IMU installation angle fusion method and system combined with heading installation angle avoidance

Abstract

The application provides an IMU installation angle fusion method and system combined with heading installation angle reverse avoidance, and belongs to the technical field of vehicle-mounted integrated navigation, and comprises the following steps: establishing a vehicle-mounted IMU static gravity vector observation model, determining the vehicle static state by using a zero-speed detection algorithm, and solving and correcting the roll installation angle and the pitch installation angle through navigation system and carrier system gravity projection constraints; establishing an IMU dynamic inertial navigation recursive motion model, solving a spatial displacement vector based on a vehicle gentle straight driving condition, solving a heading installation angle through a horizontal plane vector angle deviation, and correcting a three-axis attitude full dimension; obtaining a single set of installation angle results according to sequence convergence determination and confidence checking, adaptively updating the three-axis installation angle in combination with the confidence of the installation angle and the integrated navigation new information; and after a new installation angle is calculated each time, the heading installation angle is subjected to reverse avoidance based on double inertial navigation recursion, and the correct installation angle is output.

Description

Technical Field

[0001] This invention relates to the field of vehicle-mounted integrated navigation technology, and in particular to an IMU mounting angle fusion method and system that combines heading mounting angle reverse avoidance. Background Technology

[0002] An Inertial Measurement Unit (IMU) is the core sensing component of an inertial navigation system (INS). It typically works in conjunction with a Global Navigation Satellite System (GNSS) to form a GNSS-INS integrated navigation system. This integrated navigation system, by fusing angular motion parameters collected by gyroscopes and linear motion data acquired by accelerometers, can calculate the vehicle's attitude, position, and various motion states in real time with high precision. It plays an irreplaceable and crucial role in core areas such as autonomous driving and precise positioning and navigation of special vehicles.

[0003] GNSS provides absolute position and velocity references, effectively suppressing the problem of IMU measurement errors accumulating over time. IMU, on the other hand, possesses autonomous navigation capabilities; even if GNSS signals fail due to obstruction or electromagnetic interference, the IMU can still output high-precision navigation data within a short time. The two complement each other efficiently. A key prerequisite for deep integration is ensuring the uniformity of the data spatial reference, with the IMU's installation angle relative to the vehicle's coordinate system being a core parameter determining the overall performance of the integrated navigation system.

[0004] Any deviation in the installation angle will directly cause projection distortion of the physical quantities output by the IMU during coordinate system transformation. For example, when the heading installation angle is deviated, the heading data measured by the IMU cannot match the actual driving direction of the vehicle, making it difficult for the GNSS position correction information and the IMU motion prediction results to accurately match. Errors in the roll and pitch installation angles will reduce the accelerometer's accuracy in distinguishing gravity components and motion acceleration, further causing attitude drift and the continuous accumulation of position estimation errors, ultimately significantly reducing the operational stability and navigation accuracy of the GNSS-INS integrated navigation system. Therefore, after the IMU is installed on the vehicle, its installation angle must be accurately calibrated to ensure a unified data spatial reference, providing solid support for stable and reliable navigation under complex driving conditions.

[0005] Existing implementation methods include: (1) External reference fusion method This method uses the standard GNSS onboard equipment as a base reference, and additionally deploys hardware such as wheel speed odometers and visual sensors. Through data fusion algorithms such as Kalman filtering and extended Kalman filtering, it establishes an error model between IMU measurements and external reference information (such as GNSS positioning results and odometer speed data), and then inversely calculates the roll, pitch, and yaw angles. Its core logic is to utilize the absolute reference characteristics of external sensors to correct the relative measurement deviation of the IMU, thereby achieving high-precision estimation of the installation angle.

[0006] (2) Displacement recursion method for IMU stationary-start To avoid relying on additional hardware, a method has emerged that relies solely on IMU data to solve for the three-axis mounting angles. First, using IMU data during the vehicle's stationary phase, the roll and pitch mounting angles are calculated based on the projected geometric relationship between the accelerometer output and the gravity vector. Then, leveraging the vehicle's linear acceleration from a standstill to start, the horizontal acceleration component output by the IMU is integrated twice to recursively calculate the vehicle's trajectory vector. Finally, the heading mounting angle is calculated using the angle between the trajectory vector's direction and the IMU's forward measurement axis. The core logic relies on the motion characteristics of a stationary start with an initial velocity of 0, using the uniqueness of the displacement trajectory direction to estimate the heading angle, ultimately achieving a complete solution for the three-axis mounting angles without external hardware assistance.

[0007] Both of the aforementioned existing technologies have significant limitations and are difficult to adapt to the actual needs of complex in-vehicle scenarios: (1) Deficiencies of external reference fusion method Although GNSS is a standard module in vehicle navigation, this method requires the additional deployment of hardware such as wheel speed odometers and visual sensors. This not only increases the hardware procurement cost of the vehicle, but also requires additional installation and debugging (such as matching the odometer with the vehicle wheel diameter and calibrating the visual sensor), increasing the complexity of system integration. At the same time, its performance is heavily dependent on the quality of GNSS signals. In scenarios where GNSS signals are blocked or fail, such as tunnels, underground parking garages, and densely populated areas with tall buildings, external reference information is interrupted, and the installation angle estimation will completely fail, making it impossible to meet the calibration requirements in complex environments.

[0008] (2) Defects of the IMU static-start displacement recursion method While this method requires no additional hardware and can completely solve for the three-axis mounting angles, it suffers from stringent application limitations and significant technical shortcomings: The method must strictly start the recursive calculation from a stationary vehicle state. The acceleration during the initial starting phase may be too small, and the weak acceleration signal is easily masked by sensor noise. To obtain an effective and identifiable displacement trajectory, a lengthy integration recursive calculation is required. This not only results in a cumbersome calculation process and a large computational load but also introduces more accumulated errors due to the increased integration time, leading to distortion of the displacement trajectory and a significant increase in the estimation error of the heading mounting angle. Furthermore, the calculated mounting angles are only for single-use and do not consider how to weight and fuse multiple sets of mounting angles, resulting in low system reliability.

[0009] In addition, both types of existing technologies share common technical drawbacks: existing vehicle-mounted IMUs calculate the heading angle based on the core assumption that "the direction of motion acceleration is consistent with the direction of vehicle movement". This assumption no longer holds true under hidden conditions such as sudden deceleration and braking, and reversing. This can easily lead to a systematic error of 180° reversal in the heading angle calculation. Moreover, the motion characteristics of such hidden conditions are very similar to those of conventional effective calculation conditions. Conventional verification methods cannot identify such anomalies, which will directly affect the accuracy of subsequent navigation calculations and cannot guarantee the stability of vehicle navigation.

[0010] In summary, existing technologies generally suffer from the following drawbacks: high hardware costs, poor adaptability to different scenarios, long inertial navigation recursion time, easy accumulation of errors, and failure to consider how to weight and fuse installation angles and avoid reverse calculation. Summary of the Invention

[0011] This invention provides an IMU installation angle fusion method and system that combines heading installation angle reverse avoidance to solve the shortcomings of existing technologies, such as high hardware cost, poor scenario adaptability, long inertial navigation recursion time, easy error accumulation, and failure to consider how to weightedly fuse and reverse the avoidance calculation of installation angles. The invention adopts a method of reverse avoidance for each newly calculated installation angle before adaptively fusing and updating the installation angle, and simultaneously considers the weighted fusion and reverse avoidance calculation of installation angles, resulting in higher system reliability.

[0012] In a first aspect, the present invention provides an IMU mounting angle fusion method that combines heading mounting angle reverse avoidance, comprising: A static gravity vector observation model for the vehicle-mounted IMU is established. The zero-speed detection algorithm is used to determine the stationary state of the vehicle. The roll and pitch installation angles are calculated and corrected by the gravity projection constraints of the navigation system and the vehicle system. A dynamic inertial navigation recursive motion model of IMU is established. The spatial displacement vector is calculated based on the smooth straight-line driving condition of the vehicle. The heading installation angle is calculated by the deviation of the horizontal plane vector angle, and the three-axis attitude is corrected in all dimensions. Based on the sequence convergence determination and confidence verification, a single set of installation angle results are obtained. Combining the confidence of the installation angle with the integrated navigation information, the three-axis installation angle is adaptively updated. After each new installation angle is calculated, the heading installation angle is reversed based on the dual inertial navigation recursion to output the correct installation angle.

[0013] According to the present invention, an IMU mounting angle fusion method combining heading mounting angle reverse avoidance is provided. A static gravity vector observation model of the vehicle-mounted IMU is established, a zero-speed detection algorithm is used to determine the vehicle's stationary state, and the roll and pitch mounting angles are calculated and corrected through gravity projection constraints between the navigation system and the vehicle system. The method includes: The angular velocity modulus of the carrier was calculated from the real-time angular velocity of the three axes using the IMU gyroscope, and the acceleration modulus of the carrier was calculated from the real-time specific force of the three axes using the IMU accelerometer. Determine the gyroscope zero-speed angular velocity threshold, acceleration amplitude deviation threshold, and local gravitational acceleration amplitude. If the vehicle's angular velocity magnitude is less than the gyroscope zero-speed angular velocity threshold, the absolute value of the difference between the vehicle's acceleration magnitude and the local gravitational acceleration amplitude is less than the acceleration amplitude deviation threshold, and the duration of the current state is greater than or equal to the system's minimum stationary duration, then the vehicle is determined to be in a completely stationary state. When the vehicle is stationary, the standard gravity vector in the navigation coordinate system only includes the local gravitational acceleration amplitude. Multiple frames of static observation data from the accelerometer are collected and the mean value is extracted to eliminate random noise. The attitude transformation matrix from the navigation coordinate system to the IMU carrier coordinate system is used to obtain the measured gravity observation vector in the IMU carrier coordinate system. Based on the roll and pitch installation angles to be calculated, the attitude transformation matrix is ​​expanded, the gravity vector transformation relationship is established, and the standard gravity vector in the navigation coordinate system is substituted to obtain the IMU's roll and pitch installation angles. Update the attitude transformation matrix to complete the correction of the roll and pitch installation angles.

[0014] According to the present invention, an IMU mounting angle fusion method combining heading mounting angle reverse avoidance is provided. This method establishes a dynamic inertial navigation recursive motion model for the IMU, calculates the spatial displacement vector based on the vehicle's smooth straight-line driving condition, calculates the heading mounting angle through the deviation of the horizontal plane vector angle, and corrects the three-axis attitude in all dimensions. The method includes: Determine the dynamic calibration operating conditions; Under dynamic calibration conditions, navigation quaternions are obtained, and attitude quaternions are obtained by recursively using quaternion multiplication operations in combination with the three-axis angular velocities of the carrier coordinate system output by the IMU gyroscope. The attitude matrix from the carrier coordinate system to the navigation coordinate system is obtained by solving the measured gravity observation vector and attitude quaternion in the carrier coordinate system. The measured gravity observation vector in the navigation coordinate system is obtained. After subtracting the standard gravity vector in the navigation coordinate system, multiple velocity integrations are performed to obtain the measured spatial displacement vector of the IMU in the navigation system during the calibration time period. Extract the north and east vectors of the measured spatial displacement vectors, obtain the forward horizontal reference vector from the north and east vectors, and calculate the heading installation deviation angle by solving the vector angle.

[0015] According to the present invention, an IMU mounting angle fusion method combining heading and mounting angle reverse avoidance is provided. When the vehicle acceleration exceeds a threshold, a model for recursively solving the velocity change is established, and the heading and mounting angle of the IMU is solved at each epoch using the angle between the velocity changes. The method includes: Establish a horizontal reference coordinate system to eliminate roll and pitch deviations. The horizontal reference coordinate system is parallel to the horizontal plane of the vehicle coordinate system and has rotational deviation only in the heading direction. A first-direction cosine matrix is ​​established to transform the IMU carrier coordinate system to the horizontal reference coordinate system. Based on the first-direction cosine matrix, the coordinate system is transformed on the IMU measured vector to obtain the measured acceleration in the IMU carrier coordinate system and the measured acceleration in the horizontal reference coordinate system. Determine the acceleration threshold. When the magnitude of the resultant acceleration in the horizontal reference coordinate system is greater than or equal to the acceleration threshold, initiate the integral recursion of the measured acceleration in the horizontal reference coordinate system. Establish a second direction cosine matrix from the vehicle coordinate system to the horizontal reference coordinate system. The second direction cosine matrix only contains the simplified direction cosine matrix of the heading installation angle. The velocity change vector of the vehicle coordinate system is formed by the actual forward velocity change of the vehicle, wherein the actual forward velocity change of the vehicle is greater than 0. The velocity change vector in the horizontal reference coordinate system is obtained by the second direction cosine matrix and the velocity change vector of the vehicle coordinate system. The second direction cosine matrix is ​​expanded and multiplied with the velocity change vector of the vehicle coordinate system to obtain the X-axis velocity change, Y-axis velocity change and Z-axis velocity change in the horizontal plane. Based on the changes in velocity along the X-axis and Y-axis, the sine and cosine expressions for the heading angle are obtained. Combining these with the trigonometric identities of the arctangent functions in the four quadrants, the optimal heading angle is solved.

[0016] According to the present invention, an IMU installation angle fusion method combining heading installation angle reverse avoidance is provided. Based on sequence convergence determination and confidence verification, a single set of installation angle results is obtained. Combining the confidence of the installation angle with integrated navigation information, the three-axis installation angle is adaptively updated, including: Determine the heading angle deviation threshold between adjacent epochs, extract the most recent several epochs including the current epoch, and construct a heading installation angle sequence. If the heading installation angle sequence satisfies that the difference between adjacent heading installation angles is less than the heading angle deviation threshold, and the difference between the maximum and minimum heading installation angles in the heading installation angle sequence is less than twice the heading angle deviation threshold, then the heading installation angle sequence is determined to be a converged heading installation angle sequence. Determine the horizontal attitude disturbance threshold and the heading and turning disturbance threshold, obtain the epoch interval of the converged heading installation angle sequence, and obtain quantization constraints based on the fact that the sum of the variances of the roll angle sequence and the pitch angle sequence within the epoch interval is less than or equal to the horizontal attitude disturbance threshold, and the sum of the products of the effective output value of the Z-axis of the gyroscope at any epoch in the horizontal reference coordinate system and the epoch sampling interval is less than or equal to the heading and turning disturbance threshold. Based on the horizontal attitude disturbance threshold, the heading and turning disturbance threshold, and the threshold positive weighting coefficient, construct the installation angle confidence, and use the confidence threshold to filter the installation angle confidence to obtain a converged heading installation angle sequence that has passed the confidence verification. The mean filtering is performed on the calculated heading and mounting angle values ​​of multiple epochs in the converged and confidence-verified heading and mounting angle sequence. The mean value corresponding to the second half of the epochs in the sequence is used as the optimal estimated value of the heading and mounting angle. This value is then integrated with the optimal heading and mounting angle to obtain the three-axis mounting angle vector of the IMU relative to the vehicle coordinate system. Based on the three-axis mounting angle vector, multiple sets of mounting angle calculation results are collected and filtered. The confidence level of each set of mounting angle calculation results is used as the weight to perform weighted fusion of multiple sets of fusionable mounting angles. Outlier detection and adaptive updating are performed on the new calculated mounting angle set to remove abnormal working condition deviations and obtain the final optimal estimate of the IMU three-axis mounting angle.

[0017] According to the present invention, an IMU mounting angle fusion method combining heading mounting angle reverse avoidance is provided. Based on the three-axis mounting angle vector, multiple sets of mounting angle calculation results are collected and filtered. Using the confidence level of each set of mounting angle calculation results as weight, multiple sets of fusionable mounting angles are weighted and fused. Outlier detection and adaptive updating are performed on the newly calculated mounting angle set to remove abnormal operating condition deviations, and the final optimal estimate of the IMU three-axis mounting angle is obtained, including: Calculate the GNSS-INS integrated navigation position information modulus of any set of installation angle calculation results. If it is determined that the GNSS-INS integrated navigation position information modulus is less than the position information threshold, it is determined that there is no abnormality in the settlement condition. If it is verified that the modulus of the difference between multiple consecutive sets of installation angle calculation results is less than the installation angle deviation threshold, then the multiple consecutive sets of installation angle calculation results are fusionable results. Multiple sets of fusionable results are collected cumulatively. Based on the total number of fusionable results, the installation angle obtained by weighted fusion of multiple fusionable results, and the cumulative confidence weight, an initial fusion result is obtained. Each time a new set of effective installation angle calculation results that has been screened by working conditions is added, the initial fusion result is dynamically iterated and fused to obtain an iterative fusion result. If the magnitude of the difference between the newly calculated installation angle group and the installation angle obtained by weighted fusion of multiple fusion results does not meet the requirement of being less than the installation angle deviation threshold, the newly calculated installation angle group is determined to be an outlier. If two consecutive newly calculated installation angle groups are both outliers, the installation angle change is determined to have temporal anomaly consistency. If the magnitude of the newly calculated installation angle group is greater than the position information threshold, the GNSS-INS integrated navigation position information is determined to be out of normal range. When both the temporal anomaly consistency and the GNSS-INS integrated navigation position information being out of normal range conditions are met, the iterative fusion result is updated to the latest valid installation angle calculation result. When only one of the temporal anomaly consistency and the GNSS-INS integrated navigation position information being out of normal range conditions is met, it is determined to be a normal anomaly, the outlier is removed and the current fusion result remains unchanged. A fusion weight stability threshold is determined. If the cumulative confidence weight is greater than or equal to the fusion weight stability threshold, the installation angle is determined to be stable, dynamic updates are stopped, and the current fusion value is used. Otherwise, the effective new installation angle group is included in the fusion, the weighted fusion value is dynamically updated, and the final optimal estimate of the IMU three-axis installation angle is obtained.

[0018] According to the present invention, an IMU mounting angle fusion method combining heading and mounting angle reverse avoidance is provided. After each calculation of a new mounting angle, the heading and mounting angle is reversed based on dual inertial navigation recursion to output the correct mounting angle, including: The system acquires the northward velocity component, eastward velocity component, and groundward velocity component of the GNSS output. The northward velocity component and the eastward velocity component constitute the GNSS horizontal velocity. The system determines the GNSS horizontal velocity threshold and the GNSS horizontal velocity standard deviation threshold. If the GNSS horizontal velocity magnitude is greater than the GNSS horizontal velocity threshold and the ratio of the GNSS horizontal velocity standard deviation to the GNSS horizontal velocity magnitude is less than the GNSS horizontal velocity standard deviation threshold, the system enters the recursive process. The initial heading angle is calculated using the GNSS horizontal velocity, the initial pitch angle is calculated using the ground velocity component and the GNSS horizontal velocity magnitude, the initial roll angle is set to zero, the GNSS observation value is used as the GNSS initial velocity, and a recursive initial state is constructed. If it is determined that the above conditions are not met, the verification process is retried at the next moment. Based on the recursive initial state, the estimated value of the current calculated heading installation angle is used as the original installation angle. The original installation angle is reversed by 180 degrees to obtain the reverse installation angle. The original recursive position and the reverse recursive position corresponding to the original installation angle and the reverse installation angle are determined respectively. After reaching the preset shortest recursive time, the GNSS quantization and discrimination process is entered. Based on the actual working performance of GNSS, if the standard deviation modulus of GNSS horizontal position is determined to be less than or equal to the threshold value of the standard deviation modulus of GNSS horizontal position, then the GNSS position reliability judgment process is carried out; otherwise, the recursion continues. The difference magnitudes between the original recursive position and the reverse recursive position and the GNSS position at the same time are obtained as the original position deviation and the reverse position deviation, respectively. The position deviation ratio is constructed by the ratio of the reverse position deviation to the original position deviation, and the discrimination threshold and the abnormal deviation threshold are determined. If the original position deviation is greater than the abnormal deviation threshold, and the reverse position deviation is greater than the abnormal deviation threshold, then clear the current recursive result and wait for the next initial state verification to be satisfied before re-executing the process; otherwise, proceed to the subsequent judgment process. If the original position deviation is greater than the discrimination threshold, the original installation angle is determined to be correct, and the current calculated heading installation angle estimate is used for subsequent navigation calculations. If the original position deviation is less than the reciprocal of the discrimination threshold, it is determined that the original installation angle has a 180-degree reversal. The currently calculated estimated heading installation angle is then corrected to the reverse installation angle and input into the integrated navigation system. If none of the above conditions are met, the recursion continues, and the GNSS location reliability judgment process and subsequent discrimination process are re-executed.

[0019] Secondly, the present invention also provides an IMU mounting angle fusion system that combines heading mounting angle reverse avoidance, comprising: The first module is used to establish the vehicle's gyroscope zero-bias estimation model and the accelerometer-gravity projection correlation model in a stationary state, and to solve for the roll and pitch installation angles. The second module is used to establish a model for recursively solving the velocity change when the vehicle acceleration exceeds the threshold. The model is then used to solve the heading installation angle of the IMU at each epoch using the angle between the velocity changes. The first calculation module is used to obtain a single set of installation angle results based on the sequence convergence judgment and confidence verification, and adaptively update the three-axis installation angle by combining the confidence of the installation angle with the integrated navigation information. The second calculation module is used to reverse the heading installation angle based on the dual inertial navigation recursion after each new installation angle is calculated, and output the correct installation angle.

[0020] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the IMU mounting angle fusion method for reverse avoidance combined with heading mounting angle as described above.

[0021] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the IMU mounting angle fusion method combined with heading mounting angle reverse avoidance as described above.

[0022] The IMU mounting angle fusion method and system provided by this invention, which combines heading mounting angle reverse avoidance, achieves adaptive fusion update of the three mounting angles by calculating the roll and pitch mounting angles when stationary and calculating the heading mounting angle through velocity increment and attitude constraints. It is suitable for vehicle inertial navigation calibration scenarios, does not require external references such as odometers, and significantly simplifies the implementation conditions of mounting angle calibration. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating the IMU mounting angle fusion method combining heading mounting angle reverse avoidance provided by the present invention. Figure 2 This is a schematic diagram of the installation angle of the GNSS / INS integrated navigation vehicle IMU provided by the present invention; Figure 3 This is a schematic diagram of the confidence-weighted adaptive update process for the installation angle of an on-board IMU provided by the present invention; Figure 4 This is a schematic diagram of the reverse avoidance process based on dual inertial navigation recursion for heading installation angle provided by the present invention; Figure 5 This is a schematic diagram of the IMU mounting angle fusion system that combines heading mounting angle reverse avoidance provided by the present invention; Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0026] Figure 1 This is a flowchart illustrating the IMU installation angle fusion method combining heading installation angle reverse avoidance provided in an embodiment of the present invention, as shown below. Figure 1 As shown, it includes: Step 100: Establish a static gravity vector observation model for the vehicle-mounted IMU, use a zero-speed detection algorithm to determine the vehicle's stationary state, and calculate and correct the roll and pitch installation angles through gravity projection constraints between the navigation system and the vehicle system. Step 200: Establish an IMU dynamic inertial navigation recursive motion model, calculate the spatial displacement vector based on the vehicle's smooth straight-line driving condition, calculate the heading installation angle through the deviation of the horizontal plane vector angle, and correct the three-axis attitude in all dimensions. Step 300: Based on the sequence convergence determination and confidence verification, obtain the single set of installation angle results, and adaptively update the three-axis installation angle by combining the confidence of the installation angle with the integrated navigation information; Step 400: After each new installation angle is calculated, the heading installation angle is reversed based on the dual inertial navigation recursion to output the correct installation angle.

[0027] Specifically, this embodiment of the invention first constructs a MEMS error identification model based on the vehicle's static condition, establishes the mapping relationship between the accelerometer sensor output and the local gravity field in the IMU carrier coordinate system, and solves the carrier tilt installation deviation angle by simultaneously solving the attitude constraint equations; then, it sets a vehicle motion acceleration discrimination threshold, constructs an inertial integral incremental solution model under the carrier's dynamic motion, and uses the geometric angle characteristics of the horizontal plane velocity increment vector to iteratively solve the heading installation deviation angle epoch by epoch; by performing convergence discrimination and credibility quantification screening on the attitude solution sequence, it outputs the single effective installation angle estimate; based on the solution credibility index and integrated navigation information constraints, it realizes online adaptive iterative correction of the IMU three-axis installation deviation angle; finally, after each round of IMU three-axis installation angle solution is completed, a dual-path inertial solution parallel deduction mechanism is adopted to perform direction discrimination and reverse error correction on the heading deviation angle, ensuring that the installation angle data sent into the fusion filtering link is true and reliable.

[0028] Based on the above embodiments, step 100 includes: Static calibration: IMU roll and pitch installation angle calculation and correction The core of this step is to utilize the static observation characteristics of gravity vectors when the vehicle is stationary, and combine the vector projection relationship between the navigation system and the IMU carrier coordinate system to accurately calculate the roll and pitch installation angles in the vertical plane, thereby eliminating the installation attitude deviations in the pitch and roll directions of the IMU. The entire calibration process requires no vehicle movement and relies on purely static data.

[0029] (1) Vehicle stationary state determination A zero-speed detection algorithm is employed to identify the vehicle's motion state, avoiding interference from minor vibrations and low-speed slippage on static data. By collecting real-time output data from IMU accelerometers and gyroscopes, a quantified zero-speed determination criterion is constructed. The vehicle is determined to be stationary by combining dual threshold constraints of angular velocity magnitude and acceleration magnitude with the duration of motion. The relevant core formulas and determination logic for zero-speed detection are as follows.

[0030] Assume the real-time angular velocity output of the IMU gyroscope on three axes is , , Formula for calculating the angular velocity modulus of the carrier:

[0031] Assume the real-time force output of the IMU accelerometer on three axes is , , Formula for calculating the acceleration modulus of a carrier:

[0032] Define the zero-speed threshold: The threshold value for the zero-velocity angular velocity of the gyroscope. For acceleration amplitude deviation threshold, This represents the local gravitational acceleration amplitude. The conditions for determining if a vehicle is stationary are: the angular velocity magnitude is less than a threshold, the acceleration magnitude is close to the gravitational acceleration, and the duration of the state meets the set requirements. The complete criteria are as follows:

[0033] In the formula: The duration of the current state. Set a minimum stationary duration for the system. When all the above conditions are met consecutively, the vehicle is determined to be completely stationary, and static data acquisition can be initiated, effectively filtering out false judgments caused by short-term shaking or slight bumps.

[0034] (2) Static gravity vector observation When the vehicle is stationary, there is no motion acceleration in the IMU carrier coordinate system (b frame), and the accelerometer observations only contain the gravitational acceleration component. Assume the standard gravity vector is a fixed vector in the navigation coordinate system (n frame, N-E).

[0035] in This represents the local gravitational acceleration amplitude. After the vehicle comes to a stable stop, multiple frames of static accelerometer observation data are collected, and the average value is taken to eliminate random noise, yielding the measured gravity observation vector in the vehicle coordinate system. When at rest, the coordinate transformation relationship is satisfied:

[0036] In the formula: This is the attitude transformation matrix from the navigation system to the IMU carrier system, including the roll installation angle to be calculated. Pitch installation angle .

[0037] (3) There is no approximate solution for roll and pitch installation angles. Based on the complete expansion of the attitude transformation matrix (without small-angle approximations) and combined with gravity vector projection constraints, the precise calculation formulas for roll and pitch angles are derived. The core expansions related to roll and pitch angles in the attitude transformation matrix are as follows:

[0038] Establish the simultaneous gravity vector transformation relationship and substitute it into... It can be deduced that:

[0039] By eliminating trigonometric coupling terms through fractional operations, the formula for calculating the installation angle without approximation is obtained:

[0040]

[0041] After obtaining the accurate IMU roll and pitch installation angles by solving the above formulas, the attitude transformation matrix is ​​updated to complete the correction of the IMU's installation attitude deviation in the vertical plane, thus eliminating pitch and roll attitude coupling interference for subsequent heading angle calibration.

[0042] Based on the above embodiments, step 200 includes: Dynamic calibration: IMU heading and mounting angle calculation and all-attitude correction This step is based on the dynamic condition of the vehicle traveling smoothly in a straight line. The measured spatial displacement vector of the IMU is solved by inertial navigation attitude recursion and velocity integration. The front of the vehicle body is taken as the ideal forward reference. The heading installation angle is calculated by the angle between the horizontal plane vectors. This makes up for the horizontal heading installation deviation that cannot be identified by static calibration. Finally, the three-axis mounting angle is fully calibrated.

[0043] like Figure 2In the schematic diagram of the GNSS / INS integrated navigation vehicle-mounted IMU installation angle shown, the vehicle coordinate system is called the v system, and the carrier (IMU) coordinate system is called the b system. The XYZ axes of the v system point to the lower right front of the vehicle body, while the XYZ axes of the b system do not coincide with the lower right front of the vehicle body, but form an angle. Therefore, it is necessary to construct a rotation matrix from the vehicle system to the IMU system.

[0044] (1) Dynamic calibration of operating conditions control The vehicle is controlled to maintain a low speed and smooth straight-line driving state without steering or significant acceleration or deceleration throughout the process. This avoids the impact of vehicle sideslip and violent attitude fluctuations on the displacement vector, ensuring that the actual forward direction of the vehicle is consistent with the geometric front of the vehicle body, and providing accurate benchmark conditions for the calculation of the heading angle.

[0045] (2) Inertial navigation attitude and displacement vector calculation During dynamic calibration, high-precision inertial navigation attitude and displacement recursive calculations are required based on the original IMU observation data to provide a true spatial motion vector reference for the heading installation angle calculation. This paper adopts a quaternion attitude update algorithm to complete the real-time recursive calculation of the vehicle attitude. This algorithm has no singularities, is applicable to all angles, and matches the calibration principle of no small angle approximations throughout the process. The vehicle attitude is iteratively updated by the real-time angular velocity observations of the gyroscope, which represents the spatial attitude change of the IMU relative to the NE-G coordinate system in real time. Then, combined with the specific force observations of the accelerometer, after coordinate transformation, gravity compensation, and frame-by-frame integration, the navigation system velocity update and position displacement update are completed sequentially, accurately calculating the actual motion trajectory vector of the vehicle and avoiding the coupling effect of attitude calculation errors on heading calibration. Let the navigation quaternion be... The attitude quaternion recursive update formula is:

[0046] In the formula: The angular velocities of the carrier coordinate system output by the IMU gyroscope. This refers to quaternion multiplication operations.

[0047] By combining the accelerometer force observations, the navigation system force conversion and velocity integral update are completed. (Carrier system force) The formula for converting to the navigation system is:

[0048] In the formula: The attitude matrix from the carrier frame to the navigation frame is solved using quaternions. Velocity integration is performed after subtracting the effects of gravity.

[0049]

[0050] Integrating the velocity again yields the measured spatial displacement vector of the IMU in the navigation system during the calibration time period. .

[0051] (3) Calculation of heading installation angle and full attitude correction Following the inertial navigation recursive calculation described above, precise vehicle displacement information in the navigation coordinate system can be obtained. Under ideal installation conditions, the IMU's measured forward direction should perfectly coincide with the vehicle's geometric forward direction. If there is a heading installation deviation, the horizontal displacement vector calculated by the IMU will form a fixed angle with the vehicle's ideal forward direction. This section, based on the vehicle's uniform straight-line motion characteristics without lateral slippage, ignores vehicle maneuvering attitude interference and utilizes the azimuth deviation of the horizontal displacement vector in the navigation system to accurately calculate the IMU's heading installation deviation angle. Small-angle approximations are not used throughout to ensure calibration accuracy in scenarios with large installation deviation angles. When the vehicle is traveling in a straight line, the ideal forward vector of the vehicle is the north direction vector of the navigation system. The ideal forward horizontal reference vector of the vehicle is defined as... Extract the horizontal plane component of the IMU measured displacement vector. The heading installation deviation angle is calculated using the vector angle formula. The formula has no angular approximation, resulting in higher accuracy:

[0052] that angle This refers to the yaw installation deviation angle in the horizontal plane of the IMU. The roll and pitch angles obtained from static calibration and the yaw angle obtained from dynamic calibration are uniformly substituted into the attitude transformation matrix to complete the overall correction of the IMU's roll, pitch, and yaw three-axis installation angles, completely eliminating the IMU's installation attitude deviation and completing the entire calibration process.

[0053] Based on the above embodiments, step 300 includes: (1) Convergence determination of heading installation angle sequence To filter out measurement noise and random disturbances and improve the accuracy and stability of the heading installation angle calculation, a convergence test is performed on the continuous heading installation angle sequence calculated epoch by epoch. The current epoch is then used. Including the recent Each epoch, constructing the heading installation angle sequence. Set the heading angle deviation threshold between adjacent epochs. A sequence is considered convergent if it simultaneously satisfies the following constraints:

[0054] (2) Installation angle confidence test After the heading and installation angle sequence converges, the confidence level of the installation angles is verified to quantitatively evaluate the degree of horizontal attitude disturbance and heading change, eliminating low-confidence operating conditions. A horizontal attitude disturbance threshold is set. Heading and turning disturbance threshold Let the epoch interval of the convergent sequence be... The quantification constraints are:

[0055] In the formula, For the epoch interval corresponding to the convergence sequence of the heading installation angle; for The sum of variances of the roll and pitch angle sequences within the interval quantifies the overall perturbation of the horizontal attitude (the smaller the variance, the more stable the attitude). for The total heading angle within the interval reflects whether the vehicle is maintaining straight-line travel; The variance calculation operator is expressed as follows:

[0056] ( For the interval epoch number, For sequence (mean) , Horizontal reference coordinate system Next The roll and pitch angles of the epoch. For the first in this coordinate system Epoch Gyroscope Effective output value of axis The epoch sampling interval.

[0057] Construct an installation angle confidence model to assess the reliability of the solution:

[0058] In the formula, For the installation angle confidence, , It is a positive weighting coefficient, which can be adjusted according to the degree of influence of the operating condition disturbance; The larger the value, the higher the reliability of the solution. Set a confidence threshold. , The solution is deemed invalid and the result is discarded. It is determined to be valid and retained.

[0059] (3) Calculation of final value of heading installation angle and integration of three-axis installation angle After the heading installation angle sequence converges and the confidence check passes, The calculated heading and installation angle values ​​for each epoch are filtered using a sliding window mean filter, and then... The mean of each epoch is used as the optimal estimate of the heading angle:

[0060] Roll installation angle calculated using the gravity projection method Pitch installation angle With respect to the above-mentioned optimal heading installation angle By integrating the data, we obtain the three-axis mounting angle vectors of the IMU relative to the vehicle coordinate system:

[0061] (4) Adaptive update of integrated navigation information judgment 1) Collection and filtering of multiple sets of installation angle calculation results Definition of the first The effective solution result is: ,right First, a validity screening of the operating conditions is performed to ensure that the calculation results correspond to normal driving conditions. The specific judgment criteria are as follows: 1. Preliminary Position Information Assessment: Calculation of Position Information Module Length for GNSS-INS Integrated Navigation ,like ( If the location information threshold is used, then the solution condition is determined to be without obvious abnormalities. 2. Stability determination of continuous solutions: To verify the consistency of three consecutive valid solutions, the following conditions must be met simultaneously. and ( (This is the installation angle deviation threshold) to ensure that the solution results are free from sudden fluctuations.

[0062] When both of the above conditions are met, the determination is made. , , This is a fusion-compatible result. Definition For the first If there are three fusionable results, then these three fusionable results are denoted as... , , Only such fusion-compatible results are retained for subsequent weighted fusion.

[0063] 2) Confidence-based weighted fusion calculation Using the confidence level of each set of mounting angles as weight, a two-step weighted fusion of multiple sets of fusionable mounting angles is performed to improve the accuracy and robustness of the final mounting angle: 1. Initial Fusion: After accumulating 3 sets of fusionable results, initial fusion is triggered to build a stable baseline. The fusion formula is:

[0064] In the formula, This represents the total number of fusionable results; for The installation angle is obtained by weighted fusion of the fusion results; This is used as the cumulative confidence weight; this step effectively avoids random disturbances in single-batch calculations by redundantly fusing multiple sets of data.

[0065] 2. Dynamic Iterative Fusion: After the initial fusion is completed, for each newly added set of valid solution results selected by the working conditions... , decide first If this condition is met, lightweight iterative fusion will be initiated, eliminating the need to store historical data. The iterative formula is as follows:

[0066] In the formula, For the first The mounting angle after secondary fusion; For the front The cumulative confidence weight of the sub-fusion The updated cumulative weights are used; the adaptive fusion update of the installation angle is achieved through binary iteration, taking into account the reliability of new data and historical fusion results.

[0067] 3) Outlier detection and adaptive update Outlier detection and adaptive updates are performed on the newly calculated installation angle group to eliminate abnormal operating condition deviations: 1. Outlier detection: If the newly added valid solution does not meet the requirements... If it is, then it is judged as an outlier. They will not be included in the fusion process for the time being, in order to avoid affecting the overall estimation accuracy due to a single abnormal operating condition.

[0068] 2. IMU Movement Detection: A physical movement of the IMU is determined to have occurred if and only if both of the following conditions are met simultaneously: (1) Temporal anomaly consistency: Two consecutive sets of solution results are outliers, and satisfy the following conditions: This indicates that the installation angle variation is consistent, thus ruling out random anomalies; (2) Location-based supporting evidence: If This indicates that the GNSS-INS integrated navigation position information is outside the normal range, which can further verify that the anomaly was caused by IMU movement.

[0069] When both conditions are met, the fusion benchmark is updated to the latest valid solution; if only one condition is met, it is judged as a normal anomaly, the abnormal data is removed and the current fusion result remains unchanged.

[0070] 3. Adaptive and stable update: Set a stable threshold for fusion weights. If the cumulative confidence weights satisfy If the installation angle is determined to be stable, dynamic updates are stopped and the current fusion value is used; if the condition is not met, a valid new solution group is included in the fusion, and the weighted fusion value is dynamically updated.

[0071] When the fusion weight Reaching the preset threshold When the fusion result has fully utilized the redundancy of multiple batches of data and the installation angle estimation tends to stabilize, the iterative update is stopped, and the current weighted fusion value is taken as the final optimal estimate of the IMU's three-axis installation angle. .

[0072] Thus, the confidence-weighted adaptive fusion update of the IMU mounting angle has been completed. This method mines the redundancy value of multiple batches of data through staged weighted fusion, and improves robustness by combining outlier detection and IMU movement determination. It achieves high-precision and long-term stable estimation of the mounting angle under complex vehicle conditions, providing reliable support for the engineering application of vehicle IMU mounting angle.

[0073] like Figure 3 The diagram illustrating the confidence-weighted adaptive update process for the vehicle-mounted IMU installation angle shows the overall process as follows: 1. Initialization and First Weighted Startup After system startup, board initialization is performed first to complete basic parameter and status configuration. Then, the single-set installation angle (a) is calculated, and it is determined whether the system has entered integrated navigation and the new information is small. If the system has not entered integrated navigation or the new information is large, the latest single-set installation angle is used directly. If the system has entered integrated navigation and the new information is small, the system further checks whether three consecutive sets of installation angles are close. If they are close, the weighted result of these three sets is used as the initial installation angle. If they are not close, a new single-set installation angle (b) is calculated, and the subsequent decision-making process begins.

[0074] 2. Dynamic Update and Fusion Weight Determination After calculating the single-group installation angle (b), the system enters the dynamic update phase. First, it determines whether the new group installation angle is close to the current weighted installation angle: if not, the new group installation angle is marked as an outlier, and the current weighted installation angle is used directly without updating; if close, it further determines whether the fusion weight has reached a preset threshold. If the fusion weight reaches the threshold, a fusion update is performed, incorporating the new group installation angle into the weighted calculation and updating the fusion weight; if the threshold is not reached, the current weighted installation angle continues to be used, ensuring a smooth transition of installation angles.

[0075] 3. Outlier and IMU Movement Detection When a new set of installation angles is marked as an outlier, the system enters the anomaly protection branch. First, it checks if the condition of "two consecutive sets of outliers with similar values" is met. If so, it further checks if the integrated navigation information is significant. If the information is significant, it indicates that the IMU has been physically moved, and the original installation angle needs to be replaced with the latest single-set installation angle. If neither condition is met, it returns to recalculating the single-set installation angle to prevent outliers from affecting system stability.

[0076] 4. Reset and Closed-Loop Optimization The entire process supports a board information reset mechanism. If the reset condition is triggered, the system returns to the initialization phase and re-executes the initial installation angle calculation and weighting process. If the reset is not triggered, the dynamic update and anomaly protection logic continues to be executed, forming a complete closed-loop control. Through multiple judgments and adaptive strategies, the estimation accuracy of the installation angle and the robustness of the system under complex vehicle operating conditions are effectively guaranteed.

[0077] Based on the above embodiments, step 400 includes: After each round of IMU three-axis installation angle calculation is completed, a dual-path inertial calculation parallel extrapolation mechanism is adopted to perform direction determination and reverse error correction on the heading angle, ensuring that the installation angle data sent into the fusion filtering link is true and reliable.

[0078] The current vehicle-mounted IMU's heading angle calculation is based on the core assumption that "the direction of motion acceleration is consistent with the direction of vehicle movement". This assumption no longer holds true under hidden conditions such as sudden deceleration and braking, and reversing, which can easily lead to a systematic error of 180° reversal in the heading angle calculation. Moreover, the motion characteristics of such conditions are similar to those of conventional effective calculation conditions, and the accuracy verification cannot identify such anomalies. Therefore, a reverse avoidance method needs to be added to correct this directional deviation.

[0079] like Figure 4 The schematic diagram of the heading installation angle reverse avoidance process based on dual inertial navigation recursion shown in the figure constitutes a three-level closed-loop process through initial state validity verification, dual inertial navigation recursion, and GNSS quantization discrimination. This process can accurately identify and correct heading installation angle reverse errors, avoiding the introduction of abnormal installation angles into navigation calculations. The specific steps are as follows.

[0080] (1) Initial state validity verification make , , Let be the north, east, and ground velocity components output by GNSS, respectively. Then the GNSS horizontal velocity can be expressed as: The initial calculation yielded an estimated value for the heading installation angle. Then, the system performs a two-condition check on the initial state. Only when both conditions are met simultaneously can the subsequent recursive process proceed. 1. Velocity Amplitude Condition: The horizontal velocity magnitude of the GNSS must satisfy... ,in To ensure that the vehicle is in an effective state of motion; 2. Normalized velocity accuracy condition: The ratio of the standard deviation of GNSS horizontal velocity to the velocity modulus satisfies... .

[0081] After the verification is passed, the initial heading angle is calculated using GNSS velocity components. Initial pitch angle The initial roll angle is set to zero, and the initial velocity is determined using GNSS observations, thus completing the recursive initial state construction. If the conditions are not met, the system waits for the next moment to re-trigger the verification.

[0082] (2) Parallel recursion of dual heading installation angle inertial navigation Based on the established initial state, pure inertial navigation recursion is performed simultaneously using two sets of heading installation angles: 1. Original installation angle : Take the estimated value of the heading installation angle calculated at the current solution. ; 2 Reverse installation angle Reverse the original installation angle .

[0083] , The corresponding recursive positions are respectively , After reaching the minimum recursion time of 3 seconds, the GNSS quantization and discrimination process begins.

[0084] (3) Output of GNSS quantization discrimination and avoidance results Based on the actual operating performance of GNSS, set the criteria for determining the reliability of GNSS positioning. If the condition is not met, continue the recursion; if the condition is met, proceed to the following judgment step: definition , GNSS position at the same time The difference modulus are respectively , Construct the position deviation ratio:

[0085] The discrimination threshold is taken as Set abnormal deviation threshold. The discrimination and output rules are as follows: 1. First, determine and If the condition is met, it indicates that the recursive errors corresponding to the two installation angles are both large, and the current judgment is deemed invalid. The current recursive result is cleared, and the process is re-executed after the next initial state verification is satisfied. If the condition is not met, the process proceeds to the next judgment. 2 If ,Right now Much larger The original installation angle was determined to be correct, and it was directly adopted. Perform subsequent navigation calculations; 3 If ,Right now much smaller The original installation angle was determined to exist. In the opposite direction, Revised to Then input the integrated navigation system; 4. If none of the above conditions are met, the process continues recursively and returns to the beginning of this section to re-execute the GNSS position reliability judgment and subsequent discrimination process.

[0086] (4) Adaptive fusion update of installation angle combined with reverse avoidance of heading and installation angle For each independently calculated onboard IMU's heading, roll, and pitch three-axis mounting angle, the aforementioned heading mounting angle reverse avoidance module is first invoked to accurately identify and correct the 180° systematic reverse error that is prone to occur under hidden conditions such as rapid deceleration and reversing. Abnormal mounting angle data is strictly eliminated, and reliable mounting angle results that conform to actual operating conditions are selected. Then, the compliant mounting angle data after verification and correction is sent to the adaptive fusion update system to carry out multi-batch data weighted fusion and iterative optimization. Through this process design of "avoiding and correcting errors first, then fusion and updating," abnormal reverse mounting angles are prevented from flowing into the fusion stage from the source, ensuring that the initial data input to the adaptive fusion update system is accurate and reliable, further improving the stability and accuracy of the overall mounting angle estimation, and adapting to the long-term navigation needs of various complex onboard driving scenarios.

[0087] The IMU mounting angle fusion system combining reverse avoidance of heading and mounting angle provided by the present invention will be described below. The IMU mounting angle fusion system combining reverse avoidance of heading and mounting angle described below can be referred to in correspondence with the IMU mounting angle fusion method combining reverse avoidance of heading and mounting angle described above.

[0088] Figure 5 This is a schematic diagram of the IMU mounting angle fusion system that combines heading and mounting angle reverse avoidance provided in an embodiment of the present invention, as shown below. Figure 5 As shown, it includes: a first establishment module 51, a second establishment module 52, a first calculation module 53, and a second calculation module 54, wherein: The first setup module 51 is used to establish the vehicle's gyroscope zero-bias estimation model and the accelerometer and gravity projection correlation model in a stationary state, respectively, to solve the roll and pitch installation angles; the second setup module 52 is used to establish a model for recursively solving the velocity change when the vehicle's acceleration exceeds a threshold, and to solve the IMU's heading installation angle at each epoch using the angle between the velocity changes; the first calculation module 53 is used to obtain a single set of installation angle results based on the sequence convergence judgment and confidence verification, and to adaptively update the three-axis installation angles by combining the confidence of the installation angles with the integrated navigation information; the second calculation module 54 is used to reverse the heading installation angle based on the dual inertial navigation recursion after each calculation of the new installation angle, and output the correct installation angle.

[0089] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call logic instructions in the memory 630 to execute an IMU mounting angle fusion method that combines heading mounting angle reverse avoidance. This method includes: establishing a static gravity vector observation model of the vehicle-mounted IMU, using a zero-speed detection algorithm to determine the vehicle's stationary state, and calculating and correcting the roll mounting angle and pitch mounting angle through gravity projection constraints between the navigation system and the vehicle system; establishing a dynamic inertial navigation recursive motion model of the IMU, calculating the spatial displacement vector based on the vehicle's smooth straight-line driving condition, calculating the heading mounting angle through the deviation of the horizontal plane vector angle, and correcting the three-axis attitude in all dimensions; obtaining a single set of mounting angle results based on sequence convergence judgment and confidence verification, and adaptively updating the three-axis mounting angle by combining the confidence of the mounting angle with the integrated navigation information; after each calculation of a new mounting angle, performing reverse avoidance of the heading mounting angle based on dual inertial navigation recursion, and outputting the correct mounting angle.

[0090] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements an IMU mounting angle fusion method for reverse avoidance of heading mounting angle provided by the above methods. The method includes: establishing a static gravity vector observation model of the vehicle-mounted IMU, using a zero-speed detection algorithm to determine the stationary state of the vehicle, and calculating and correcting the roll mounting angle and pitch mounting angle through gravity projection constraints of the navigation system and the vehicle system; establishing a dynamic inertial navigation recursive motion model of the IMU, calculating the spatial displacement vector based on the smooth straight-line driving condition of the vehicle, calculating the heading mounting angle through the deviation of the horizontal plane vector angle, and correcting the three-axis attitude in all dimensions; obtaining a single set of mounting angle results based on sequence convergence judgment and confidence verification, and adaptively updating the three-axis mounting angle by combining the confidence of the mounting angle and the integrated navigation information; after each calculation of a new mounting angle, reverse avoidance of the heading mounting angle is performed based on dual inertial navigation recursion, and the correct mounting angle is output.

[0092] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0093] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for IMU mounting angle fusion combining heading and mounting angle reverse avoidance, characterized in that, include: A static gravity vector observation model for the vehicle-mounted IMU is established. The zero-speed detection algorithm is used to determine the stationary state of the vehicle. The roll and pitch installation angles are calculated and corrected by the gravity projection constraints of the navigation system and the vehicle system. A dynamic inertial navigation recursive motion model of IMU is established. The spatial displacement vector is calculated based on the smooth straight-line driving condition of the vehicle. The heading installation angle is calculated by the deviation of the horizontal plane vector angle, and the three-axis attitude is corrected in all dimensions. Based on the sequence convergence determination and confidence verification, a single set of installation angle results are obtained. Combining the confidence of the installation angle with the integrated navigation information, the three-axis installation angle is adaptively updated. After each new installation angle is calculated, the heading installation angle is reversed based on the dual inertial navigation recursion to output the correct installation angle.

2. The IMU mounting angle fusion method combining heading and mounting angle reverse avoidance as described in claim 1, characterized in that, A static gravity vector observation model for the vehicle-mounted IMU is established. A zero-speed detection algorithm is used to determine the vehicle's stationary state. Through gravity projection constraints between the navigation system and the vehicle system, the roll and pitch installation angles are calculated and corrected, including: The angular velocity modulus of the carrier was calculated from the real-time angular velocity of the three axes using the IMU gyroscope, and the acceleration modulus of the carrier was calculated from the real-time specific force of the three axes using the IMU accelerometer. Determine the gyroscope zero-speed angular velocity threshold, acceleration amplitude deviation threshold, and local gravitational acceleration amplitude. If the vehicle's angular velocity magnitude is less than the gyroscope zero-speed angular velocity threshold, the absolute value of the difference between the vehicle's acceleration magnitude and the local gravitational acceleration amplitude is less than the acceleration amplitude deviation threshold, and the duration of the current state is greater than or equal to the system's minimum stationary duration, then the vehicle is determined to be in a completely stationary state. When the vehicle is stationary, the standard gravity vector in the navigation coordinate system only includes the local gravitational acceleration amplitude. Multiple frames of static observation data from the accelerometer are collected and the mean value is extracted to eliminate random noise. The attitude transformation matrix from the navigation coordinate system to the IMU carrier coordinate system is used to obtain the measured gravity observation vector in the IMU carrier coordinate system. Based on the roll and pitch installation angles to be calculated, the attitude transformation matrix is ​​expanded, the gravity vector transformation relationship is established, and the standard gravity vector in the navigation coordinate system is substituted to obtain the IMU's roll and pitch installation angles. Update the attitude transformation matrix to complete the correction of the roll and pitch installation angles.

3. The IMU mounting angle fusion method combining heading and mounting angle reverse avoidance according to claim 1, characterized in that, A dynamic inertial navigation recursive motion model for the IMU is established. Based on the vehicle's smooth straight-line driving condition, the spatial displacement vector is calculated. The heading installation angle is calculated using the deviation of the horizontal plane vector angle, and the three-axis attitude is corrected in all dimensions, including: Determine the dynamic calibration operating conditions; Under dynamic calibration conditions, navigation quaternions are obtained, and attitude quaternions are obtained by recursively using quaternion multiplication operations in combination with the three-axis angular velocities of the carrier coordinate system output by the IMU gyroscope. The attitude matrix from the carrier coordinate system to the navigation coordinate system is obtained by solving the measured gravity observation vector and attitude quaternion in the carrier coordinate system. The measured gravity observation vector in the navigation coordinate system is obtained. After subtracting the standard gravity vector in the navigation coordinate system, multiple velocity integrations are performed to obtain the measured spatial displacement vector of the IMU in the navigation system during the calibration time period. Extract the north and east vectors of the measured spatial displacement vectors, obtain the forward horizontal reference vector from the north and east vectors, and calculate the heading installation deviation angle by solving the vector angle.

4. The IMU mounting angle fusion method combining heading and mounting angle reverse avoidance according to claim 1, characterized in that, Once the vehicle acceleration exceeds a threshold, a model for recursively solving for the velocity change is established using the inertial navigation system. The heading angle of the IMU is then calculated at each epoch using the angle between the velocity changes, including: Establish a horizontal reference coordinate system to eliminate roll and pitch deviations. The horizontal reference coordinate system is parallel to the horizontal plane of the vehicle coordinate system and has rotational deviation only in the heading direction. A first-direction cosine matrix is ​​established to transform the IMU carrier coordinate system to the horizontal reference coordinate system. Based on the first-direction cosine matrix, the coordinate system is transformed on the IMU measured vector to obtain the measured acceleration in the IMU carrier coordinate system and the measured acceleration in the horizontal reference coordinate system. Determine the acceleration threshold. When the magnitude of the resultant acceleration in the horizontal reference coordinate system is greater than or equal to the acceleration threshold, initiate the integral recursion of the measured acceleration in the horizontal reference coordinate system. Establish a second direction cosine matrix from the vehicle coordinate system to the horizontal reference coordinate system. The second direction cosine matrix only contains the simplified direction cosine matrix of the heading installation angle. The velocity change vector of the vehicle coordinate system is formed by the actual forward velocity change of the vehicle, wherein the actual forward velocity change of the vehicle is greater than 0. The velocity change vector in the horizontal reference coordinate system is obtained by the second direction cosine matrix and the velocity change vector of the vehicle coordinate system. The second direction cosine matrix is ​​expanded and multiplied with the velocity change vector of the vehicle coordinate system to obtain the X-axis velocity change, Y-axis velocity change and Z-axis velocity change in the horizontal plane. Based on the changes in velocity along the X-axis and Y-axis, the sine and cosine expressions for the heading angle are obtained. Combining these with the trigonometric identities of the arctangent functions in the four quadrants, the optimal heading angle is solved.

5. The IMU mounting angle fusion method combining heading and mounting angle reverse avoidance according to claim 1, characterized in that, Based on the sequence convergence determination and confidence check, a single set of installation angle results is obtained. Combining the confidence of the installation angle with the integrated navigation information, the three-axis installation angle is adaptively updated, including: Determine the heading angle deviation threshold between adjacent epochs, extract the most recent several epochs including the current epoch, and construct a heading installation angle sequence. If the heading installation angle sequence satisfies that the difference between adjacent heading installation angles is less than the heading angle deviation threshold, and the difference between the maximum and minimum heading installation angles in the heading installation angle sequence is less than twice the heading angle deviation threshold, then the heading installation angle sequence is determined to be a converged heading installation angle sequence. Determine the horizontal attitude disturbance threshold and the heading and turning disturbance threshold, obtain the epoch interval of the converged heading installation angle sequence, and obtain quantization constraints based on the fact that the sum of the variances of the roll angle sequence and the pitch angle sequence within the epoch interval is less than or equal to the horizontal attitude disturbance threshold, and the sum of the products of the effective output value of the Z-axis of the gyroscope at any epoch in the horizontal reference coordinate system and the epoch sampling interval is less than or equal to the heading and turning disturbance threshold. Based on the horizontal attitude disturbance threshold, the heading and turning disturbance threshold, and the threshold positive weighting coefficient, construct the installation angle confidence, and use the confidence threshold to filter the installation angle confidence to obtain a converged heading installation angle sequence that has passed the confidence verification. The mean filtering is performed on the calculated heading and mounting angle values ​​of multiple epochs in the converged and confidence-verified heading and mounting angle sequence. The mean value corresponding to the second half of the epochs in the sequence is used as the optimal estimated value of the heading and mounting angle. This value is then integrated with the optimal heading and mounting angle to obtain the three-axis mounting angle vector of the IMU relative to the vehicle coordinate system. Based on the three-axis mounting angle vector, multiple sets of mounting angle calculation results are collected and filtered. The confidence level of each set of mounting angle calculation results is used as the weight to perform weighted fusion of multiple sets of fusionable mounting angles. Outlier detection and adaptive updating are performed on the new calculated mounting angle set to remove abnormal working condition deviations and obtain the final optimal estimate of the IMU three-axis mounting angle.

6. The IMU mounting angle fusion method combining heading and mounting angle reverse avoidance according to claim 5, characterized in that, Based on the triaxial mounting angle vector, multiple sets of mounting angle calculation results are collected and filtered. Using the confidence level of each set of mounting angle calculation results as weight, multiple sets of fusionable mounting angles are weighted and fused. Outlier detection and adaptive updates are performed on the newly calculated mounting angle sets to remove abnormal operating condition deviations, obtaining the final optimal estimate of the IMU triaxial mounting angle, including: Calculate the GNSS-INS integrated navigation position information modulus of any set of installation angle calculation results. If it is determined that the GNSS-INS integrated navigation position information modulus is less than the position information threshold, it is determined that there is no abnormality in the settlement condition. If it is verified that the modulus of the difference between multiple consecutive sets of installation angle calculation results is less than the installation angle deviation threshold, then the multiple consecutive sets of installation angle calculation results are fusionable results. Multiple sets of fusionable results are collected cumulatively. Based on the total number of fusionable results, the installation angle obtained by weighted fusion of multiple fusionable results, and the cumulative confidence weight, an initial fusion result is obtained. Each time a new set of effective installation angle calculation results that has been screened by working conditions is added, the initial fusion result is dynamically iterated and fused to obtain an iterative fusion result. If the magnitude of the difference between the newly calculated installation angle group and the installation angle obtained by weighted fusion of multiple fusion results does not meet the requirement of being less than the installation angle deviation threshold, the newly calculated installation angle group is determined to be an outlier. If two consecutive newly calculated installation angle groups are both outliers, the installation angle change is determined to have temporal anomaly consistency. If the magnitude of the newly calculated installation angle group is greater than the position information threshold, the GNSS-INS integrated navigation position information is determined to be out of normal range. When both the temporal anomaly consistency and the GNSS-INS integrated navigation position information being out of normal range conditions are met, the iterative fusion result is updated to the latest valid installation angle calculation result. When only one of the temporal anomaly consistency and the GNSS-INS integrated navigation position information being out of normal range conditions is met, it is determined to be a normal anomaly, the outlier is removed and the current fusion result remains unchanged. A fusion weight stability threshold is determined. If the cumulative confidence weight is greater than or equal to the fusion weight stability threshold, the installation angle is determined to be stable, dynamic updates are stopped, and the current fusion value is used. Otherwise, the effective new installation angle group is included in the fusion, the weighted fusion value is dynamically updated, and the final optimal estimate of the IMU three-axis installation angle is obtained.

7. The IMU mounting angle fusion method combining heading and mounting angle reverse avoidance according to claim 1, characterized in that, After each new installation angle is calculated, based on the dual inertial navigation recursion, the heading installation angle is reversed to avoid errors and output the correct installation angle, including: The system acquires the northward velocity component, eastward velocity component, and groundward velocity component of the GNSS output. The northward velocity component and the eastward velocity component constitute the GNSS horizontal velocity. The system determines the GNSS horizontal velocity threshold and the GNSS horizontal velocity standard deviation threshold. If the GNSS horizontal velocity magnitude is greater than the GNSS horizontal velocity threshold and the ratio of the GNSS horizontal velocity standard deviation to the GNSS horizontal velocity magnitude is less than the GNSS horizontal velocity standard deviation threshold, the system enters the recursive process. The initial heading angle is calculated using the GNSS horizontal velocity, the initial pitch angle is calculated using the ground velocity component and the GNSS horizontal velocity magnitude, the initial roll angle is set to zero, the GNSS observation value is used as the GNSS initial velocity, and a recursive initial state is constructed. If it is determined that the above conditions are not met, the verification process is retried at the next moment. Based on the recursive initial state, the estimated value of the current calculated heading installation angle is used as the original installation angle. The original installation angle is reversed by 180 degrees to obtain the reverse installation angle. The original recursive position and the reverse recursive position corresponding to the original installation angle and the reverse installation angle are determined respectively. After reaching the preset shortest recursive time, the GNSS quantization and discrimination process is entered. Based on the actual working performance of GNSS, if the standard deviation modulus of GNSS horizontal position is determined to be less than or equal to the threshold value of the standard deviation modulus of GNSS horizontal position, then the GNSS position reliability judgment process is carried out; otherwise, the recursion continues. The difference magnitudes between the original recursive position and the reverse recursive position and the GNSS position at the same time are obtained as the original position deviation and the reverse position deviation, respectively. The position deviation ratio is constructed by the ratio of the reverse position deviation to the original position deviation, and the discrimination threshold and the abnormal deviation threshold are determined. If the original position deviation is greater than the abnormal deviation threshold, and the reverse position deviation is greater than the abnormal deviation threshold, then clear the current recursive result and wait for the next initial state verification to be satisfied before re-executing the process; otherwise, proceed to the subsequent judgment process. If the original position deviation is greater than the discrimination threshold, the original installation angle is determined to be correct, and the current calculated heading installation angle estimate is used for subsequent navigation calculations. If the original position deviation is less than the reciprocal of the discrimination threshold, it is determined that the original installation angle has a 180-degree reversal. The currently calculated estimated heading installation angle is then corrected to the reverse installation angle and input into the integrated navigation system. If none of the above conditions are met, the recursion continues, and the GNSS location reliability judgment process and subsequent discrimination process are re-executed.

8. An IMU mounting angle fusion system that combines heading and mounting angle reverse avoidance, characterized in that, include: The first module is used to establish the vehicle's gyroscope zero-bias estimation model and the accelerometer-gravity projection correlation model in a stationary state, and to solve for the roll and pitch installation angles. The second module is used to establish a model for recursively solving the velocity change when the vehicle acceleration exceeds the threshold. The model is then used to solve the heading installation angle of the IMU at each epoch using the angle between the velocity changes. The first calculation module is used to obtain a single set of installation angle results based on the sequence convergence judgment and confidence verification, and adaptively update the three-axis installation angle by combining the confidence of the installation angle with the integrated navigation information. The second calculation module is used to reverse the heading installation angle based on the dual inertial navigation recursion after each new installation angle is calculated, and output the correct installation angle.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the IMU mounting angle fusion method as described in any one of claims 1 to 7, which combines heading mounting angle reverse avoidance.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the IMU mounting angle fusion method as described in any one of claims 1 to 7, which combines heading mounting angle reverse avoidance.

Images