A vehicle positioning method system, device, medium and product based on motion constraint adaptive filtering

Abstract

The application discloses a vehicle positioning method system, device, medium and product based on motion constraint adaptive filtering, and relates to the technical field of vehicle navigation positioning. The method comprises the following steps: based on historical preprocessed data, a virtual dead reckoning state model and a virtual dead reckoning measurement model are constructed, and a virtual dead reckoning state vector is determined, so as to determine an INS installation angle and an NHC rod arm; based on INS measurement data, a combined system state model and a mechanical arrangement model are constructed, and three-dimensional positions and three-dimensional velocities of the INS installation angle and the NHC rod arm are determined; original data of a current epoch are acquired; when the original data do not contain GNSS measurement data, a three-dimensional position after filtering update is determined according to a first positioning process; and when the original data contain GNSS measurement data, a three-dimensional position after filtering update of the vehicle is determined according to a second positioning process. The application improves the positioning precision of the vehicle in a GNSS denial environment.

Description

Technical Field

[0001] This application relates to the field of vehicle navigation and positioning technology, and in particular to a vehicle positioning method, system, device, medium, and product based on motion-constrained adaptive filtering. Background Technology

[0002] Continuous, reliable, and real-time location information is essential for vehicle navigation. Integrated navigation based on the Global Navigation Satellite System (GNSS) and Inertial Navigation System (INS) is one of the most widely used positioning solutions, providing vehicles with accurate location information. However, in complex environments such as tunnels, overpasses, and urban canyons, GNSS signals are easily blocked and interfered with, leading to discontinuous and unreliable positioning. Without adding extra sensors, nonholonomic constraints (NHC), an effective and low-cost vehicle GNSS / INS enhancement technology, can significantly limit the error accumulation rate of INS in GNSS-denied environments. However, accurate compensation of the mounting angle and lever arm, along with proper configuration of noise parameters, are prerequisites for NHC to function effectively; imprecise parameter configuration will reduce vehicle positioning accuracy.

[0003] However, in practical applications, the INS installation angle, NHC lever arm, and model noise parameters all exhibit complex nonlinear coupling relationships with factors such as road conditions, vehicle motion state, observation environment, and INS installation location. Manual parameter tuning alone is insufficient to obtain the optimal configuration. Estimating the INS installation angle and NHC lever arm, and accurately determining the noise statistics of the NHC during information fusion, significantly limits the potential of implicit information to improve the navigation accuracy of the integrated system, resulting in lower GNSS positioning accuracy for vehicles in denied environments. Summary of the Invention

[0004] The purpose of this application is to provide a vehicle positioning method, system, device, medium, and product based on motion-constrained adaptive filtering to solve the problem of low positioning accuracy of GNSS for vehicles in denied environments.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] Firstly, this application provides a vehicle localization method based on motion-constrained adaptive filtering, comprising:

[0007] Historical measurement data is acquired and preprocessed to obtain historical preprocessed data. This historical measurement data includes GNSS and INS measurement data from multiple historical epochs. The GNSS measurement data includes multi-frequency, multi-mode pseudorange and carrier phase measurements from the GNSS receiver. The INS measurement data includes three-dimensional velocity increments measured by the accelerometer and three-dimensional angle increments measured by the gyroscope in the INS. Preprocessing includes quality assessment and integrated navigation solution. The historical preprocessed data also includes GNSS / INS navigation information from multiple historical epochs, including the vehicle's three-dimensional position increment, three-dimensional position, pitch angle, roll angle, and heading angle.

[0008] Based on historical preprocessed data, a virtual dead reckoning status model and a virtual dead reckoning measurement model are constructed.

[0009] Using Kalman filtering, the virtual dead reckoning state vector is determined based on the virtual dead reckoning state model and the virtual dead reckoning measurement model.

[0010] Based on the virtual dead reckoning state vector, the INS installation angle and NHC lever arm are determined; the INS installation angle is the Euler angle from the vehicle system to the load system, and the NHC lever arm is the distance vector from the center of the INS to the effective point of the NHC;

[0011] Obtain the raw data for the current epoch; the raw data includes: GNSS measurement data and / or INS measurement data;

[0012] Based on INS measurement data, a combined system state model and a mechanical arrangement model are constructed.

[0013] Based on the INS mounting angle, NHC lever arm, and mechanical arrangement model, the three-dimensional position and three-dimensional velocity of the vehicle's compensated INS mounting angle and NHC lever arm are determined.

[0014] When the original data does not contain GNSS measurement data, the filtered and updated three-dimensional position of the vehicle is determined according to the first positioning process to achieve vehicle positioning; wherein, the first positioning process includes:

[0015] Based on the vehicle's compensated INS mounting angle and the three-dimensional and nominal velocities of the NHC lever arm, an NHC measurement model is constructed.

[0016] Based on the NHC measurement model, the measurement noise covariance matrix is ​​reconstructed using the inverse gamma distribution, resulting in an NHC measurement model that includes the reconstructed measurement noise covariance matrix.

[0017] Based on the combined system state model and the NHC measurement model containing the reconstructed measurement noise covariance matrix, variational Bayesian adaptive filtering is used to determine the combined system state vector.

[0018] Based on the combined system state vector and the vehicle's compensated INS mounting angle and NHC lever arm 3D position, the vehicle's filtered updated 3D position is determined.

[0019] When the raw data includes GNSS measurement data, the filtered and updated three-dimensional position of the vehicle is determined according to the second positioning process to achieve vehicle positioning; wherein the second positioning process includes:

[0020] Based on the vehicle's compensated INS installation angle and the three-dimensional position of the NHC lever arm, and the GNSS measurement data in the original data, a combined system measurement model is constructed.

[0021] Based on the combined system state model and the combined system measurement model, Kalman filtering is used to determine the combined system state vector.

[0022] Based on the combined system state vector and the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC lever arm, the filtered updated three-dimensional position of the vehicle is determined.

[0023] Optionally, the virtual dead reckoning state model includes:

[0024]

[0025] in, δP represents the k-th epoch. n δP n This represents the three-dimensional position error of the vehicle in the n-frame, where the n-frame is the navigation frame. δP represents the k-1 epoch. n ; Representing the k-1 epoch Let represent the direction cosine matrix from the b-system to the n-system, where the b-system is the load system; Representing the k-1 epoch The direction cosine matrix from the v-frame to the b-frame is represented by M; the auxiliary matrix is ​​represented by α. k-1 Let α represent the pitch and yaw installation angle errors of the INS at epoch k-1. φ represents the three-dimensional position increment of the vehicle in the n-frame at epoch k; k-1 φ represents the pitch misalignment error, roll misalignment error, and yaw misalignment error of the vehicle at epoch k-1; δk k-1 Let δk represent the k-1 epoch, where δk represents the virtual odometer calibration factor error.

[0026] Optionally, the virtual dead reckoning measurement model includes:

[0027]

[0028] Among them, zDR,k This represents the measurement vector of the virtual dead reckoning model at epoch k; P represents the k-th epoch obtained through the positional recursion equation. n P n This represents the three-dimensional position of the vehicle in the n-frame. P represents the k-th epoch of the historical preprocessed data. n ; l represents the k-1 epoch. b , l b Indicates the NHC lever arm; × indicates the vector cross product operation; δl represents the k-th epoch. b ,δl b Indicates the NHC lever arm error; e DR,k H represents the measurement noise vector of the virtual dead reckoning measurement model at epoch k; DR,k This represents the coefficient matrix of the virtual dead reckoning measurement model at epoch k; x k Let x represent the k-th epoch, where x represents the state vector of the combined system.

[0029] Optionally, the combined system state model includes:

[0030]

[0031] in, Represents the differential of φ; This represents the three-dimensional rotational angular velocity of the n-frame relative to the i-frame, where the i-frame is an inertial frame. express The calculation error; This indicates the three-dimensional measurement error of the gyroscope; Indicates δv n The differential; δv n f represents the three-dimensional velocity error of the vehicle in the n-frame; b This represents the three-dimensional velocity increment measured by the accelerometer; express The calculation error; This represents the three-dimensional angular velocity of Earth's rotation in the n-frame. express The calculation error; δf represents the angular velocity of a three-dimensional carrier in the n-system. b δg represents the three-dimensional measurement error of the accelerometer. n Indicates three-dimensional gravity error; Indicates δP n The differential.

[0032] Optionally, the mechanical arrangement model includes:

[0033]

[0034] in, P represents the k-1 epoch obtained through the position recursion equation. n ; τ represents the three-dimensional velocity of the vehicle in the n-frame at epoch k; τ represents the sampling interval of the INS. This represents the three-dimensional velocity of the vehicle in the n-th frame at epoch k-1; Represents the specific force integral term for epoch k; This represents the harmful acceleration integral term at epoch k; Representing the k-epoch This represents the direction cosine matrix from epoch k-1 to epoch k in the n-system. Let represent the direction cosine matrix from epoch k to epoch k-1 in the b system.

[0035] Optionally, the formulas for calculating the vehicle's compensated INS mounting angle and the three-dimensional velocity of the NHC lever arm under the v-frame include:

[0036]

[0037] in, This represents the three-dimensional velocity of the vehicle in the v-frame at epoch k, where the INS mounting angle and NHC lever arm are the compensated values. This represents the direction cosine matrix from the b-system to the v-system calculated based on the INS installation angle; This represents the three-dimensional angle increment measured by the gyroscope at epoch k; Representing the k-epoch

[0038] Secondly, this application provides a vehicle positioning system based on motion-constrained adaptive filtering, used to implement the vehicle positioning method based on motion-constrained adaptive filtering described in any of the above claims, wherein the vehicle positioning system based on motion-constrained adaptive filtering includes:

[0039] The data acquisition module is used to acquire historical measurement data and preprocess it to obtain historical preprocessed data. The historical measurement data includes GNSS and INS measurement data from multiple historical epochs. The GNSS measurement data includes multi-frequency, multi-mode pseudorange and carrier phase measurements from the GNSS receiver. The INS measurement data includes three-dimensional velocity increments measured by the accelerometer and three-dimensional angle increments measured by the gyroscope in the INS. Preprocessing includes quality assessment and integrated navigation solution. The historical preprocessed data includes GNSS / INS navigation information from multiple historical epochs, including the vehicle's three-dimensional position increment, three-dimensional position, pitch angle, roll angle, and heading angle.

[0040] The first model building module is used to build a virtual dead reckoning status model and a virtual dead reckoning measurement model based on historical preprocessed data.

[0041] The virtual dead reckoning state vector determination module is used to determine the virtual dead reckoning state vector based on the virtual dead reckoning state model and the virtual dead reckoning measurement model using Kalman filtering.

[0042] The INS installation angle and NHC lever arm determination module is used to determine the INS installation angle and NHC lever arm based on the state vector calculated from the virtual dead reckoning. The INS installation angle is the Euler angle from the vehicle system to the load system, and the NHC lever arm is the distance vector from the center of the INS to the effective point of the NHC.

[0043] The raw data acquisition module is used to acquire the raw data of the current epoch; the raw data includes: GNSS measurement data and / or INS measurement data;

[0044] The second model building module is used to build a combined system state model and a mechanical arrangement model based on INS measurement data;

[0045] The speed compensation module is used to determine the three-dimensional position and three-dimensional speed of the vehicle's compensated INS mounting angle and NHC lever arm based on the INS mounting angle, NHC lever arm and mechanical arrangement model.

[0046] The first positioning module is used to determine the filtered and updated three-dimensional position of the vehicle according to the first positioning process when the original data does not contain GNSS measurement data, thereby achieving vehicle positioning; wherein the first positioning process includes:

[0047] Based on the vehicle's compensated INS mounting angle and the three-dimensional and nominal velocities of the NHC lever arm, an NHC measurement model is constructed.

[0048] Based on the NHC measurement model, the measurement noise covariance matrix is ​​reconstructed using the inverse gamma distribution, resulting in an NHC measurement model that includes the reconstructed measurement noise covariance matrix.

[0049] Based on the combined system state model and the NHC measurement model containing the reconstructed measurement noise covariance matrix, variational Bayesian adaptive filtering is used to determine the combined system state vector.

[0050] Based on the combined system state vector and the vehicle's compensated INS mounting angle and NHC lever arm 3D position, the vehicle's filtered updated 3D position is determined.

[0051] The second positioning module is used to determine the filtered and updated three-dimensional position of the vehicle according to the second positioning process when the original data includes GNSS measurement data, thereby achieving vehicle positioning; wherein the second positioning process includes:

[0052] Based on the vehicle's compensated INS installation angle and the three-dimensional position of the NHC lever arm, and the GNSS measurement data in the original data, a combined system measurement model is constructed.

[0053] Based on the combined system state model and the combined system measurement model, Kalman filtering is used to determine the combined system state vector.

[0054] Based on the combined system state vector and the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC lever arm, the filtered updated three-dimensional position of the vehicle is determined.

[0055] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the vehicle positioning method based on motion constraint adaptive filtering as described in any of the preceding claims.

[0056] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the vehicle positioning method based on motion constraint adaptive filtering as described in any of the preceding claims.

[0057] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the vehicle positioning method based on motion constraint adaptive filtering as described in any of the preceding claims.

[0058] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0059] This application discloses a vehicle positioning method, system, device, medium, and product based on motion-constrained adaptive filtering. First, historical preprocessed data undergoes a series of processing and modeling steps to determine the INS mounting angle and NHC lever arm. Then, based on INS measurement data, a combined system state model and a mechanical arrangement model are constructed. Based on the INS mounting angle, NHC lever arm, and mechanical arrangement model, the three-dimensional position and three-dimensional velocity of the vehicle's compensated INS mounting angle and NHC lever arm are determined. Finally, during vehicle positioning in the current epoch: when the original data includes GNSS measurement data, a combined system measurement model is constructed based on the vehicle's compensated INS mounting angle and NHC lever arm's three-dimensional position and the GNSS measurement data in the original data. Based on the combined system state model and the combined system measurement model, Kalman filtering is used to determine the combined system state vector. Based on the combined system state vector and the vehicle's compensated INS mounting angle and NHC lever arm's three-dimensional position, the filtered and updated three-dimensional position of the vehicle is determined. When the original data does not include GNSS measurement data, an NHC measurement model is constructed based on the vehicle's compensated INS installation angle and the 3D and nominal velocities of the NHC arm. Based on this NHC model, the measurement noise covariance matrix is ​​reconstructed using an inverse gamma distribution, resulting in an NHC measurement model containing the reconstructed measurement noise covariance matrix. Based on the combined system state model and the NHC measurement model containing the reconstructed measurement noise covariance matrix, variational Bayesian adaptive filtering is used to determine the combined system state vector. Based on the combined system state vector and the vehicle's compensated INS installation angle and the 3D position of the NHC arm, the vehicle's filtered and updated 3D position is determined. Because the variational Bayesian method can adaptively adjust the NHC measurement noise covariance matrix based on prior and observation information, it can improve the positioning accuracy and reliability of vehicles in GNSS-denied environments such as urban canyons, overpasses, and tunnels. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 A schematic flowchart of a vehicle localization method based on motion-constrained adaptive filtering provided in an embodiment of this application;

[0062] Figure 2 Schematic diagram of INS installation angle and NHC lever arm;

[0063] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0064] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0065] The purpose of this application is to provide a vehicle positioning method, system, device, medium, and product based on motion-constrained adaptive filtering, which aims to improve the positioning accuracy of vehicles in GNSS-denied environments.

[0066] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, this application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0067] In one exemplary embodiment, such as Figure 1 As shown, the vehicle localization method based on motion-constrained adaptive filtering in this embodiment includes:

[0068] Step 1: Obtain historical measurement data and preprocess the historical measurement data to obtain historical preprocessed data.

[0069] The historical measurement data includes: GNSS measurement data and INS measurement data from multiple historical epochs; the GNSS measurement data includes: multi-frequency and multi-mode pseudorange and carrier phase measurements measured by the GNSS receiver; the INS measurement data includes: three-dimensional velocity increments measured by the accelerometer in the INS and three-dimensional angle increments measured by the gyroscope in the INS; preprocessing includes: quality assessment and integrated navigation solution; the historical preprocessed data includes: GNSS / INS navigation information from multiple historical epochs; the GNSS / INS navigation information includes: vehicle three-dimensional position increments, three-dimensional position, pitch angle, roll angle, and heading angle.

[0070] Specifically, preprocessing mainly involves quality assessment of historical measurement data and extracting a segment of historical measurement data containing both straight-line and turning data for combined navigation calculation. The purpose is to ensure that the extracted historical measurement data has good observability so that it can be used to determine the INS installation angle and NHC lever arm in subsequent steps.

[0071] Step 2: Based on historical preprocessed data, construct a virtual dead reckoning status model and a virtual dead reckoning measurement model.

[0072] As an optional implementation method, the virtual dead reckoning state model includes:

[0073]

[0074] in, δP represents the k-th epoch. n δP n This represents the three-dimensional position error of the vehicle in the n-frame, where the n-frame is the navigation frame. δP represents the k-1 epoch. n ; Representing the k-1 epoch Let represent the direction cosine matrix from the b-system to the n-system, where the b-system is the load system; Representing the k-1 epoch The direction cosine matrix from the v-frame to the b-frame is represented by M; the auxiliary matrix is ​​represented by α. k-1 Let α represent the pitch and yaw installation angle errors of the INS at epoch k-1. φ represents the three-dimensional position increment of the vehicle in the n-frame at epoch k; k-1 φ represents the pitch misalignment error, roll misalignment error, and yaw misalignment error of the vehicle at epoch k-1; δk k-1 Let δk represent the k-1 epoch, where δk represents the virtual odometer calibration factor error.

[0075] As an optional implementation method, the virtual dead reckoning measurement model includes:

[0076]

[0077] Among them, z DR,k This represents the measurement vector of the virtual dead reckoning model at epoch k; P represents the k-th epoch obtained through the positional recursion equation. n P n This represents the three-dimensional position of the vehicle in the n-frame. P represents the k-th epoch of the historical preprocessed data. n ; l represents the k-1 epoch. b , l b Indicates the NHC lever arm; × indicates the vector cross product operation; δl represents the k-th epoch. b ,δl b Indicates the NHC lever arm error; e DR,k H represents the measurement noise vector of the virtual dead reckoning measurement model at epoch k; DR,k This represents the coefficient matrix of the virtual dead reckoning measurement model at epoch k; x k Let x represent the k-th epoch, where x represents the state vector of the combined system.

[0078] Specifically, the position recursion equation is as follows:

[0079]

[0080] in, P represents the k-1 epoch obtained through the position recursion equation. n ; This represents the increment of the vehicle's three-dimensional position in the v frame at epoch k.

[0081] The expression for x is:

[0082] x=[(δP n ) T α T φ T (δl b ) T δk] T .

[0083] For α, φ, δl b Both δk and δk are modeled as random walk processes.

[0084] Step 3: Using Kalman filtering, determine the virtual dead reckoning state vector based on the virtual dead reckoning state model and the virtual dead reckoning measurement model.

[0085] Specifically, the expression for step 3 is:

[0086]

[0087]

[0088] Among them, K DR,k This represents the gain matrix for virtual dead reckoning at epoch k; H represents the one-step prediction covariance matrix of the virtual dead reckoning state model at epoch k; DR,k C represents the coefficient matrix of the virtual dead reckoning measurement model at epoch k; DR,k This represents the measurement noise covariance matrix of the virtual dead reckoning measurement model at epoch k; Represents the virtual dead reckoning state vector at epoch k; This represents the one-step predicted state vector of the virtual dead reckoning state model at epoch k; This represents the error covariance matrix for virtual dead reckoning at epoch k.

[0089] Step 4: Calculate the state vector based on the virtual dead reckoning and determine the INS installation angle and NHC lever arm.

[0090] Among them, such as Figure 2As shown, the INS mounting angle is the Euler angle from the vehicle system (V-series) to the load system (B-series). The essence of the INS mounting angle is due to the incomplete alignment of the V-series and B-series during INS installation, and it is generally a small angle. The NHC lever arm is the distance vector from the center of the INS to the effective point of the NHC. Figure 2 In the middle, x v Let x be the x-axis of the v-frame, and y be the y-axis. v Let z be the y-axis of the v-frame. v Let x be the z-axis of the v-frame. b Let x be the x-axis of the b-system, y b Let z be the y-axis of the b-system. b Let z be the z-axis of the b-system.

[0091] Specifically, for Perform Euler angle conversion to obtain the INS mounting angle. The calculation formulas for the NHC lever arm are as follows:

[0092]

[0093] in, Representing the k-epoch The direction cosine matrix from epoch k-1 to epoch k in the v system can be calculated from the pitch and heading installation angle errors of the INS estimated by filtering. Indicates l under epoch k b Since the above two equations are updated recursively epoch-by-epoch, considering the convergence and steady-state nature of the filter, the INS installation angle and NHC arm calculated in the last epoch of the historical preprocessed data are saved here as the INS installation angle and NHC arm in step 4.

[0094] Step 5: Obtain the raw data for the current epoch; the raw data includes: GNSS measurement data and / or INS measurement data.

[0095] Specifically, if the current epoch is obstructed by urban high-rise buildings, roadside trees, overpasses, and tunnels, and the GNSS receiver cannot receive GNSS measurement data, then the raw data for the current epoch will not contain GNSS measurement data.

[0096] Step 6: Based on INS measurement data, construct the combined system state model and mechanical arrangement model.

[0097] As an optional implementation, the combined system state model includes:

[0098]

[0099] in, Represents the differential of φ; This represents the three-dimensional rotational angular velocity of the n-frame relative to the i-frame, where the i-frame is an inertial frame. express The calculation error; This indicates the three-dimensional measurement error of the gyroscope; Indicates δv n The differential; δv n f represents the three-dimensional velocity error of the vehicle in the n-frame; b This represents the three-dimensional velocity increment measured by the accelerometer; express The calculation error; This represents the three-dimensional angular velocity of Earth's rotation in the n-frame. express The calculation error; δf represents the angular velocity of a three-dimensional carrier in the n-system. b δg represents the three-dimensional measurement error of the accelerometer. n Indicates three-dimensional gravity error; Indicates δP n The differential.

[0100] As an alternative implementation, the mechanical choreography model includes:

[0101]

[0102] in, P represents the k-1 epoch obtained through the position recursion equation. n ; τ represents the three-dimensional velocity of the vehicle in the n-frame at epoch k; τ represents the sampling interval of the INS. This represents the three-dimensional velocity of the vehicle in the n-th frame at epoch k-1; Represents the specific force integral term for epoch k; This represents the harmful acceleration integral term at epoch k; Representing the k-epoch This represents the direction cosine matrix from epoch k-1 to epoch k in the n-system. Let represent the direction cosine matrix from epoch k to epoch k-1 in the b system.

[0103] Step 7: Based on the INS mounting angle, NHC lever arm, and mechanical arrangement model, determine the three-dimensional position and three-dimensional velocity of the vehicle's compensated INS mounting angle and NHC lever arm.

[0104] As an optional implementation, the formulas for calculating the three-dimensional velocities of the vehicle's compensated INS mounting angle and NHC lever arm under v-frame conditions include:

[0105]

[0106] in, This represents the three-dimensional velocity of the vehicle in the v-frame at epoch k, where the INS mounting angle and NHC lever arm are the compensated values. This represents the direction cosine matrix from the b-system to the v-system calculated based on the INS installation angle; This represents the three-dimensional angle increment measured by the gyroscope at epoch k; Representing the k-epoch

[0107] Step 8: When the original data does not contain GNSS measurement data, determine the filtered and updated three-dimensional position of the vehicle according to the first positioning process to achieve vehicle positioning.

[0108] The first positioning process includes:

[0109] Step 81: Construct the NHC measurement model based on the vehicle's compensated INS mounting angle and the three-dimensional and nominal velocities of the NHC lever arm.

[0110] Specifically, when the raw data does not contain GNSS measurement data, the combined system state vector is:

[0111]

[0112] in, This indicates that the gyroscope has three-dimensional zero bias in the b-frame. This indicates that the accelerometer has three-dimensional zero bias in the b-frame.

[0113] The three-dimensional zero bias error of the gyroscope and accelerometer is modeled as a first-order Markov process:

[0114]

[0115] in, express The differential; express The derivative; ζ represents the correlation time; w ε This represents the noise drive vector for zero bias of the gyroscope; This represents the noise driving vector for the accelerometer's zero bias.

[0116] When the raw data does not include GNSS measurement data, NHC-assisted vehicle navigation is implemented. Under the NHC assumption, the nominal speed of the vehicle is:

[0117] v NHC,k =[0v y,k 0] T .

[0118] Among them, v NHC,kv represents the nominal speed of the vehicle at epoch k. y,k This represents the forward velocity of the vehicle in the v system at epoch k.

[0119] Based on the vehicle's compensated INS mounting angle and the three-dimensional and nominal velocities of the NHC lever arm, an NHC measurement model is constructed as follows:

[0120]

[0121] Among them, z k This represents the measurement vector of the NHC measurement model at epoch k; Representing the k-epoch φ represents the direction cosine matrix from the n-system to the b-system. k φ represents the k-th epoch; δv represents the k-th epoch. n ; Representing the k-epoch e v,k This represents the measurement noise vector of the NHC measurement model at epoch k.

[0122] Without wheel odometers, the NHC measurement model expressed in the above formula can only constrain the vehicle's lateral and axial velocities, which can be transformed into the following matrix form:

[0123] z k =H v,k x k +e v,k .

[0124] Among them, H v,k This represents the coefficient matrix of the NHC measurement model at epoch k. v,k Calculated using the following formula:

[0125]

[0126] Among them, h 1,1 H represents v,k The element in the first row and first column; h 1,2 H represents v,k The element in the first row and second column; h 1,4 H represents v,k The element in the 1st row and 4th column; h 2,1 H represents v,k The element in the 2nd row and 1st column; h 2,2 H represents v,k The element in the second row and second column; h 2,4 H represents v,k The element in the 2nd row and 4th column; The first '·' of the matrix [·] OK.

[0127] Considering the NHC assumption that lateral velocity constraints and axial velocity constraints are unrelated, i.e.:

[0128] C v,k =diag(C x C z ).

[0129] Among them, C v,k The measurement noise covariance matrix of the NHC measurement model at epoch k is represented by diag(·); C represents the diagonal matrix. x Indicates the variance of NHC measurement noise in the lateral direction of the vehicle; C z This represents the variance of the NHC measurement noise in the vehicle's orientation.

[0130] Step 82: Based on the NHC measurement model, the measurement noise covariance matrix is ​​reconstructed using the inverse gamma distribution to obtain the NHC measurement model containing the reconstructed measurement noise covariance matrix.

[0131] Specifically, the NHC measurement noise covariance matrix is ​​reconstructed using the inverse gamma distribution based on the NHC measurement model, that is, the NHC measurement noise covariance matrix C... v,k Also considered as an unknown parameter. This is for the joint estimation of the NHC measurement noise covariance matrix C. v,k With state vector x k It is necessary to ensure C v,k The prior probability density function and the posterior probability density function have the same functional form.

[0132] First, perform the prior probability density function p(C) v,k |z 1:k-1 The variational update of ) is as follows:

[0133]

[0134] Among them, z 1:k-1 Represents the measurement vector of the NHC measurement model from epoch 1 to epoch k-1; IG(·,α) k|k-1,i ,β k|k-1,i ) indicates that the shape parameter is α k|k-1,i The scale parameter is β k|k-1,i The inverse gamma probability density function, α k|k-1,i Let β represent the i-th prior shape parameter at epoch k. k|k-1,i This represents the i-th prior scale parameter at epoch k; C represents v,k The i-th element on the diagonal; m represents C v,k The dimension of.

[0135] According to the Chapman-Kolmogorov equation, p(C v |z1:k-1 This can be represented as:

[0136] p(C v,k |z 1:k-1 )=∫p(C v,k |C v,k-1 )p(C v,k-1 |z 1:k-1 )dC v,k-1 .

[0137] Wherein, p(C v,k |C v,k-1 ) represents the dynamic model from epoch k-1 to epoch k; p(C v,k-1 |z 1:k-1 ) represents the measurement noise covariance matrix C of the NHC measurement model at epoch k-1. v,k-1 The posterior probability density function.

[0138] Because in practice p(C) v,k |C v,k-1 The parameters are generally unknown, and a forgetting factor ρ is usually used to pass the posterior estimated parameters from the previous epoch, as follows:

[0139]

[0140] Wherein, ρ takes values ​​in the range (0,1], reflecting the degree of fluctuation of measurement noise over time; α k-1|k-1,i Let β represent the posterior shape parameter at epoch k-1. k-1|k-1,i Let represent the i-th posterior scale parameter in epoch k-1.

[0141] Considering that the lateral and axial velocities of the vehicle under the v-series can be regarded as the fluctuation of measurement noise, the following elastic forgetting factors ρ1 and ρ2 are constructed:

[0142]

[0143] Where |·| represents the modulo operation; v x This represents the lateral velocity of the vehicle in the v-frame; v v Represents the three-dimensional velocity modulus of the vehicle; v z This indicates the vehicle's upward velocity in the v-frame.

[0144] Next, proceed with C. v,k The posterior probability density function p(C) v,k |z 1:k The variational update of ) is performed because the prior probability density function is modeled as an inverse gamma distribution, therefore the posterior probability density function p(C) v,k |z 1:k It also follows an inverse gamma distribution, as follows:

[0145]

[0146] Among them, z 1:k α represents the measurement vector of the NHC measurement model from epoch 1 to epoch k; k|k,i Let β represent the i-th posterior shape parameter at epoch k. k|k,i Let represent the i-th posterior scale parameter at epoch k.

[0147] Step 83: Based on the combined system state model and the NHC measurement model containing the reconstructed measurement noise covariance matrix, variational Bayesian adaptive filtering is used to determine the combined system state vector.

[0148] Specifically, based on variational Bayesian theory, C v,k The approximate posterior probability density Q (j+1) (C v,k ) and state vector x k The approximate posterior probability density Q (j+1) (x k The optimal solutions to all satisfy the following equation:

[0149]

[0150] Where ln[·] represents the natural logarithm; Q(θ) represents Q (j+1) (C v,k ) or Q (j+1) (x k E[·] represents the expectation operation; Represents Ξ k All elements except θ; c θ This represents a constant that is independent of θ.

[0151] Because of C v,k With x k They are mutually coupled, and Q cannot be directly calculated. (j+1) (C v,k ) and Q (j+1) (x k Therefore, a fixed-point iteration method is used to obtain the analytical solution for Q. (j+1) (C v,k ) and Q (j+1) (x k The local optimal solution of ).

[0152] First, let θ = C v,k According to variational Bayesian theory, Q (j+1) (C v,k It can be updated to the following inverse gamma distribution:

[0153]

[0154] Among them, Q(j+1) (·) represents the approximate posterior probability density of the (j+1)th iteration; shape parameter α k|k,i and scale parameter β k|k,i It can be updated to:

[0155]

[0156] Wherein, the auxiliary matrix for the j-th iteration at epoch k is... It can be calculated using the following formula:

[0157]

[0158] in, express The i-th element on the diagonal; E (j) [·] represents the expectation of the j-th iteration; This represents the state vector of the combined system in the j-th iteration. Let represent the error covariance matrix of the combined system in the j-th iteration.

[0159] The measurement noise covariance matrix of the NHC measurement model after the (j+1)th iteration at epoch k. Represented as;

[0160]

[0161] Where, α k|k,1 Let β represent the first posterior shape parameter at epoch k. k|k,1 α represents the first posterior scale parameter at epoch k; k|k,2 β represents the second posterior shape parameter at epoch k. k|k,2 This represents the second posterior scale parameter at epoch k.

[0162] Secondly, let θ = x k According to variational Bayesian theory, Q (j+1) (x k It can be updated to the following Gaussian distribution:

[0163]

[0164] in, The state vector is represented as The error covariance matrix is Gaussian probability density function; state vector of the combined system in the (j+1)th iteration With error covariance matrix Represented as:

[0165]

[0166] in, P represents the gain matrix of the (j+1)th iteration of the combined system at epoch k; k|k-1 Let represent the one-step prediction covariance matrix of the state model of the combined system at epoch k; This represents the one-step predicted state vector of the combined system state model at epoch k.

[0167] Step 84: Based on the combined system state vector and the vehicle's compensation INS mounting angle and the three-dimensional position of the NHC lever arm, determine the vehicle's filtered updated three-dimensional position.

[0168] Specifically, when the number of iterations reaches the preset maximum number of iterations, the α value of the current iteration is output. k|k β k|k , and Based on the obtained combined system state vector The three-dimensional position of the vehicle after filtering update is determined by the three-dimensional position of the compensated INS mounting angle and NHC lever arm.

[0169] Step 9: When the raw data contains GNSS measurement data, determine the filtered and updated three-dimensional position of the vehicle according to the second positioning process to achieve vehicle positioning.

[0170] The second positioning process includes:

[0171] Step 91: Based on the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC arm, and the GNSS measurement data in the original data, construct a combined system measurement model.

[0172] Specifically, when the raw data includes GNSS measurement data, the combined system state vector is:

[0173]

[0174] Where, N rb This represents the single-difference ambiguity vector between the base station r and the rover b.

[0175] Based on the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC boom, along with GNSS measurement data from the original data, a combined system measurement model is constructed. The combined system measurement model is as follows:

[0176]

[0177] H k x k +e k

[0178] in, This represents the difference operator, which means that a first difference is calculated between the base station and the rover station, and then a second difference is calculated between the reference satellite i and the non-reference satellite g. Representing the k-epoch This represents the pseudorange measurement value obtained by calculating the single difference between the base station and the rover station, and the quadratic difference between the reference satellite and the non-reference satellite. Representing the k-epoch This represents the pseudorange measurement value obtained by calculating the single difference between the base station and the rover station, and the quadratic difference between the reference satellite and the non-reference satellite, based on the three-dimensional position prediction of the vehicle's compensated INS installation angle and NHC arm. Representing the k-epoch This represents the carrier phase measurement value obtained by calculating the single difference between the base station and the rover station, and the quadratic difference between the reference satellite and the non-reference satellite; Representing the k-epoch This represents the carrier phase measurement value obtained by calculating the single difference between the base station and the rover station, and the quadratic difference between the reference satellite and the non-reference satellite, based on the three-dimensional position prediction of the vehicle's compensated INS installation angle and the NHC arm; μ i μ represents the reference satellite, where μ is the unit line-of-sight vector; μ g μ represents a non-reference satellite; Representing the k-epoch L represents the direction cosine matrix from the n-system to the e-system, where the e-system represents the Earth system; b This represents the distance vector from the INS center to the GNSS antenna phase center; λ represents the double-difference pseudorange measurement noise at k epochs, including unmodeled errors; λ represents the wavelength. H represents the double-difference carrier phase measurement noise at epoch k, including unmodeled errors; k This represents the coefficient matrix of the measurement model of the combined system at epoch k; e k This represents the measurement noise vector of the combined system measurement model at epoch k.

[0179] Step 92: Based on the combined system state model and the combined system measurement model, Kalman filtering is used to determine the combined system state vector.

[0180] Specifically, the state vector of the combined system is determined by Kalman filtering based on the combined system state model and the combined system measurement model, as follows:

[0181]

[0182] Among them, K k C represents the gain matrix of the combined system at epoch k; k This represents the measurement noise covariance matrix of the combined system measurement model at epoch k; Let represent the one-step prediction covariance matrix of the state model of the combined system at epoch k; This represents the state vector of the combined system at epoch k; This represents the one-step predicted state vector of the combined system state model at epoch k; Let represent the error covariance matrix of the combined system at epoch k.

[0183] Step 93: Based on the combined system state vector and the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC lever arm, determine the vehicle's filtered updated three-dimensional position.

[0184] In one exemplary embodiment, a vehicle positioning system based on motion constraint adaptive filtering is provided to implement any of the above-described vehicle positioning methods based on motion constraint adaptive filtering. The vehicle positioning system based on motion constraint adaptive filtering includes:

[0185] The data acquisition module is used to acquire historical measurement data and preprocess it to obtain historical preprocessed data. The historical measurement data includes GNSS measurement data and INS measurement data from multiple historical epochs. The GNSS measurement data includes multi-frequency and multi-mode pseudorange and carrier phase measurements measured by the GNSS receiver. The INS measurement data includes three-dimensional velocity increments measured by the accelerometer in the INS and three-dimensional angle increments measured by the gyroscope in the INS. The preprocessing includes quality assessment and integrated navigation solution. The historical preprocessed data includes GNSS / INS navigation information from multiple historical epochs. The GNSS / INS navigation information includes the vehicle's three-dimensional position increment, three-dimensional position, pitch angle, roll angle, and heading angle.

[0186] The first model building module is used to construct a virtual dead reckoning status model and a virtual dead reckoning measurement model based on historical preprocessed data.

[0187] The virtual dead reckoning state vector determination module is used to determine the virtual dead reckoning state vector based on the virtual dead reckoning state model and the virtual dead reckoning measurement model using Kalman filtering.

[0188] The INS installation angle and NHC lever arm determination module is used to calculate the state vector based on the virtual dead reckoning and determine the INS installation angle and NHC lever arm. The INS installation angle is the Euler angle from the vehicle system to the load system, and the NHC lever arm is the distance vector from the center of the INS to the effective point of the NHC.

[0189] The raw data acquisition module is used to acquire the raw data of the current epoch; the raw data includes: GNSS measurement data and / or INS measurement data.

[0190] The second model building module is used to build a combined system state model and a mechanical arrangement model based on INS measurement data.

[0191] The speed compensation module is used to determine the three-dimensional position and three-dimensional speed of the vehicle's compensated INS mounting angle and NHC lever arm based on the INS mounting angle, NHC lever arm, and mechanical arrangement model.

[0192] The first positioning module is used to determine the filtered and updated three-dimensional position of the vehicle according to the first positioning process when the original data does not contain GNSS measurement data, thereby achieving vehicle positioning; wherein the first positioning process includes:

[0193] An NHC measurement model is constructed based on the vehicle's compensated INS mounting angle and the three-dimensional and nominal velocities of the NHC lever arm.

[0194] Based on the NHC measurement model, the measurement noise covariance matrix is ​​reconstructed using the inverse gamma distribution, resulting in an NHC measurement model that includes the reconstructed measurement noise covariance matrix.

[0195] Based on the combined system state model and the NHC measurement model including the reconstructed measurement noise covariance matrix, variational Bayesian adaptive filtering is used to determine the combined system state vector.

[0196] Based on the combined system state vector and the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC lever arm, the filtered updated three-dimensional position of the vehicle is determined.

[0197] The second positioning module is used to determine the filtered and updated three-dimensional position of the vehicle according to the second positioning process when the original data includes GNSS measurement data, thereby achieving vehicle positioning; wherein the second positioning process includes:

[0198] Based on the vehicle's compensated INS installation angle and the three-dimensional position of the NHC arm, along with the GNSS measurement data from the original data, a combined system measurement model is constructed.

[0199] Based on the combined system state model and the combined system measurement model, Kalman filtering is used to determine the combined system state vector.

[0200] Based on the combined system state vector and the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC lever arm, the filtered updated three-dimensional position of the vehicle is determined.

[0201] In one exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a vehicle positioning method based on motion constraint adaptive filtering.

[0202] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements a vehicle localization method based on motion constraint adaptive filtering.

[0203] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements a vehicle localization method based on motion constraint adaptive filtering.

[0204] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 3 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a vehicle positioning method based on motion constraint adaptive filtering.

[0205] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0206] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0207] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0208] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0209] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0210] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0211] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A vehicle localization method based on motion-constrained adaptive filtering, characterized in that, The vehicle localization method based on motion-constrained adaptive filtering includes: Historical measurement data is acquired and preprocessed to obtain historical preprocessed data. This historical measurement data includes GNSS and INS measurement data from multiple historical epochs. The GNSS measurement data includes multi-frequency, multi-mode pseudorange and carrier phase measurements from the GNSS receiver. The INS measurement data includes three-dimensional velocity increments measured by the accelerometer and three-dimensional angle increments measured by the gyroscope in the INS. Preprocessing includes quality assessment and integrated navigation solution. The historical preprocessed data also includes GNSS / INS navigation information from multiple historical epochs, including the vehicle's three-dimensional position increment, three-dimensional position, pitch angle, roll angle, and heading angle. Based on historical preprocessed data, a virtual dead reckoning status model and a virtual dead reckoning measurement model are constructed. Using Kalman filtering, the virtual dead reckoning state vector is determined based on the virtual dead reckoning state model and the virtual dead reckoning measurement model. Based on the virtual dead reckoning state vector, the INS installation angle and NHC lever arm are determined; the INS installation angle is the Euler angle from the vehicle system to the load system, and the NHC lever arm is the distance vector from the center of the INS to the effective point of the NHC; Obtain the raw data for the current epoch; the raw data includes: GNSS measurement data and / or INS measurement data; Based on INS measurement data, a combined system state model and a mechanical arrangement model are constructed. Based on the INS mounting angle, NHC lever arm, and mechanical arrangement model, the three-dimensional position and three-dimensional velocity of the vehicle's compensated INS mounting angle and NHC lever arm are determined. When the original data does not contain GNSS measurement data, the filtered and updated three-dimensional position of the vehicle is determined according to the first positioning process to achieve vehicle positioning; wherein, the first positioning process includes: Based on the vehicle's compensated INS mounting angle and the three-dimensional and nominal velocities of the NHC lever arm, an NHC measurement model is constructed. Based on the NHC measurement model, the measurement noise covariance matrix is ​​reconstructed using the inverse gamma distribution, resulting in an NHC measurement model that includes the reconstructed measurement noise covariance matrix. Based on the combined system state model and the NHC measurement model containing the reconstructed measurement noise covariance matrix, variational Bayesian adaptive filtering is used to determine the combined system state vector. Based on the combined system state vector and the vehicle's compensated INS mounting angle and NHC lever arm 3D position, the vehicle's filtered updated 3D position is determined. When the raw data includes GNSS measurement data, the filtered and updated three-dimensional position of the vehicle is determined according to the second positioning process to achieve vehicle positioning; wherein the second positioning process includes: Based on the vehicle's compensated INS installation angle and the three-dimensional position of the NHC lever arm, and the GNSS measurement data in the original data, a combined system measurement model is constructed. Based on the combined system state model and the combined system measurement model, Kalman filtering is used to determine the combined system state vector. Based on the combined system state vector and the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC lever arm, the filtered updated three-dimensional position of the vehicle is determined.

2. The vehicle localization method based on motion-constrained adaptive filtering according to claim 1, characterized in that, The virtual dead reckoning state model includes: in, δP represents the k-th epoch. n δP n This represents the three-dimensional position error of the vehicle in the n-frame, where the n-frame is the navigation frame. δP represents the k-1 epoch. n ; Representing the k-1 epoch Let represent the direction cosine matrix from the b-system to the n-system, where the b-system is the load system; Representing the k-1 epoch The direction cosine matrix from the v-frame to the b-frame is represented by M; the auxiliary matrix is ​​represented by α. k-1 Let α represent the pitch and yaw installation angle errors of the INS at epoch k-1. φ represents the three-dimensional position increment of the vehicle in the n-frame at epoch k; k-1 φ represents the pitch misalignment error, roll misalignment error, and yaw misalignment error of the vehicle at epoch k-1; δk k-1 Let δk represent the k-1 epoch, where δk represents the virtual odometer calibration factor error.

3. The vehicle localization method based on motion-constrained adaptive filtering according to claim 2, characterized in that, The virtual dead reckoning measurement model includes: Among them, z DR,k This represents the measurement vector of the virtual dead reckoning model at epoch k; P represents the k-th epoch obtained through the positional recursion equation. n P n This represents the three-dimensional position of the vehicle in the n-frame. P represents the k-th epoch of the historical preprocessed data. n ; l represents the k-1 epoch. b , l b Indicates the NHC lever arm; × indicates the vector cross product operation; δl represents the k-th epoch. b ,δl b Indicates the NHC lever arm error; e DR,k H represents the measurement noise vector of the virtual dead reckoning measurement model at epoch k; DR,k This represents the coefficient matrix of the virtual dead reckoning measurement model at epoch k; x k Let x represent the k-th epoch, where x represents the state vector of the combined system.

4. The vehicle localization method based on motion-constrained adaptive filtering according to claim 3, characterized in that, The combined system state model includes: in, Represents the differential of φ; This represents the three-dimensional rotational angular velocity of the n-frame relative to the i-frame, where the i-frame is an inertial frame. express The calculation error; This indicates the three-dimensional measurement error of the gyroscope; Indicates δv n The differential; δv n f represents the three-dimensional velocity error of the vehicle in the n-frame; b This represents the three-dimensional velocity increment measured by the accelerometer; express The calculation error; This represents the three-dimensional angular velocity of Earth's rotation in the n-frame. express The calculation error; δf represents the angular velocity of a three-dimensional carrier in the n-system. b δg represents the three-dimensional measurement error of the accelerometer. n Indicates three-dimensional gravity error; Indicates δP n The differential.

5. The vehicle localization method based on motion-constrained adaptive filtering according to claim 4, characterized in that, The mechanical arrangement model includes: in, P represents the k-1 epoch obtained through the position recursion equation. n ; τ represents the three-dimensional velocity of the vehicle in the n-frame at epoch k; τ represents the sampling interval of the INS. This represents the three-dimensional velocity of the vehicle in the n-th frame at epoch k-1; Represents the specific force integral term for epoch k; This represents the harmful acceleration integral term at epoch k; Representing the k-epoch This represents the direction cosine matrix from epoch k-1 to epoch k in the n-system. Let represent the direction cosine matrix from epoch k to epoch k-1 in the b system.

6. The vehicle localization method based on motion-constrained adaptive filtering according to claim 5, characterized in that, The formulas for calculating the vehicle's compensated INS mounting angle and the three-dimensional velocity of the NHC lever arm under v-frame conditions include: in, This represents the three-dimensional velocity of the vehicle in the v-frame at epoch k, where the INS mounting angle and NHC lever arm are the compensated values. This represents the direction cosine matrix from the b-system to the v-system calculated based on the INS installation angle; This represents the three-dimensional angle increment measured by the gyroscope at epoch k; Representing the k-epoch 7. A vehicle positioning system based on motion-constrained adaptive filtering, used to implement the vehicle positioning method based on motion-constrained adaptive filtering as described in any one of claims 1-6, characterized in that, The vehicle positioning system based on motion-constrained adaptive filtering includes: The data acquisition module is used to acquire historical measurement data and preprocess it to obtain historical preprocessed data. The historical measurement data includes GNSS and INS measurement data from multiple historical epochs. The GNSS measurement data includes multi-frequency, multi-mode pseudorange and carrier phase measurements from the GNSS receiver. The INS measurement data includes three-dimensional velocity increments measured by the accelerometer and three-dimensional angle increments measured by the gyroscope in the INS. Preprocessing includes quality assessment and integrated navigation solution. The historical preprocessed data includes GNSS / INS navigation information from multiple historical epochs, including the vehicle's three-dimensional position increment, three-dimensional position, pitch angle, roll angle, and heading angle. The first model building module is used to build a virtual dead reckoning status model and a virtual dead reckoning measurement model based on historical preprocessed data. The virtual dead reckoning state vector determination module is used to determine the virtual dead reckoning state vector based on the virtual dead reckoning state model and the virtual dead reckoning measurement model using Kalman filtering. The INS installation angle and NHC lever arm determination module is used to determine the INS installation angle and NHC lever arm based on the state vector calculated from the virtual dead reckoning. The INS installation angle is the Euler angle from the vehicle system to the load system, and the NHC lever arm is the distance vector from the center of the INS to the effective point of the NHC. The raw data acquisition module is used to acquire the raw data of the current epoch; the raw data includes: GNSS measurement data and / or INS measurement data; The second model building module is used to build a combined system state model and a mechanical arrangement model based on INS measurement data; The speed compensation module is used to determine the three-dimensional position and three-dimensional speed of the vehicle's compensated INS mounting angle and NHC lever arm based on the INS mounting angle, NHC lever arm and mechanical arrangement model. The first positioning module is used to determine the filtered and updated three-dimensional position of the vehicle according to the first positioning process when the original data does not contain GNSS measurement data, thereby achieving vehicle positioning; wherein the first positioning process includes: Based on the vehicle's compensated INS mounting angle and the three-dimensional and nominal velocities of the NHC lever arm, an NHC measurement model is constructed. Based on the NHC measurement model, the measurement noise covariance matrix is ​​reconstructed using the inverse gamma distribution, resulting in an NHC measurement model that includes the reconstructed measurement noise covariance matrix. Based on the combined system state model and the NHC measurement model containing the reconstructed measurement noise covariance matrix, variational Bayesian adaptive filtering is used to determine the combined system state vector. Based on the combined system state vector and the vehicle's compensated INS mounting angle and NHC lever arm 3D position, the vehicle's filtered updated 3D position is determined. The second positioning module is used to determine the filtered and updated three-dimensional position of the vehicle according to the second positioning process when the original data includes GNSS measurement data, thereby achieving vehicle positioning; wherein the second positioning process includes: Based on the vehicle's compensated INS installation angle and the three-dimensional position of the NHC lever arm, and the GNSS measurement data in the original data, a combined system measurement model is constructed. Based on the combined system state model and the combined system measurement model, Kalman filtering is used to determine the combined system state vector. Based on the combined system state vector and the vehicle's compensated INS mounting angle and the three-dimensional position of the NHC lever arm, the filtered updated three-dimensional position of the vehicle is determined.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the vehicle positioning method based on motion constraint adaptive filtering as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the vehicle positioning method based on motion constraint adaptive filtering as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the vehicle positioning method based on motion constraint adaptive filtering as described in any one of claims 1-6.

Images