IMU internal reference calibration method

By acquiring the initial rest time and segmented data of the IMU for calibration, the problem of low calibration accuracy of the IMU is solved, and low-cost and efficient IMU internal reference calibration is achieved, which simplifies the calibration process and improves the accuracy.

CN118913309APending Publication Date: 2024-11-08FORYOU GENERAL ELECTRONICS
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410968496.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In the existing SLAM technology, the IMU calibration accuracy is not high, and the commonly used high-precision calibration methods are costly and not practical. The calibration cost of the turntable is expensive and the accuracy is not high in the existing calibration solution.

Method used

By obtaining the initial rest time of the target IMU, calculating the Allan standard deviation, collecting accelerometer and gyroscope data in segments, and using static and dynamic data for IMU internal calibration, it realizes accurate calibration of the accelerometer and gyroscope, and reduces calibration costs.

Benefits of technology

It realizes high-precision IMU internal reference calibration without high-precision equipment, reduces calibration costs, improves calibration accuracy, and simplifies calibration process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118913309A_ABST
    Figure CN118913309A_ABST
Patent Text Reader

Abstract

The invention provides an IMU (Inertial Measurement Unit) internal reference calibration method. The method comprises the following steps: step 1, acquiring initial static time of a target IMU; step 2, calculating Allan standard deviation in the initial static time according to the initial static time; step 3, collecting second calibration data of the target IMU under a preset condition, wherein the second calibration data comprises accelerometer data and gyroscope data; 4, calibrating the accelerometer of the target IMU according to the second calibration data, recording the residual error and the internal reference of the accelerometer, and taking the internal reference corresponding to the minimum residual error as the final internal reference of the accelerometer; and 5, calibrating the gyroscope of the target IMU by using the dynamic data and the static data corresponding to the minimum residual error. The method realizes calibration without using high-precision equipment such as a rotary table and the like, and is low in scheme cost, simple, efficient and high in calibration precision.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of SLAM technology, and in particular to an IMU internal reference calibration method. Background Art

[0002] In SLAM (Simultaneous Localization and Mapping), if only cameras or IMUs (Inertial Measurement Units) are used for mapping or positioning, the accuracy is not high and the robustness is poor. If only cameras are used, mapping cannot be continued or the mapping error is large in areas with less texture or fast movement speed (causing image blur); if only IMUs are used, since IMUs are easily disturbed by noise, there will be large errors over time. Therefore, a solution for the fusion of cameras and IMUs has been proposed. When the camera is in an area with better texture, the camera is used to calibrate the IMU. When the texture is weak or there is no texture or the image is blurred, the IMU is used to provide information for mapping and positioning, so as to improve the accuracy and robustness of mapping and positioning.

[0003] Both the camera and IMU need to be calibrated before data fusion. The accuracy of the commonly used vehicle-mounted IMU (basically MEMS-IMU type) is not as high as that of aerospace-grade IMU, so there is no need to use high-precision instruments and equipment for calibration. Even if high-precision instruments and equipment are used for calibration, the results obtained will still have certain errors due to their own low accuracy. At present, some calibration schemes use the turntable calibration method, but the cost is relatively high; there are also some that do not require the use of a turntable and can be calibrated by different rotations or flat placement, but the calibration accuracy is not high. Summary of the invention

[0004] The present invention provides an IMU internal parameter calibration method, which aims to solve the defects in the prior art and realize calibration without using high-precision equipment such as a turntable. The solution has low cost, simplicity and efficiency, and high calibration accuracy.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] The present invention provides an IMU internal parameter calibration method, comprising:

[0007] Step 1: Get the initial stationary time of the target IMU;

[0008] Step 2: Calculate the Allan standard deviation within the initial stationary time according to the initial stationary time;

[0009] Step 3: collecting second calibration data of the target IMU under preset conditions, where the second calibration data includes accelerometer data and gyroscope data;

[0010] Step 4: calibrate the accelerometer of the target IMU according to the second calibration data, and record the residual and internal parameter of the accelerometer, and use the internal parameter corresponding to the minimum residual as the final internal parameter of the accelerometer;

[0011] Step 5: Use the dynamic data and static data corresponding to the minimum residual to calibrate the gyroscope of the target IMU.

[0012] Specifically, the step 1 includes:

[0013] Step 101: placing a target IMU in a preset environment and keeping it still, and collecting data on the target IMU to obtain first calibration data, where the first calibration data includes accelerometer data and gyroscope data;

[0014] Step 102: Determine an initial stationary time according to the accelerometer data in the first calibration data.

[0015] Specifically, the step 102 includes:

[0016] Step 1021, setting an adjustable data segmentation time length, wherein the data segmentation time length takes values ​​in sequence from a set adjustable time range;

[0017] Step 1022: segment the first calibration data according to the data segmentation duration;

[0018] Step 1023, calculating the Allan variance according to the first preset relationship;

[0019] Step 1024, determine whether the current data segment duration makes the Allan variance converge, if yes, determine the initial static time according to the second preset relationship, otherwise take the next data segment duration and return to step 1022.

[0020] Specifically, the first preset relationship is:

[0021]

[0022] Among them, δ 2 (t) represents the Allan variance, t represents the length of the data segment, K represents the number of data segments, x(t,k) represents the average value of the kth segment data, and x(t,k-1) represents the average value of the k-1th segment data;

[0023] The second preset relationship is:

[0024] t init =m*t, where m≤K is the minimum value for the Allan variance to converge, and t is the length of the data segmentation.

[0025] Specifically, the step 3 includes:

[0026] Step 301, making the target IMU maintain the initial static time in any posture;

[0027] Step 302: Continue to collect second calibration data under preset conditions until the collected second calibration data meets the calibration accuracy. The preset conditions are: control the target IMU to rotate to any angle greater than a preset threshold and then remain stationary for any time, and repeat a preset number of times.

[0028] Specifically, step 4 includes:

[0029] Step 401, selecting decision coefficients in order from a first numerical interval;

[0030] Step 402: Determine the second calibration data as static data or dynamic data according to a first rule, wherein the first rule is: Then the second calibration data is determined to be static data, otherwise the second calibration data is determined to be dynamic data, wherein is the Allan standard deviation at time t;

[0031] Step 403: calibrate the accelerometer of the target IMU using the static data, and record the residual and internal parameters of the accelerometer;

[0032] Step 404, determine whether the determination coefficients within the first numerical interval have been taken, if yes, take the internal parameter corresponding to the minimum residual as the final internal parameter of the accelerometer, otherwise take the next determination coefficient within the first numerical interval and return to step 402.

[0033] Specifically, step 403 includes:

[0034] Step 4031, initializing the accelerometer internal parameters;

[0035] Step 4032, calculating a predicted calibration value based on the static data;

[0036] Step 4033, constructing an acceleration error function;

[0037] Step 4034: Use the LM algorithm to update the parameters to be calibrated in the fifth preset relationship, and determine whether the error value is less than the preset threshold value according to the error function. If yes, use the parameters obtained at this time as the calibration parameters of the accelerometer, otherwise return to step 4032.

[0038] The fifth preset relationship is:

[0039]

[0040] Specifically, the step 5 includes:

[0041] Step 501, initializing the gyroscope internal parameters;

[0042] Step 502: Calculate a first rotation vector, and use the first rotation vector as a true value, where the first rotation vector is a rotation vector corresponding to the rotation calculated by using all calibration data according to a sixth preset relationship through accelerometer information;

[0043] Step 503, calculating a second rotation vector, where the second rotation vector is a rotation vector obtained by using the angular velocity information corresponding to the acceleration information in step 502 and calculating through RK4 integration;

[0044] Step 504: construct a rotation vector error model;

[0045] Step 505: Perform iterative optimization through the LM algorithm until a gyroscope internal parameter is calculated that makes the rotation vector error less than a preset threshold.

[0046] Specifically, step 501 includes:

[0047] Step A1, using the average of the angular velocity data obtained when the target IMU is stationary as the initial bias value of the gyroscope;

[0048] Step A2: obtaining the average angular velocity of each group according to the acceleration information of the three groups of rotations of the target IMU, and using the average angular velocity as the calibration value of the angular velocity of each group of the corresponding gyroscope;

[0049] Step A3, using the average value of the uncalibrated gyroscope measurement values ​​of the target IMU corresponding to each group of rotation stages as the measurement value of the angular velocity of each group, and obtaining an analytical solution through the seventh preset relationship as the initial value of the gyroscope axis deviation and scale;

[0050] The step A2 comprises:

[0051] Step A201, obtaining corresponding rotation vectors from three groups of rotation acceleration information of the target IMU through a sixth preset relationship;

[0052] Step A202: Divide the rotation vector by the rotation time corresponding to each group to obtain the average angular velocity of each group, and use the average angular velocity as the calibration value of the angular velocity of each group of the corresponding gyroscope.

[0053] Specifically, the sixth preset relationship is:

[0054] a i = i · ipre

[0055]

[0056] Among them, a i represents the calibrated acceleration obtained in the i-th measurement, a ipre Indicates the calibrated acceleration obtained in the last measurement, R i Represents the rotation matrix from the last measurement to the current measurement, v i is the corresponding rotation vector.

[0057] The seventh preset relationship is:

[0058]

[0059] The beneficial effects of the present invention are as follows: the present invention obtains the initial stationary time of the target IMU, calculates the Allan standard deviation within the initial stationary time according to the initial stationary time, and divides the obtained calibration data into static data and dynamic data according to the Allan standard deviation, and then calibrates the target IMU using the static data and the dynamic data, thereby achieving calibration without the need to use high-precision equipment such as a turntable, and the solution has low cost, simplicity, efficiency, and high calibration accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flow chart of the IMU internal reference calibration method of the present invention. DETAILED DESCRIPTION

[0061] The following specifically describes the embodiments of the present invention in conjunction with the accompanying drawings, which are only for reference and illustration and do not constitute a limitation on the scope of patent protection of the present invention.

[0062] In the process described in the specification, claims or drawings of the present invention, the serial numbers of each step (such as step 10, 20, etc.) are included. The serial numbers are only used to distinguish the steps, and the serial numbers themselves do not represent any execution order. It should be noted that the descriptions such as "first" and "second" in this article are only used to distinguish the description objects, etc., and do not represent the order of precedence, nor do they mean that "first", "second", etc. are different types.

[0063] like Figure 1 As shown, this embodiment provides an IMU internal parameter calibration method, including:

[0064] Step 1: Get the initial stationary time t of the target IMU init .

[0065] In this embodiment, the target IMU is the IMU whose internal parameters are to be calibrated.

[0066] The IMU in this embodiment is a MEMS-IMU.

[0067] In this embodiment, step 1 includes:

[0068] Step 101: place a target IMU in a preset environment and keep it still, and collect data from the target IMU to obtain first calibration data, where the first calibration data includes accelerometer data and gyroscope data.

[0069] In a specific implementation, the preset environment is an environment with relatively stable temperature and magnetic field.

[0070] The collection time t0 of the first calibration data can be set to about 5 minutes. If the collection time t0 of the first calibration data cannot converge the Allan variance in the subsequent steps, it should be appropriately increased.

[0071] Step 102: Determine the initial stationary time t according to the accelerometer data in the first calibration data. init .

[0072] In this embodiment, step 102 includes:

[0073] Step 1021: Set an adjustable data segmentation time length t, wherein the data segmentation time length t takes values ​​in sequence from a set adjustable time range.

[0074] Step 1022: segment the first calibration data according to the data segmentation time t.

[0075] In the specific implementation, the data segmentation time length t takes values ​​in sequence from a set adjustable time range (for example, 1s~10s or 1s~30s, the time length is adjustable) (that is, 1s, 2s, 3s...10s, etc.), and the first calibration data is divided into K segments, then K=t0 / t.

[0076] Step 1023: Calculate the Allan variance δ according to the first preset relationship 2 (t).

[0077] In this embodiment, the first preset relationship is:

[0078]

[0079] Among them, δ 2 (t) represents the Allan variance, t represents the length of the data segment, K represents the number of data segments, x(t,k) represents the average value of the kth segment data, and x(t,k-1) represents the average value of the k-1th segment data.

[0080] Step 1024: determine whether the current data segmentation time length t makes the Allan variance converge. If so, determine the initial static time t according to the second preset relationship. init , otherwise take the next data segment length t and return to step 1022.

[0081] In this embodiment, the second preset relationship is:

[0082] t init =m*t, where m≤K is the minimum value for the Allan variance to converge, and t is the length of the data segmentation.

[0083] Step 2: According to the initial stationary time t init Calculate the Allan standard deviation during the initial stationary time

[0084] In this embodiment, step 2 includes:

[0085] The initial static time t init Assign to t w , and calculate the Allan standard deviation according to the following third preset relationship That is, the Allan standard deviation during the initial stationary time:

[0086]

[0087] in, represents the Allan standard deviation, represents t centered at time t w Allan variance of the x, y, and z axes over the time period.

[0088] Easy to understand,

[0089] Step 3: Collect second calibration data of the target IMU under preset conditions, where the second calibration data includes accelerometer data and gyroscope data.

[0090] In this embodiment, step 3 includes:

[0091] Step 301: Make the target IMU maintain the initial static time t at any position init ;

[0092] For example, if the initial stationary time t init If the target IMU is held for 5 seconds, data is collected again.

[0093] Step 302: Continue to collect second calibration data under preset conditions until the collected second calibration data meets the calibration accuracy. The preset conditions are: control the target IMU to rotate to any angle greater than a preset threshold and then remain stationary for any time, and repeat a preset number of times n.

[0094] In a specific implementation, the preset threshold is 10°.

[0095] In specific implementation, the number of repeated operations n on the target IMU is about 40. If n is too small, the calibration accuracy will be affected.

[0096] For example, after keeping the target IMU in position 1 for 5 seconds, rotate it around any axis (or a combination of multiple axes) by angle 1 to position 2, and keep the target IMU stationary in position 2 for 3 seconds; then rotate it around any axis (or a combination of multiple axes) by angle 2 again to position 3, and keep the target IMU stationary in position 3 for 5 seconds; then rotate it around any axis (or a combination of multiple axes) by angle 3 again to position 4, and keep the target IMU stationary in position 4 for 2 seconds; ... and so on, until the set number of acquisitions (for example, 40 times) is completed. The above rotation angles 1, 2, 3, etc. can be greater than 10°, and no precise control is required.

[0097] It is easy to understand that in the above-mentioned repeated process of "rotation angle-stationary-rotation angle-stationary ...", the target IMU is always outputting the second calibration data.

[0098] In this step, the target IMU can be stationary for different times each time, and the rotation angle can also be different each time without the need for precise control. Compared with the current calibration scheme that uses high-precision equipment to control fixed rotation, stationary time and angle, the scheme of the present invention is simple, low-cost and highly efficient.

[0099] Step 4: calibrate the accelerometer of the target IMU according to the second calibration data, and record the residual and internal parameters of the accelerometer, and use the internal parameter corresponding to the minimum residual as the final internal parameter of the accelerometer.

[0100] In this embodiment, step 4 includes:

[0101] Step 401: Select decision coefficients k in order from a first numerical interval.

[0102] In this embodiment, the first numerical interval is [1, K], where K is the number of data segments.

[0103] In this embodiment, the decision coefficient k starts from 1.

[0104] Step 402: Determine the second calibration data as static data or dynamic data according to a first rule, wherein the first rule is: Then the second calibration data is determined to be static data, otherwise the second calibration data is determined to be dynamic data, wherein is the Allan standard deviation at time t.

[0105] In this step, the The calculation method is: the initial static time t init Assign to t w , and calculate the Allan standard deviation at time t according to the first preset relationship and the third preset relationship

[0106] Step 403: Use the static data to calibrate the accelerometer of the target IMU, and record the residual and internal parameters of the accelerometer.

[0107] The uncalibrated data and the calibrated data of the accelerometer comply with the model represented by the following fourth preset relationship:

[0108]

[0109] in, is the uncalibrated data, d is the calibrated data, C represents scale calibration, T represents axis deviation calibration, and b represents offset.

[0110] C, T, and b are the parameters to be calibrated of the target IMU.

[0111] The specific expressions of C and T are as follows:

[0112]

[0113] The fourth preset relationship can be converted into the following fifth preset relationship:

[0114]

[0115] In this embodiment, step 403 includes:

[0116] Step 4031, initialize the accelerometer internal parameters.

[0117] The specific initialization methods include:

[0118] ① Place the six faces of the target IMU facing down in a horizontal position and set [0, 0, g] T , [0, 0, -] T ,[0,g,0 T ,[0,-

[0119] g,0 T ,[g,0,0 T ,[-,0,0 TAs the true values ​​corresponding to the z-axis pointing downwards and upwards, the y-axis pointing downwards and upwards, and the x-axis pointing downwards and upwards respectively (g represents the local gravitational acceleration);

[0120] ② Obtain an analytical solution through the fifth preset relationship as the initial value of the accelerometer.

[0121] For example, if Y=AX, you can use X=(A T A) -1 A T Y to find the analytical solution.

[0122] Step 4032: Calculate a predicted calibration value d based on the static data.

[0123] The specific method is: substitute the static data obtained in step 3 into the fourth preset relationship Calculate the predicted calibration value d.

[0124] Step 4033: Construct acceleration error function L a .

[0125] In this embodiment, the acceleration error model L a for:

[0126]

[0127] Among them, L a is the acceleration error, is the calibrated value of the target IMU accelerometer.

[0128] The acceleration error model L in this step a It is constructed based on the fact that at the same location, when stationary, the magnitude of acceleration is consistent with the magnitude of gravity.

[0129] Step 4034: Use the LM algorithm to update the parameters to be calibrated in the fifth preset relationship, and determine whether the error value is less than the preset threshold value according to the error function. If yes, use the parameters obtained at this time as the calibration parameters of the accelerometer, otherwise return to step 4032.

[0130] The parameter to be calibrated in the fifth preset relational expression is the variable in the last item of the fifth preset relational expression.

[0131] Step 404, determine whether the decision coefficient k in the first numerical interval has been taken. If yes, take the internal parameter corresponding to the minimum residual as the final internal parameter of the accelerometer; otherwise, take the next decision coefficient k in the first numerical interval and return to step 402.

[0132] Step 5: Use the dynamic data and static data corresponding to the minimum residual to calibrate the gyroscope of the target IMU.

[0133] The minimum residual in this step is the minimum residual of the accelerometer recorded during the calibration process in step 4.

[0134] In this embodiment, step 5 includes:

[0135] Step 501: Initialize the gyroscope internal parameters.

[0136] The specific initialization methods include:

[0137] Step A1, using the average of the angular velocity data obtained when the target IMU is stationary as the initial bias value of the gyroscope;

[0138] The stationary state in this step refers to the stationary state of the target IMU in step 302 .

[0139] Step A2: according to the acceleration information of the three groups of rotations of the target IMU, obtain the average angular velocity of each group, and use the average angular velocity as the calibration value d of the angular velocity of each group of the corresponding gyroscope.

[0140] In this embodiment, step A2 includes:

[0141] Step A201, obtain the corresponding rotation vector of the three groups of rotation acceleration information of the target IMU through the sixth preset relationship.

[0142] In this embodiment, the sixth preset relationship is:

[0143] a i = i · ipre

[0144]

[0145] Among them, a i represents the calibrated acceleration obtained in the i-th measurement, a ipre Indicates the calibrated acceleration obtained in the last measurement, R i Represents the rotation matrix from the last measurement to the current measurement, v i is the corresponding rotation vector.

[0146] Step A202: Divide the rotation vector by the rotation time corresponding to each group to obtain the average angular velocity of each group, and use the average angular velocity as the calibration value d of the angular velocity of each group of the corresponding gyroscope.

[0147] Step A3: using the average value of the uncalibrated gyroscope measurement values ​​of the target IMU corresponding to each group of rotation stages as the measurement value of the angular velocity of each group And the analytical solution is obtained through the seventh preset relationship as the initial value of the gyroscope axis deviation and scale.

[0148] In this embodiment, the seventh preset relationship is:

[0149]

[0150] Step 502: Calculate a first rotation vector and use the first rotation vector as a true value. The first rotation vector is a rotation vector corresponding to the rotation calculated by using all calibration data according to a sixth preset relationship through accelerometer information.

[0151] Among them, the rotation vector corresponding to the stationary data is zero.

[0152] Step 503: Calculate a second rotation vector, where the second rotation vector is a rotation vector calculated by RK4 integration using the angular velocity information corresponding to the acceleration information in step 502.

[0153] RK4 is the fourth-order Runge-Kutta.

[0154] Step 504: Construct a rotation vector error model L g .

[0155] In this embodiment, the rotation vector error model L g for:

[0156]

[0157] in, represents the first rotation vector, Represents the second rotation vector.

[0158] Step 505: Perform iterative optimization using the LM algorithm until the rotation vector error L is calculated. g The gyroscope internal parameters C, T, b are less than the preset threshold.

[0159] The above disclosure is only a preferred embodiment of the present invention, which cannot be used to limit the scope of protection of the present invention. Therefore, equivalent changes made according to the scope of the patent application of the present invention are still within the scope covered by the present invention.

Claims

1. An IMU internal reference calibration method, characterized in that: include: Step 1: Get the initial stationary time of the target IMU; Step 2: Calculate the Allan standard deviation within the initial stationary time according to the initial stationary time; Step 3: collecting second calibration data of the target IMU under preset conditions, where the second calibration data includes accelerometer data and gyroscope data; Step 4: calibrate the accelerometer of the target IMU according to the second calibration data, and record the residual and internal parameter of the accelerometer, and use the internal parameter corresponding to the minimum residual as the final internal parameter of the accelerometer; Step 5: Use the dynamic data and static data corresponding to the minimum residual to calibrate the gyroscope of the target IMU.

2. The IMU internal reference calibration method according to claim 1, characterized in that: The step 1 comprises: Step 101: placing a target IMU in a preset environment and keeping it still, and collecting data on the target IMU to obtain first calibration data, where the first calibration data includes accelerometer data and gyroscope data; Step 102: Determine an initial stationary time according to the accelerometer data in the first calibration data.

3. The IMU internal reference calibration method according to claim 2, characterized in that: The step 102 includes: Step 1021, setting an adjustable data segmentation time length, wherein the data segmentation time length takes values ​​in sequence from a set adjustable time range; Step 1022: segment the first calibration data according to the data segmentation duration; Step 1023, calculating the Allan variance according to the first preset relationship; Step 1024, determine whether the current data segment duration makes the Allan variance converge, if yes, determine the initial static time according to the second preset relationship, otherwise take the next data segment duration and return to step 1022.

4. The IMU internal reference calibration method according to claim 3, characterized in that: The first preset relationship is: Among them, δ 2 (t) represents the Allan variance, t represents the length of the data segment, K represents the number of data segments, x(t,k) represents the average value of the kth segment data, and x(t,k-1) represents the average value of the k-1th segment data; The second preset relationship is: t init =m*t, where m≤K is the minimum value for the Allan variance to converge, and t is the length of the data segmentation.

5. The IMU internal reference calibration method according to claim 4, characterized in that: The step 3 comprises: Step 301, making the target IMU maintain the initial static time in any posture; Step 302: Continue to collect second calibration data under preset conditions until the collected second calibration data meets the calibration accuracy. The preset conditions are: control the target IMU to rotate to any angle greater than a preset threshold and then remain stationary for any time, and repeat a preset number of times.

6. The IMU internal reference calibration method according to claim 5, characterized in that: The step 4 comprises: Step 401, selecting decision coefficients in order from a first numerical interval; Step 402: Determine the second calibration data as static data or dynamic data according to a first rule, wherein the first rule is: Then the second calibration data is determined to be static data, otherwise the second calibration data is determined to be dynamic data, wherein is the Allan standard deviation at time t; Step 403: calibrate the accelerometer of the target IMU using the static data, and record the residual and internal parameters of the accelerometer; Step 404, determine whether the determination coefficients within the first numerical interval have been taken, if yes, take the internal parameter corresponding to the minimum residual as the final internal parameter of the accelerometer, otherwise take the next determination coefficient within the first numerical interval and return to step 402.

7. The IMU internal reference calibration method according to claim 6, characterized in that: The step 403 includes: Step 4031, initializing the accelerometer internal parameters; Step 4032, calculating a predicted calibration value based on the static data; Step 4033: construct an acceleration error function; Step 4034: Use the LM algorithm to update the parameters to be calibrated in the fifth preset relationship, and determine whether the error value is less than the preset threshold value according to the error function. If yes, use the parameters obtained at this time as the calibration parameters of the accelerometer, otherwise return to step 4032. The fifth preset relationship is:

8. The IMU internal reference calibration method according to claim 1, characterized in that: The step 5 comprises: Step 501, initializing the gyroscope internal parameters; Step 502: Calculate a first rotation vector, and use the first rotation vector as a true value, where the first rotation vector is a rotation vector corresponding to the rotation calculated by using all calibration data according to a sixth preset relationship through accelerometer information; Step 503, calculating a second rotation vector, where the second rotation vector is a rotation vector obtained by using the angular velocity information corresponding to the acceleration information in step 502 and calculating through RK4 integration; Step 504: construct a rotation vector error model; Step 505: Perform iterative optimization through the LM algorithm until a gyroscope internal parameter is calculated that makes the rotation vector error less than a preset threshold.

9. The IMU internal reference calibration method according to claim 8, characterized in that: The step 501 includes: Step A1, using the average of the angular velocity data obtained when the target IMU is stationary as the initial bias value of the gyroscope; Step A2: obtaining the average angular velocity of each group according to the acceleration information of the three groups of rotations of the target IMU, and using the average angular velocity as the calibration value of the angular velocity of each group of the corresponding gyroscope; Step A3, using the average value of the uncalibrated gyroscope measurement values ​​of the target IMU corresponding to each group of rotation stages as the measurement value of the angular velocity of each group, and obtaining an analytical solution through the seventh preset relationship as the initial value of the gyroscope axis deviation and scale; The step A2 comprises: Step A201, obtaining corresponding rotation vectors from three groups of rotation acceleration information of the target IMU through a sixth preset relationship; Step A202: Divide the rotation vector by the rotation time corresponding to each group to obtain the average angular velocity of each group, and use the average angular velocity as the calibration value of the angular velocity of each group of the corresponding gyroscope.

10. The IMU internal reference calibration method according to claim 9, characterized in that: The sixth preset relationship is: a i =R i ·a ipre Among them, a i represents the calibrated acceleration obtained in the i-th measurement, a ipre Indicates the calibrated acceleration obtained in the last measurement, R i Represents the rotation matrix from the last measurement to the current measurement, v i is the corresponding rotation vector. The seventh preset relationship is:

Citation Information

Patent Citations

  • Panorama video anti-shaking method and system, and portable terminal

    CN107040694A

  • IMU (Inertial Measurement Unit) continuous calibration method, device and equipment and storage medium

    CN117589200A