based on the navigation error state prediction results, the prior corrected navigation state is obtained and provided to the pseudorange information calculation module; During the Kalman filter measurement update stage, the navigation error state estimation result of BDS / INS fusion is obtained, and the navigation state calculated by INS is corrected by feedback to realize BDS / INS integrated navigation. The pseudorange information calculation module is used to determine the predicted BDS pseudorange measurements for each visible satellite based on the BDS ephemeris and the prior correction navigation state provided by the Kalman filter module. ;Predicted values ​​based on BDS pseudorange measurements and pseudorange measurement Calculate N pseudo-distance information , forming the current information set Provided to the new interest difference statistics calculation module; The new interest difference statistics calculation module is used to calculate the current new interest set. and historical preferred pseudo-distance new information set To form a dynamic information dataset ; Set the current new information With dynamic information dataset The pseudorange information is pairwise differencing, and the information difference statistics of N visible satellites are calculated based on the difference values. Provided to the fault detection module; The fault detection threshold calculation module is used to calculate the fault detection threshold Th by combining the INS cumulative error, the degree of environmental influence, and the background noise of the BDS receiver, and provide it to the fault detection module. The fault detection module is used for iterative fault detection: in each iteration, it determines the current set of information. New information differential statistics corresponding to each visible satellite in China Is the maximum value greater than the fault detection threshold Th? If so, then remove the pseudorange information of the satellite corresponding to the maximum value of the new information difference statistic from the current information set. and dynamic information dataset Removed from the set of new information. Recalculate the differences and update the new difference statistics. Then based on the new current information set Perform fault detection; otherwise, complete the iteration and set the current information set. The remaining r available pseudorange information are added to the set of historical preferred pseudorange information. And according to time from oldest to newest Delete the r pseudo-distance information from the old time; The weighting module is used to reweight the r available pseudorange information using a robust function to obtain the weight vector in the Kalman filter measurement update formula. Update the Kalman filter algorithm in the Kalman filter module.

8. The BDS / INS navigation device based on information differential fault detection as described in claim 7, characterized in that, The fault detection threshold Th used in the fault detection threshold calculation module is constructed as follows: Using the state error covariance matrix predicted in the first step of Kalman filtering Calculate the position error information entropy EP, which is used to characterize the impact of INS cumulative error on position estimation; Environmental factors calculated using BDS-based signal-to-noise ratio measurement This characterizes the degree of influence of the environment on BDS measurement information and the degree of contamination of navigation state variables; The fault detection threshold Th is then constructed as follows: ； ； in, For correction factor, The baseline coefficient characterizing the noise floor of the BDS receiver, where N is the number of currently visible satellites. Let be the signal-to-noise ratio of the BDS measurement for the i-th visible satellite.

9. The BDS / INS navigation device based on differential information fault detection as described in claim 8, characterized in that, In the fault detection threshold calculation module, the location error information entropy EP is calculated as follows: ； in, for The size of the top left corner is submatrix, For determinant operations, It is a natural constant; benchmark coefficient Calculated based on the average pseudorange error of BDS fault-free measurements collected in an open environment: ； in, To obtain the expected value, This is the pseudorange error vector in an open scene.

Images