Adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy

Abstract

The present invention relates to an adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy, belonging to the field of integrated navigation; comprising: step S1, establishing a discrete state equation and an observation equation of the GNSS / INS tight combination system INS error; and using INS sensor data to recurse the carrier motion state; step S2, based on the observation quantity of the GNSS / INS tight combination system, using a multi-channel heterogeneous kernel generalized correlation entropy Kalman filter to estimate the INS recursive error, and using the estimated error value to feedback correct the recursive carrier motion state; step S3, using a moment matching estimator to update the kernel function parameters of each channel in the filter to adapt to the observation noise that varies dynamically with the carrier motion. The present invention improves the robustness of the GNSS / INS integrated navigation system when facing various non-Gaussian noises.

Description

Technical Field

[0001] The invention belongs to the field of integrated navigation, and in particular relates to an adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy. Background Art

[0002] In actual navigation applications, both GNSS and INS have their own limitations. The combined navigation solution of INS and GNSS positioning information can make full use of the complementarity between the various information sources and overcome the limitations of navigation using a single information source. This is also the development trend of navigation technology today, that is, the use of multi-source information fusion navigation and positioning technology using multiple sensors. The GNSS / INS combined navigation system can combine the observation information of GNSS with the recursive information of INS to obtain navigation results with high accuracy, fast refresh rate and strong stability.

[0003] In the research of GNSS / INS integrated navigation system, the information fusion algorithm that fuses the two navigation information is the key technology. The mainstream information fusion algorithm is Kalman filtering, but its robustness is poor. Its performance is optimal only when the noise is Gaussian distributed. In actual application scenarios, there are often tunnels, canyons, urban building shielding, and unknown electromagnetic interference, which leads to the GNSS observations often containing non-Gaussian noise. In the current research on information fusion algorithms for dealing with non-Gaussian noise, there is a method that introduces correlation entropy into Kalman filtering as a robust factor to replace the second-order moment, thereby improving the robustness of the information fusion algorithm, which is called maximum correlation entropy Kalman filtering (MCKF). Because correlation entropy is not sensitive to outliers and can be equivalent to traditional KF under certain conditions. However, this type of method can only deal with noise with obvious outliers without considering its specific distribution. If the GNSS observation noise exhibits other complex non-Gaussian distributions, these methods lack flexibility. In addition, the observation noise of the GNSS / INS integrated navigation system has certain differences in different observation channels (i.e., observations of different satellites), and even only some observation channels show abnormal distribution. The existing methods generally use the method of correlation entropy to weight all channels, which affects the use of normal observations to a certain extent and has obvious conservatism. In addition, with the movement of the carrier and the changes in the surrounding environment, the GNSS observation noise may change dynamically, and the fixed kernel parameters used in MCKF do not have sufficient adaptability to the environment. Summary of the invention

[0004] In view of the above analysis, the present invention aims to disclose an adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy to improve the robustness of the GNSS / INS integrated navigation system when facing various non-Gaussian noises.

[0005] The present invention discloses an adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy, comprising:

[0006] Step S1, establishing the discrete state equation and observation equation of INS error of the GNSS / INS tightly integrated system; and using INS sensor data to perform carrier motion state recursion;

[0007] Step S2: Based on the observations of the GNSS / INS tightly integrated system, a multi-channel heterogeneous kernel generalized correlation entropy Kalman filter is used to estimate the INS recursive error, and the estimated error value is used to perform feedback correction on the recursive carrier motion state;

[0008] Step S3: using a moment matching estimator to update the kernel function parameters of each channel in the filter to adapt to the observation noise that varies dynamically with the motion of the carrier.

[0009] Furthermore, in step S1, the state vector in the discrete state equation is:

[0010] ;

[0011] in, , , , , , and They are the position vector error, velocity error, misalignment angle, gyroscope bias, accelerometer bias, gyroscope scale factor error and accelerometer scale factor error of the INS state recursion result in the ECEF coordinate system;

[0012] The observed quantity in the observation equation is:

[0013] ;

[0014] in, for Time to The pseudoranges observed by the observable satellites, For the The three-dimensional coordinates of the satellite positions, is the three-dimensional coordinate of the recursive position of INS at time k; i=1,…,m, where m is the number of observable satellites.

[0015] Furthermore, the multi-channel hetero-kernel generalized correlation entropy Kalman filter introduces generalized correlation entropy into Kalman filtering; adopts a multi-channel hetero-kernel structure, and the shape parameters of the kernel function in each observation channel are isolated from each other and determined by the statistical characteristics of the observation noise in each channel.

[0016] Furthermore, the process of using the multi-channel heterogeneous kernel generalized correlation entropy Kalman filter to perform INS recursive error estimation includes:

[0017] 1) Perform one-step state prediction to obtain One-step estimation The covariance matrix of :

[0018] 2) According to One-step estimation The covariance matrix of and the covariance matrix of the observation noise Perform Cholesky decomposition;

[0019] 3) According to the Cholesky decomposition results, based on One-step estimation Filter residual Iterative calculation of ; Indicates the number of iterations, the initial value is 1;

[0020] 4) According to The filter residual of the iteration is calculated The generalized correlation entropy weighted values ​​of each observation channel of the iteration constitute the generalized correlation entropy weighted matrix ;

[0021] In the iterative process, the collected The observation error samples of the observation window near the moment are used to estimate the moment, and the weighted matrix is ​​calculated based on the moment estimation result. The shape parameters of the kernel functions of each channel in are updated;

[0022] 5) Using the generalized correlation entropy weighted matrix Calculate Kalman gain ;

[0023] 6) According to Kalman gain Perform state estimation to obtain Iteration value ; When the The iteration value is the same as the When the difference of the iteration value is greater than the threshold, return to step 3) and execute again; otherwise, get the current moment The best estimate of the state quantity .

[0024] Further, time The weight matrix of the iteration :

[0025] ;

[0026] in, For the Channel kernel function,

[0027] Kernel Function The expression is:

[0028] ;

[0029] For the The shape parameter of the kernel function of the observation channel, is the core bandwidth.

[0030] Furthermore, the estimated error value is used to perform feedback correction on the recursive carrier motion state.

[0031] use The best estimate of the state quantity at the moment In , , Correction of the position of the INS recursive state ,speed , direction cosine matrix , get the current position after the estimated error correction ,speed , direction cosine matrix ;

[0032] ;

[0033] .

[0034] Further, in step S3, Updates to the channel kernel function parameters include:

[0035] 1) Set the length of the near-time observation window to , when the system is running greater than or equal to At each moment, the corrected position is obtained. ;

[0036] 2) According to the corrected position at each moment Calculate the Observation channels approaching The error sample sequence observed at each moment;

[0037] 3) Calculate the Observation channels approaching The second-order moment and fourth-order moment estimates of the observation error sample sequence at each moment;

[0038] 4) Use the kurtosis coefficient as the moment matching estimator to calculate the estimate of the shape parameter based on the second-order moment and fourth-order moment estimates ;use Update step S2 Channel kernel function parameter.

[0039] Furthermore, when the filtering process runs to At time , calculate the Observation channels approaching The observation error sequence ;in,

[0040] ;

[0041] for Moment The a posteriori error of the satellite observations is for Time to The pseudoranges observed by the satellites, For the Satellites in The optimal estimate of the pseudorange at time is obtained by Three-dimensional coordinates of satellite position at the moment And the corrected 3D coordinates of the user position It is obtained by taking the difference modulo the vector.

[0042] Furthermore, the second-order moment estimate of the sample sequence is and the fourth moment estimate for:

[0043] ;

[0044] ;

[0045] is the mean of samples at N moments;

[0046] Furthermore, the kurtosis coefficient of the generalized Gaussian distribution function is:

[0047] ;

[0048] pass The inverse function of , by the second-order moment and the fourth moment Calculate shape parameters Estimated value of ; is the Laplace function;

[0049] ;

[0050] Depend on The expression of is calculated by linear interpolation; the second-order moment and the fourth moment Using the second moment estimate and the fourth moment estimate Substitute, calculate Then update the shape parameters of the kernel function in step S2.

[0051] The present invention can achieve the following beneficial effects:

[0052] The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy disclosed in the present invention improves the robustness of the GNSS / INS integrated navigation system in the face of various non-Gaussian noises, and can perform adaptive parameter adjustment to cope with dynamic time-varying noises. In addition, the isolation of the kernel parameters of each observation channel makes the filter more advantageous when facing the situation where each channel contains differential noise. Specifically,

[0053] The multi-channel heterogeneous kernel generalized correlation entropy Kalman filter (MCMK-GMCKF) is used as the GNSS / INS integrated navigation information fusion filter, so that the integrated navigation system can cope with a variety of complex noises with non-Gaussian overall distribution;

[0054] In view of the difference in noise between different GNSS observations, the kernel function shape parameters of different observation channels of the filter are isolated from each other to form a heterogeneous kernel structure, which can not only perform targeted robust processing on abnormal observation channels, but also make full use of normal observations.

[0055] The moment matching estimator is used to evaluate the statistical characteristics of noise in each channel in real time and adjust the kernel parameters so that the integrated navigation system can autonomously adapt to the observation noise that varies dynamically with the motion of the carrier. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0057] Figure 1 The flowchart of the adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy in an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used to illustrate the principles of the present invention together with the embodiments of the present invention.

[0059] An embodiment of the present invention discloses an adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy, such as Figure 1 As shown, including:

[0060] Step S1, establishing the discrete state equation and observation equation of INS error of the GNSS / INS tightly integrated system; and using INS sensor data to perform carrier motion state recursion;

[0061] Step S2: Based on the observations of the GNSS / INS tightly integrated system, a multi-channel heterogeneous kernel generalized correlation entropy Kalman filter is used to estimate the INS recursive error, and the estimated error value is used to perform feedback correction on the recursive carrier motion state;

[0062] Step S3: using a moment matching estimator to update the kernel function parameters of each channel in the filter to adapt to the observation noise that varies dynamically with the motion of the carrier.

[0063] Specifically, in step S1, the state vector of the established GNSS / INS tight integration system is:

[0064] ;

[0065] in, , , , , , and They are the position vector error, velocity error, misalignment angle, gyroscope bias, accelerometer bias, gyroscope scale factor error and accelerometer scale factor error of the INS state recursion result in the ECEF coordinate system;

[0066] According to the error disturbance theory, the discrete state equation of INS error is established as:

[0067] ;

[0068] in, , They are , The state vector at the moment; for The system noise at the moment, is the state transfer matrix;

[0069] Among them, the state transfer matrix for:

[0070]

[0071] in, is the system discrete time interval, is the Earth's rotation angular velocity vector, is the direction cosine matrix from the carrier coordinate system b to the ECEF coordinate system, is the specific force output by the carrier accelerometer, , , , The gyro bias error, accelerometer bias error, gyro scale factor error, and accelerometer scale factor error are modeled as the relevant time parameters of the first-order Markov process respectively;

[0072] The observation equation is:

[0073] ;

[0074] Observable Assuming that there is observable satellites, that is, In the case of multiple observation channels, the specific observation quantity for:

[0075] ;

[0076] in, for Time to The pseudoranges observed by the observable satellites, Based on At this moment, INS delivers the position and The geometric distance calculated from the positions of the observable satellites; ; is the number of observable satellites;

[0077] for Observe the noise at all times; for The moment observation matrix is ​​known as Optimal estimation of the state at the moment and calculated from the ephemeris The satellite position at this moment is calculated and is a known quantity.

[0078] The relationship between the satellite pseudorange observation and the INS recursive state error (i.e., state quantity X) is established through the observation equation; Observable quantity at a time:

[0079] ;

[0080] in, For the The three-dimensional coordinates of the satellite positions, is the three-dimensional coordinate of the recursive position of INS at time k.

[0081] Specifically, when using INS sensor data to recursively infer the carrier motion state, at any time when the tightly integrated navigation system is running , it is generally known The optimal estimate of the state at time: position ,speed , direction cosine matrix (At the first moment when the integrated navigation system is started, k=1, then the optimal estimate at k=0 is the initial state of the system); and according to the current Gyroscope output angle increment at time and speed increment Perform pure inertial navigation state recursion:

[0082] The specific recursive process includes:

[0083] 1) Posture update:

[0084]

[0085]

[0086]

[0087]

[0088]

[0089] in, is the direction cosine matrix of the vector from the carrier b system to the ECEF coordinate system, and is also the target matrix for attitude update. The attitude update is the previous moment Update to current time . Yes The intermediate matrices are the direction cosine matrices for transforming the vector from the ECEF coordinate system at time k-1 to the ECEF coordinate system at time k and from the carrier b system at time k to the carrier b system at time k-1. for The Earth's rotation angle vector in the time interval is given by the Earth's rotation angular velocity (Known constants) are calculated. is the gyro output angle increment The three-axis angular increment of the carrier corrected by the equivalent rotating vector twin-sample algorithm.

[0090] 2) Speed ​​update:

[0091] First, according to The carrier's position latitude is known at all times ,longitude calculate :

[0092]

[0093]

[0094] The goal of velocity update is to find The velocity in the ECEF coordinate system at the moment , here we first The time represents the velocity in the ECEF coordinate system Convert to navigation coordinate system n and find , and then update in the n system to get , and finally switch to the ECEF coordinate system:

[0095] ;

[0096] And ask Time direction cosine matrix :

[0097] ;

[0098] It can be divided into three components: ;

[0099] The first component The second component has been obtained The expression is:

[0100] ;

[0101] is the vector representing the angular velocity of the Earth's rotation in the n system, represents the carrier speed, To represent the angular velocity vector of the carrier in the n system rotating relative to the ECEF coordinate system, the expression is:

[0102] ;

[0103] in, is the meridian curvature radius, is the radius of curvature of the Maoyou circle, From speed The easting and northing components of .

[0104] is the height component in the carrier position, is the carrier latitude. In theory, these state quantities need to be To reduce the error, use these state quantities in Moment and The value of time is linearly recursive, and we get The value of the moment is used instead.

[0105] The third component of , whose expression is:

[0106] ;

[0107] in, The expressions are:

[0108]

[0109]

[0110] Linear recursive value of time; They are the angular increment and velocity increment output by the sensor respectively.

[0111] get After that, it is transferred to the ECEF coordinate system to obtain the updated carrier velocity:

[0112] ;

[0113] In a short time and The difference is very small and they can be substituted for each other.

[0114] 3) Location Update:

[0115] The velocity is integrated by trapezoidal shape to update the position of the carrier in the ECEF coordinate system:

[0116] ;

[0117] In addition, the carrier position in the ECEF coordinate system is converted to the geographic coordinate system for use at the next moment, and is also output as more intuitive carrier position information:

[0118] ;

[0119] in, , e is the eccentricity constant of the earth ellipsoid, and N is the radius of curvature of the earth reference ellipsoid. and The calculations of h_k need to reference each other, so iterative calculations are required.

[0120] The above is the basic process of pure inertial navigation ground push, and the recursive result is is the theoretical value without error. However, due to the inevitable errors in the output of the gyroscope and accelerometer, the state obtained by recursion using the above method in the actual system is expressed as To indicate the state where no error correction is performed, the “e” symbol representing the projection coordinate system is ignored for the sake of convenience.

[0121] When performing integrated navigation, the INS sensor data is used to recursively infer the carrier's motion state (such as position, speed, etc.) to obtain the position after pure inertial navigation state recursion. ,speed , direction cosine matrix However, due to the existence of sensor errors, the derived state also contains errors. Then the error in the motion state is set as the state quantity of the information fusion algorithm , and establish an error model, then fuse the model's predicted error value with the navigation information brought by GNSS observations through Kalman filtering to obtain the optimal estimate of the error in the current motion state, and use this value to correct ,speed , direction cosine matrix The error in the current more accurate carrier motion state is obtained, and the optimal estimate of the state is obtained after correction using the optimal estimate of the error. , , , which will be used for system output and State recursion at each moment.

[0122] In an embodiment of the present invention, the multi-channel heterogeneous kernel generalized correlation entropy Kalman filter (MCMK-GMCKF) introduces generalized correlation entropy into Kalman filtering to replace the mean square value of the error as the optimization cost; a multi-channel heterogeneous kernel structure is adopted to achieve mutual isolation of kernel function shape parameters in different observation channels, so that the kernel function shape parameters are determined by the statistical characteristics of the observation noise in each channel; thereby, the MCMK-GMCKF can cope with various forms of non-Gaussian noise, and can adjust the shape parameters of the kernel function according to the differences in the noise statistical characteristics of GNSS observations in different channels, so as to achieve better filtering effects.

[0123] Specifically, in step S2, the process of estimating the INS state recursive error by MCMK-GMCKF includes:

[0124] 1) Perform one-step state prediction to obtain One-step estimation The covariance matrix of :

[0125]

[0126]

[0127] for The optimal estimate of the error state at time , For use and the state transition matrix Computational One-step estimate of the error state at time , One-step estimate The covariance matrix of for The moment estimation error covariance matrix, for The system noise covariance matrix at time instant;.

[0128] Here and It will be used to calculate later The optimal estimate of the error state at the moment ;

[0129] 2) According to One-step estimation The covariance matrix of and the covariance matrix of the observation noise Perform Cholesky decomposition;

[0130] ;

[0131] is the covariance matrix of the observation noise in the observation equation.

[0132] 3) According to the Cholesky decomposition results, based on One-step estimation Filter residual Iterative calculation of ;

[0133] ;

[0134] in, , for The moment observation; n is the dimension of the state vector;

[0135] , for The moment observation matrix;

[0136] When step 3) is executed for the first time at the same time, Take the , and then calculate according to the iterative value; Represents the number of iterations, with an initial value of 1.

[0137] 4) According to The filter residual of the iteration is calculated The generalized correlation entropy weighted values ​​of each observation channel of the iteration constitute the generalized correlation entropy weighted matrix ;

[0138] In the iterative process, the collected The observation error samples of the observation window near the moment are used to estimate the moment, and the weighted matrix is ​​calculated based on the moment estimation result. The shape parameters of the kernel functions of each channel in are updated;

[0139] time The weight matrix of the iteration :

[0140] ;

[0141] in, For the Channel kernel function,

[0142] Kernel Function The expression is:

[0143] ;

[0144] For the The shape parameter of the kernel function of the observation channel, is the kernel bandwidth. Before the first step 3), the shape parameters of each channel are initially set to 2, that is, the standard Gaussian distribution form. Set to .

[0145] In the subsequent iterations, each channel The parameters will be dynamically adjusted according to step 3) to reflect the characteristics of the different cores of each channel to cope with the differentiated noise of each channel.

[0146] 5) Using the generalized correlation entropy weighted matrix Calculate Kalman gain ;

[0147] ;

[0148] ;

[0149] in, is the observation error covariance matrix of each observation channel weighted by relevant entropy;

[0150] 6) According to Kalman gain Perform state estimation to obtain the Iteration value ; When the The iteration value is the same as the When the difference of the iteration value is greater than the threshold, return to step 3) and execute again; otherwise, get the current moment The best estimate of the state quantity .

[0151] Specifically, ;

[0152] For the The iteration value obtained by the first iteration is The result of the iteration is calculated.

[0153] At this time, if:

[0154]

[0155] Jump to step 3) and re-execute, otherwise get the best estimate of the state quantity at the current moment ; The threshold to set.

[0156] use The best estimate of the state quantity at the moment In , , Correction of the position of the INS recursive state ,speed , direction cosine matrix , get the current position after the estimated error correction ,speed , direction cosine matrix ;

[0157] ;

[0158] .

[0159] In step S3, the kernel function parameters of each GNSS observation channel are estimated in real time using a moment matching estimator to achieve filter adaptation to the time-varying noise of each channel. Moment It can be expressed as:

[0160] ;

[0161] in , is the Laplace function. In the formula, Only with shape parameters About and standard deviation So we can use the sample second-order moment estimate , and then use the sample Moment Estimation Then reversely estimate the shape parameters .

[0162] Specifically, the updating of the kernel function parameters of the i-th channel in step S3 includes:

[0163] 1) Set the length of the near-time observation window to , when the system is running greater than or equal to At each moment, the corrected position is obtained. ;

[0164] 2) According to the corrected position at each moment Calculate the Observation channels approaching The error sample sequence observed at each moment;

[0165] When the filtering process runs to At time , calculate the Observation channels approaching The observation error sequence at each moment ;in,

[0166] ;

[0167] for Moment The a posteriori error of the satellite observations is for Time to The pseudoranges observed by the satellites, For the Satellites in The optimal estimate of the pseudorange at time is obtained by Three-dimensional coordinates of satellite position at the moment And the corrected 3D coordinates of the user position It is obtained by taking the difference modulo the vector.

[0168] 3) Calculate the Observation channels approaching The second-order moment and fourth-order moment estimates of the observation error sample sequence at each moment;

[0169] The second-order moment estimate of the sample sequence and the fourth moment estimate for:

[0170] ;

[0171] ;

[0172] is the mean of samples at N moments;

[0173] 4) Use the kurtosis coefficient as the moment matching estimator to calculate the estimate of the shape parameter based on the second-order moment and fourth-order moment estimates ;use Update step S2 Channel kernel function parameter.

[0174] Specifically, the kurtosis coefficient of the generalized Gaussian distribution function is:

[0175] ;

[0176] pass The inverse function of , by the second-order moment and the fourth moment Calculate shape parameters Estimated value of ;Right now

[0177] ;

[0178] Depend on The expression of is calculated by linear interpolation; the second-order moment and the fourth moment Using the second moment estimate and the fourth moment estimate Substitute, calculate Then update the shape parameters of the kernel function in step S2.

[0179] According to the above-mentioned updating method of the kernel function parameters of the ith channel, the shape parameters of all channels are estimated respectively, and the kernel parameters are adjusted so that the integrated navigation system can autonomously adapt to the observation noise that varies dynamically with the motion of the carrier.

[0180] In summary, the adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy disclosed in the embodiment of the present invention improves the robustness of the GNSS / INS integrated navigation system in the face of various non-Gaussian noises, and can perform adaptive parameter adjustment to cope with dynamic time-varying noise. In addition, the isolation of the kernel parameters of each observation channel makes the filter more advantageous when facing the situation where each channel contains differential noise.

[0181] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. An adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy, characterized in that: include: Step S1, establishing the discrete state equation and observation equation of INS error of GNSS / INS tightly integrated system; And use INS sensor data to recursively infer the carrier's motion state; Step S2: Based on the observations of the GNSS / INS tightly integrated system, a multi-channel heterogeneous kernel generalized correlation entropy Kalman filter is used to estimate the INS recursive error, and the estimated error value is used to perform feedback correction on the recursive carrier motion state; Step S3, using a moment matching estimator to update the kernel function parameters of each channel in the filter to adapt to the observation noise that varies dynamically with the motion of the carrier; The multi-channel hetero-kernel generalized correlation entropy Kalman filter introduces generalized correlation entropy into Kalman filtering; adopts a multi-channel hetero-kernel structure, and the shape parameters of the kernel function in each observation channel are isolated from each other and determined by the statistical characteristics of the observation noise in each channel; The process of using a multi-channel heterogeneous kernel generalized correlation entropy Kalman filter to perform INS recursive error estimation includes: 1) Perform one-step state prediction to obtain One-step estimation The covariance matrix of : 2) According to One-step estimation The covariance matrix of and the covariance matrix of the observation noise Perform Cholesky decomposition; 3) According to the Cholesky decomposition results, based on One-step estimation Filter residual Iterative calculation of ; Indicates the number of iterations, the initial value is 1; 4) According to The filter residual of the iteration is calculated The generalized correlation entropy weighted values ​​of each observation channel of the iteration constitute the generalized correlation entropy weighted matrix ; In the iterative process, the collected The observation error samples of the observation window near the moment are used to estimate the moment, and the weighted matrix is ​​calculated based on the moment estimation result. The shape parameters of the kernel functions of each channel in are updated; 5) Using the generalized correlation entropy weighted matrix Calculate Kalman gain ; 6) According to Kalman gain Perform state estimation to obtain Iteration value ; When the The iteration value is the same as the When the difference of the iteration value is greater than the threshold, return to step 3) and execute again; otherwise, get the current moment The best estimate of the state quantity .

2. The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy according to claim 1, characterized in that: In step S1, the state vector in the discrete state equation is: ； in, , , , , , and They are the position vector error, velocity error, misalignment angle, gyroscope bias, accelerometer bias, gyroscope scale factor error and accelerometer scale factor error of the INS state recursion result in the ECEF coordinate system; The observed quantity in the observation equation is: ； in, for Time to The pseudoranges observed by the observable satellites, For the The three-dimensional coordinates of the satellite positions, is the three-dimensional coordinate of the position recursively derived by INS at time k; i=1,…,m, where m is the number of observable satellites.

3. The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy according to claim 1, characterized in that: time The weight matrix of the iteration : ; in, For the Channel kernel function, Kernel Function The expression is: ； For the The shape parameter of the kernel function of the observation channel, is the core bandwidth.

4. The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy according to claim 1, characterized in that: The estimated error value is used to perform feedback correction on the recursive carrier motion state. use The best estimate of the state quantity at the moment In , , Correction of the position of the INS recursive state ,speed , direction cosine matrix , get the current position after the estimated error correction ,speed , direction cosine matrix ; ； 。 5. The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy according to claim 1, characterized in that: In step S3, Updates to the channel kernel function parameters include: 1) Set the length of the near-time observation window to , when the system is running greater than or equal to At each moment, the corrected position is obtained. ; 2) According to the corrected position at each moment Calculate the Observation channels approaching The error sample sequence observed at each moment; 3) Calculate the Observation channels approaching The second-order moment and fourth-order moment estimates of the observation error sample sequence at each moment; 4) Use the kurtosis coefficient as the moment matching estimator to calculate the estimate of the shape parameter based on the second-order moment and fourth-order moment estimates ;use Update step S2 Channel kernel function parameter.

6. The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy according to claim 5, characterized in that: When the filtering process runs to At time , calculate the Observation channels approaching The observation error sequence at each moment ;in, ； for Moment The a posteriori error of the satellite observations is for Time to The pseudoranges observed by the satellites, For the Satellites in The optimal estimate of the pseudorange at time is obtained by Three-dimensional coordinates of satellite position at the moment And the corrected 3D coordinates of the user position It is obtained by taking the difference modulo the vector.

7. The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy according to claim 6, characterized in that: Second-order moment estimate of the sample series and the fourth moment estimate for: ； ； is the mean of samples at N moments.

8. The adaptive robust GNSS / INS integrated navigation information fusion method based on generalized maximum correlation entropy according to claim 7, characterized in that: The kurtosis coefficient of the generalized Gaussian distribution function is: ； pass The inverse function of , by the second-order moment and the fourth moment Calculate shape parameters Estimated value of ; is the Laplace function; ； Depend on The expression of is calculated by linear interpolation; the second-order moment and the fourth moment Using the second moment estimate and the fourth moment estimate Substitute, calculate Then update the shape parameters of the kernel function in step S2.

Images