Tri-axial non-magnetic rotating platform and magnetic field sensor heading error calibration method

Abstract

The application provides a three-axis non-magnetic rotating platform and a magnetic field sensor heading error calibration method. The three-axis non-magnetic rotating platform comprises a platform base and a rotating module. The platform base is provided with a horizontal adjusting mechanism and a reference indicating module. The horizontal adjusting mechanism is used for adjusting the platform top surface to a horizontal state, and the reference indicating module is used for indicating the geomagnetic north direction and the horizontal state. The rotating module comprises a rotating base and a rotating turntable. The rotating turntable is rotatably installed on the rotating base and rotates around the axis of the rotating turntable. A detachable connecting structure is arranged between the rotating module and the platform base. The detachable connecting structure enables the rotating module to be installed in three mutually orthogonal installation attitudes with respect to the platform base. In the three mutually orthogonal installation attitudes, the axis of the rotating turntable is aligned with the three mutually orthogonal directions one by one. The application can realize three-axis rotation, unify the coordinate system, and ensure the consistency of the calibration conditions of the data collected by the magnetic field sensor.

Description

Technical Field

[0001] This invention relates to the field of magnetic field sensor calibration technology, specifically to a three-axis non-magnetic rotating platform, and also to a method for calibrating the heading error of a magnetic field sensor using the three-axis non-magnetic rotating platform. Background Technology

[0002] For the calibration of heading errors of magnetic field sensors, existing technologies mainly employ the following types of devices and methods: 1. Using a single-axis or dual-axis mechanical turntable, a single-axis motor, a dual-axis motor, or a hand-cranked turntable is used to drive the magnetic field sensor to rotate around one or two axes, collecting magnetic field data under different attitudes for fitting compensation; 2. Using a metal universal joint or gimbal, a universal joint or gimbal is constructed using metal brackets and metal bearings to fix the sensor, and then it is rotated manually to achieve multi-angle measurements; 3. Using simple manual clamps, metal screws, brackets, and other simple clamps are used to fix the sensor, and then the orientation is changed manually to measure the geomagnetic field at several angles. This method is mostly used for rough calibration or teaching demonstrations.

[0003] However, existing heading error calibration technologies lack a unified and standardized spatial coordinate system reference. Traditional platforms typically only provide mechanical rotation functions and lack integration with the geomagnetic north direction and the horizontal plane. Furthermore, operators often rely on experience to roughly align with the north or depend on external separate devices such as levels and compasses. This makes it difficult to ensure the consistency of coordinate references between different locations and times when conducting the same experiment. Consequently, the conditions for each calibration experiment are inconsistent, data comparability is poor, and it is not conducive to long-term error analysis and model optimization.

[0004] Moreover, traditional platforms have limited adjustable attitude dimensions, making it difficult to achieve true three-axis spatial orientation. For example, single-axis or dual-axis turntables can only rotate in one or two planes and cannot achieve complete rotation around three mutually perpendicular axes. Although some omnidirectional structures can achieve multi-degree-of-freedom motion, precise control and repeated positioning are difficult. The calibration process cannot cover the complete three-dimensional attitude space, and the heading error under complex three-axis combined attitudes is difficult to measure comprehensively.

[0005] Furthermore, the existing heading error calibration method involves rotating and sampling magnetic field data and fitting and compensating for the magnetic field data. It lacks a structured process for triaxial consistency and makes it difficult to separate the installation deviation of the magnetic field sensor and the attitude coupling error.

[0006] Therefore, there is an urgent need for a more optimized method for calibrating heading errors. Summary of the Invention

[0007] The primary objective of this invention is to provide a three-axis non-magnetic rotation platform that enables three-axis rotation, unifies the coordinate system, and ensures consistent calibration conditions for data acquisition by the magnetic field sensor.

[0008] The second objective of this invention is to provide a method for calibrating the heading error of a magnetic field sensor that enables consistent calibration of three-axis data and improves calibration accuracy. To achieve the aforementioned first objective, the triaxial non-magnetic rotating platform provided by the present invention includes a platform base and a rotating module. The platform base is provided with a horizontal adjustment mechanism and a reference indicator module. The horizontal adjustment mechanism is used to adjust the top surface of the platform to a horizontal state, and the reference indicator module is used to indicate the geomagnetic north direction and the horizontal state. The rotating module includes a rotating base and a rotating disk. The rotating disk is rotatably mounted on the rotating base and rotates about the axis of the rotating disk. A detachable connection structure is provided between the rotating module and the platform base. The detachable connection structure allows the rotating module to be installed relative to the platform base in one of three mutually orthogonal installation postures. In the three mutually orthogonal installation postures, the rotation axis of the rotating disk is aligned with the three mutually orthogonal directions respectively.

[0009] As can be seen from the above scheme, in the three-axis non-magnetic rotary platform of the present invention, the detachable connection structure allows the same set of rotary modules to be selectively assembled into three orthogonal installation postures. The axes of the rotating turntable correspond to three sets of orthogonal directions, and the three-axis posture of the sensor can be adjusted using only a single set of rotary mechanisms. The platform base integrates a horizontal adjustment mechanism and a reference indicator module, which can quickly establish horizontal and northward references, reducing reliance on external auxiliary equipment. At the same time, the three orthogonal installation postures can provide magnetic field sampling data from multiple postures for the heading error calibration of the magnetic field sensor, unifying the coordinate system, ensuring consistent calibration conditions for the magnetic field sensor data acquisition, and improving the accuracy of the heading error calibration of the magnetic field sensor.

[0010] In a further embodiment, the leveling mechanism includes a platform top plate, a platform bottom plate, and leveling bolts. The platform top plate is mounted on the platform bottom plate via the leveling bolts, which are used to adjust the relative height between the platform top plate and the platform bottom plate to achieve leveling of the platform top surface.

[0011] As can be seen, the leveling mechanism uses a platform top plate, a platform bottom plate, and leveling bolts. The platform top surface is leveled by adjusting the relative height of the platform top plate and the platform bottom plate using the leveling bolts. The leveling operation is simple and can stably build a horizontal reference surface, avoiding calibration errors caused by platform tilt.

[0012] In a further embodiment, the reference indicator module includes a level and a compass, both of which are mounted on the platform base.

[0013] Therefore, the reference indication module uses a level and compass installed on the platform base to monitor the horizontal state and indicate the geomagnetic north direction. It is free from electromagnetic interference, can ensure the uniformity of the reference coordinate system for different batches of calibration tests, and improve the accuracy of the heading error calibration results of the magnetic field sensor.

[0014] In a further embodiment, the detachable connection structure includes a slot on the top plate of the platform and a plug-in part on the rotating base, with the plug-in part cooperating with the slot.

[0015] As can be seen, the detachable connection structure adopts a plug-in form in which the slot of the platform top plate and the plug-in part of the rotating base cooperate. It is convenient to disassemble and assemble, and facilitates quick switching between the three orthogonal installation attitudes of the rotating module. The plug-in cooperation has a good positioning effect, and the alignment accuracy is high when reassembling. It can reduce the installation deviation caused by attitude switching and improve the accuracy of heading error calibration.

[0016] In a further embodiment, the rotating module also includes a pointer mounted on a rotating base; the rotating disk is provided with circumferentially arranged angle scales, which are matched with the pointer angle scales.

[0017] Therefore, installing a pointer on the rotating base, in conjunction with the circumferential angle scale of the rotating turntable, forms a mechanical angle reading mechanism. This allows for intuitive reading of the rotation angle, supports manual fixed-point discrete sampling, facilitates precise control of the rotation position, reduces angle positioning errors, and ensures the data accuracy of the magnetic field sensor's heading error calibration.

[0018] To achieve the second objective of this invention, the magnetic field sensor heading error calibration method provided by this invention is applied to the aforementioned three-axis non-magnetic rotating platform. The method includes: adjusting the top surface of the platform to a horizontal state using a horizontal adjustment mechanism; aligning the three-axis non-magnetic rotating platform with the geomagnetic north direction using a reference indicator module to establish a unified reference coordinate system; fixing the magnetic field sensor to be calibrated onto the rotating module; sequentially installing the rotating module onto the platform base in three mutually orthogonal installation postures; in each installation posture, causing the rotating turntable to swing at a small angle near a preset reference angle, collecting first magnetic field data from the magnetic field sensor at at least two swing angle positions; and causing the rotating turntable to rotate around the current axis of rotation at full angle, collecting second magnetic field data at multiple discrete angle positions; establishing a measurement model including the magnetic field sensor's direction consistency offset matrix, heading error matrix, zero offset, and geomagnetic field modulus; establishing an optimization objective function based on multiple sets of first and second magnetic field data; solving the optimization objective function using a preset nonlinear optimization method to obtain the direction consistency offset matrix, heading error matrix, zero offset, and geomagnetic field modulus.

[0019] As can be seen from the above scheme, the magnetic field sensor heading error calibration method of the present invention utilizes a three-axis non-magnetic rotation platform to establish a reproducible unified reference coordinate system, avoiding systematic errors introduced by inconsistent benchmarks. Simultaneously, magnetic field data is acquired through three orthogonal installation postures. For each posture, first magnetic field data with small-angle oscillation and second magnetic field data with full-angle rotation are obtained. By fusing local geometric constraints of the magnetic field and spherical amplitude constraints of the geomagnetic field, the first and second magnetic field data form a complementary constraint relationship, which can suppress parameter divergence problems caused by noise and improve the stability of parameter identification. Furthermore, by establishing a measurement model including a direction consistency offset matrix, heading error matrix, zero bias, and geomagnetic field modulus, and using nonlinear optimization to achieve joint solution of multiple error parameters, sensor assembly deviations and systematic heading errors can be avoided, improving the accuracy of heading error identification.

[0020] In a further scheme, the objective function for optimization includes a small-angle swing geometric constraint term and a full-angle rotation spherical consistency term. The small-angle swing geometric constraint term and the full-angle rotation spherical consistency term correspond to the first magnetic field data and the second magnetic field data, respectively. The small-angle swing geometric constraint term is constructed based on the geometric constraint that the direction of change of the magnetic field vector during small-angle rotation is perpendicular to the magnetic field vector itself. The full-angle rotation spherical consistency term is constructed based on the spherical constraint that the geomagnetic field modulus remains constant during spatial rotation.

[0021] Therefore, the optimization objective function integrates small-angle oscillation geometric constraints and full-angle rotation spherical consistency terms corresponding to the first and second magnetic field data, respectively. The geometric constraints are established based on the geometric law that the direction of change of the magnetic field during small-amplitude rotation is orthogonal to the magnetic field vector, enhancing the ability to identify installation direction deviations. The spherical consistency term is constructed based on the constant characteristic of the geomagnetic field modulus, constraining amplitude-type errors. The joint participation of these two types of constraints in the optimization, with complementary information, can reduce the risk of multiple solutions in parameter solving, improve the convergence effect of nonlinear optimization, and enhance the accuracy of error parameter identification.

[0022] In a further proposed solution, the objective function is expressed as: ; ; The constraints are: ;in, This is the directional consistency offset matrix. Here is the heading error matrix; Zero bias, The magnitude of the geomagnetic field. and As a weighting factor, It is a small-angle swing geometric constraint term. It is a uniformity term for a spherical surface rotating at all angles. For the first The platform principal axis unit vector corresponding to the axis of rotation under a certain installation posture. This is the local directional derivative obtained from the first magnetic field data through the central difference approximation. For the first Under various installation postures at discrete angles The second magnetic field data collected at the location, This represents the number of discrete sampling points during full-angle rotation. 3 A 3-dimensional identity matrix.

[0023] In a further scheme, after obtaining the direction consistency offset matrix, heading error matrix, zero bias and geomagnetic field modulus, the scheme also includes: using the solved parameters to compensate for the new magnetic field measurement data of the magnetic field sensor, and verifying the compensation effect through the new magnetic field measurement data.

[0024] Therefore, after solving for the error parameters, further multi-dimensional error compensation is performed on the new magnetic field measurement data, and the compensation effect is verified by using independently re-measured new magnetic field measurement data, which can improve the measurement accuracy of the magnetic field sensor's heading.

[0025] In a further scheme, the steps of compensating the new magnetic field measurement data of the magnetic field sensor using the solved parameters and verifying the compensation effect using the new magnetic field measurement data include: compensating the new magnetic field measurement data using the solved direction consistency offset matrix, heading error matrix, zero bias and geomagnetic field modulus to obtain the magnetic field vector estimate in the reference coordinate system; calculating the heading angle using the horizontal component of the compensated magnetic field vector, and verifying the magnetic field modulus consistency and heading consistency respectively.

[0026] Therefore, by using the multi-dimensional error parameters obtained from calibration to jointly compensate for the independently measured new magnetic field measurement data, we can avoid assembly orientation deviation, heading system error and sensor zero bias, obtain the magnetic field vector in the reference coordinate system, and calculate the heading angle based on the horizontal magnetic field components. The compensation effect can be evaluated through a dual verification scheme of magnetic field magnitude consistency and heading consistency. This can identify calibration anomalies and improve the reliability of the heading error calibration results of the magnetic field sensor. Attached Figure Description

[0027] Figure 1 This is a structural diagram of an embodiment of the three-axis non-magnetic rotary platform of the present invention in one installation posture.

[0028] Figure 2 This is an exploded view of the structure of an embodiment of the three-axis non-magnetic rotary platform of the present invention.

[0029] Figure 3 This is a structural diagram of the rotating base in an embodiment of the three-axis non-magnetic rotating platform of the present invention.

[0030] Figure 4 This is a structural diagram of another installation posture of an embodiment of the three-axis non-magnetic rotary platform of the present invention.

[0031] Figure 5 This is a structural diagram of the three-axis non-magnetic rotary platform embodiment of the present invention under another installation posture.

[0032] Figure 6 This is a flowchart of an embodiment of the magnetic field sensor heading error calibration method of the present invention.

[0033] Figure 7 This is a schematic diagram of small-angle swing data acquisition in an embodiment of the magnetic field sensor heading error calibration method of the present invention.

[0034] The present invention will be further described below with reference to the accompanying drawings and embodiments. Detailed Implementation

[0035] Example of a three-axis non-magnetic rotary platform: like Figure 1 and Figure 2 As shown, in this embodiment, the three-axis non-magnetic rotating platform includes a platform base 1 and a rotating module 2. The platform base 1 is provided with a horizontal adjustment mechanism 11 and a reference indicator module 12. The horizontal adjustment mechanism 11 is used to adjust the top surface of the platform to a horizontal state, and the reference indicator module 12 is used to indicate the geomagnetic north direction and the horizontal state.

[0036] In this embodiment, the leveling mechanism 11 includes a platform top plate 111, a platform bottom plate 112, and leveling bolts 113. The platform top plate 111 is mounted on the platform bottom plate 112 via the leveling bolts 113. The leveling bolts 113 are used to adjust the relative height between the platform top plate 111 and the platform bottom plate 112 to achieve leveling of the platform top surface. The leveling mechanism 11 uses the platform top plate 111, platform bottom plate 112, and leveling bolts 113 to adjust the relative height between the platform top plate 111 and the platform bottom plate 112 to achieve leveling of the top surface. The leveling operation is simple, can stably construct a horizontal reference plane, and avoid calibration errors caused by platform tilt. Preferably, the leveling bolts 113 are made of pure titanium to reduce magnetic interference.

[0037] In this embodiment, the reference indication module 12 includes a level (not shown) and a compass (not shown), both of which are mounted on the platform base 1. The reference indication module 12 uses the level and compass mounted on the platform base 1 to monitor the horizontal state and indicate the geomagnetic north direction. It is free from electromagnetic interference, ensures that the reference coordinate system is the same for different batches of calibration tests, and improves the consistency of the calibration results of the magnetic field sensor heading error.

[0038] In this embodiment, the rotating module 2 includes a rotating base 21 and a rotating disk 22. The rotating disk 22 is rotatably mounted on the rotating base 21 and rotates about its axis of rotation. The rotating base 21 and the rotating disk 22 are connected by a bearing assembly 23, which uses a ceramic bearing to reduce magnetic interference.

[0039] In this embodiment, a detachable connection structure 3 is provided between the rotating module 2 and the platform base 1. The detachable connection structure 3 allows the rotating module 2 to be installed relative to the platform base 1 in one of three mutually orthogonal installation postures. In the three mutually orthogonal installation postures, the rotation axis of the rotating disk 22 is aligned with the three mutually orthogonal directions respectively.

[0040] In this embodiment, the detachable connection structure 3 includes a slot 1111 disposed on the platform top plate 111 and a plug portion 211 disposed on the rotating base 21 (see...). Figure 3 The detachable connection structure 3 uses a plug-in connection where the slot 1111 of the platform top plate 111 engages with the plug-in part 211 of the rotating base 21. This facilitates easy assembly and disassembly, allowing for quick switching between the three orthogonal installation attitudes of the rotating module 2. The plug-in connection provides excellent positioning, ensuring high alignment accuracy during repeated assembly and reducing installation deviations caused by attitude switching, thus improving the accuracy of heading error calibration. The three orthogonal installation attitudes are as follows: Figure 1 , Figure 4 and Figure 5 As shown.

[0041] In this embodiment, the rotation module 2 also includes a pointer 24, which is mounted on the rotation base 21. The rotating disk 22 is provided with circumferential angle scales (not shown), and the pointer 24 is configured to cooperate with the circumferential angle scales. The pointer 24, mounted on the rotation base 21, cooperates with the circumferential angle scales of the rotating disk 22 to form a mechanical angle reading mechanism. This allows for intuitive reading of the rotation angle, supports manual fixed-point discrete sampling, facilitates precise control of the rotation position, reduces angle positioning errors, and ensures the data accuracy of the magnetic field sensor heading error calibration.

[0042] In addition, a sensor mounting base 25 is provided on the rotating turntable 22, which is used to mount the magnetic field sensor to be calibrated.

[0043] The three-axis non-magnetic rotary platform of this embodiment allows the same set of rotary modules 2 to be selectively assembled into three orthogonal installation postures through a detachable connection structure 3. The axes of the rotating turntable 22 correspond to three sets of orthogonal directions, and the three-axis attitude adjustment of the sensor can be achieved using only a single set of rotary mechanisms. The platform base 1 integrates a horizontal adjustment mechanism 11 and a reference indicator module 12, which can quickly establish horizontal and northward references, reducing reliance on external auxiliary equipment. At the same time, the three orthogonal installation postures can provide magnetic field sampling data from multiple postures for the heading error calibration of the magnetic field sensor, unify the coordinate system, ensure the consistency of calibration conditions for the magnetic field sensor data, and improve the accuracy of the heading error calibration of the magnetic field sensor.

[0044] Example of a method for calibrating heading error of a magnetic field sensor: In this embodiment, the magnetic field sensor heading error calibration method is based on the magnetic field sensor heading error calibration method of the three-axis non-magnetic rotating platform in the above embodiment.

[0045] See Figure 5 In this embodiment, see Figure 5 When calibrating the heading error of the magnetic field sensor, step S1 is executed first. The platform top surface is adjusted to a horizontal state using the leveling mechanism 11, and the three-axis non-magnetic rotating platform is aligned with the geomagnetic north direction using the reference indicator module 12, establishing a unified reference coordinate system. The reference coordinate system uses the vertically upward direction as one axis and two mutually orthogonal directions in the horizontal plane as the other two axes. When calibrating the heading error of the magnetic field sensor, the platform base 1 is placed on the experimental platform or ground. The platform top plate 111 is leveled using the leveling bolts 113, and the platform base 1 is aligned with the geomagnetic north direction, ensuring the compass points north. This allows for the construction of a unified reference coordinate system that can be repeatedly established in any geographical location. (See also...) Figure 1 The three axes of this reference coordinate system are defined as follows: the x-axis is defined as the direction of the long side of the platform, and its positive direction is the geographic North Pole; the y-axis is defined as the direction of the short side of the platform, which is orthogonal to the x-axis in the horizontal plane; and the z-axis is defined as the direction of the vertical upward direction.

[0046] After adjusting the top surface of the platform to a horizontal position and aligning the three-axis non-magnetic rotating platform with the geomagnetic north direction, proceed to step S2 to fix the magnetic field sensor to be calibrated onto the rotating module 2. This is done by mounting the magnetic field sensor onto the sensor mounting base 25. Alternatively, step S2 can be performed before step S1.

[0047] After the magnetic field sensor is installed, step S3 is executed. The rotating module 2 is installed on the platform base 1 in three orthogonal installation postures. In each installation posture, the rotating disk 22 is oscillated at a small angle near a preset reference angle to collect the first magnetic field data of the magnetic field sensor at at least two oscillation angle positions. Then, the rotating disk 22 is rotated around the current rotation axis at a full angle to collect the second magnetic field data at multiple discrete angle positions. To complete the calibration operation, magnetic field data under three orthogonal installation postures needs to be acquired.

[0048] Specifically, when collecting magnetic field data, in the first... Installation methods ( First, perform small-angle swing data acquisition, such as... Figure 6 As shown, the rotating disk 22 revolves around the current rotation axis. Slight swaying to both sides, preferably. The angle is 5° to 10°. At two discrete angle points (discrete angle points refer to the positions corresponding to the two discrete sampling angles, i.e., the midpoint). Two points corresponding to the same angle when swaying left and right and Static sampling The first magnetic field data is obtained by averaging the results. ; ; in,, Indicates the first The average vector of the magnetic field measurements of the turntable on the negative side of the reference angle under various installation postures. Indicates the first The average vector of the magnetic field measurements of the turntable on the positive side of the reference angle under various installation postures. It is the first Magnetic field measurement under various installation orientations yes Take the average of the next two times. It is the first This installation method is placed at 0 degrees.

[0049] get and Then, the local directional derivative is approximated by the central difference: Since the small-angle oscillation is only used to obtain information about the direction of magnetic field changes, its local directional derivative term is approximated by the central difference and is only related to the direction of magnetic field changes, but not to the absolute value of the oscillation angle. Therefore, there is no need to accurately measure the oscillation angle.

[0050] Next, full-angle rotation data acquisition is performed. The rotating turntable 22 rotates around the current axis one full revolution (or rotates in steps), recording the magnetic field data at all angles. Since this device cannot automatically and continuously read the rotation angle, a manual point-by-point discrete step acquisition method is used: select the step angle to read... ( =10°), according to The turntable 22 is manually rotated point by point, and the rotation is stopped at each angle point. The angle value is read and recorded manually, and the sample is taken at rest. Taking the average of the results, we obtain the discrete mean: This forms a full-angle dataset, namely, the second magnetic field data: .

[0051] After collecting the first and second magnetic field data for three installation orientations, three sets of segmented data were obtained, namely, .

[0052] After obtaining the first magnetic field data and the second magnetic field data, step S4 is executed to establish a measurement model including the orientation consistency offset matrix, heading error matrix, zero bias and geomagnetic field modulus of the magnetic field sensor. An optimization objective function is established based on multiple sets of first magnetic field data and second magnetic field data. The optimization objective function is solved by a preset nonlinear optimization method to obtain the orientation consistency offset matrix, heading error matrix, zero bias and geomagnetic field modulus.

[0053] Specifically, first, a unified calibration function is constructed, and then in the... Under this installation method, the original triaxial output of the magnetic sensor satisfies: ,in, : Represents the measured three-axis output (from ); The orientation consistency offset matrix represents the inconsistency between the orientation of the sensor coordinate system and the reference coordinate system. : Heading error matrix; : The true vector of the geomagnetic field; : Indicates the magnitude of the geomagnetic field; Zero bias.

[0054] All first and second magnetic field data are integrated into a single nonlinear least squares optimization problem, while simultaneously solving for the orientation consistency offset matrix. Heading error matrix Zero bias and geomagnetic field modulus Heading error matrix The equivalent heading error model primarily operates on the platform's horizontal plane. It describes the systematic errors of the magnetic sensors in heading-related directions, but is not used to compensate for pitch or roll errors. It is an invertible matrix. The direction of the geomagnetic field is considered constant in one experiment, and its magnitude B is used as a scalar parameter to be estimated in the optimization.

[0055] In this embodiment, the objective function for optimization includes a small-angle swing geometric constraint term and a full-angle rotation spherical consistency term. The small-angle swing geometric constraint term and the full-angle rotation spherical consistency term correspond to the first magnetic field data and the second magnetic field data, respectively. The small-angle swing geometric constraint term is constructed based on the geometric constraint that the direction of change of the magnetic field vector during small-angle rotation is perpendicular to the magnetic field vector itself. The full-angle rotation spherical consistency term is constructed based on the spherical constraint that the geomagnetic field modulus remains constant during spatial rotation.

[0056] Specifically, the objective function is expressed as: ; ; ; The constraints are: ; in, This is the directional consistency offset matrix. Here is the heading error matrix; Zero bias, The magnitude of the geomagnetic field. and As a weighting factor, It is a small-angle swing geometric constraint term. It is a uniformity term for a spherical surface rotating at all angles. For the first The platform principal axis unit vector corresponding to the axis of rotation under a certain installation posture. , This is the local directional derivative obtained from the first magnetic field data through the central difference approximation. For the first Under various installation postures at discrete angles The second magnetic field data collected at the location, This represents the number of discrete sampling points during full-angle rotation. 3 A 3-dimensional identity matrix.

[0057] To optimize the objective function, a numerical optimization method based on nonlinear least squares is used to jointly solve for the unknown parameters. The parameters to be solved are uniformly represented as a parameter vector: To facilitate numerical solutions, It can be represented by an equivalent minimal parameterization method. For example, three independent parameters can be used to describe the corresponding spatial rotation. The parameterization method is a well-known method in this technical field. The specific parameterization form is not limited here, as long as the orthogonality and invertibility constraints are satisfied. , and It participates directly in optimization as a continuously differentiable variable.

[0058] To ensure the convergence of the optimization process, the initial parameter values ​​can be given in the following way: The initial value is set as the identity matrix, representing the installation direction; The initial value is set as the identity matrix, representing the ideal state with no heading error; The initial value can be roughly estimated from the mean of the full-angle rotation data; The initial value can be obtained from the average value of the original magnetic field measurement modulus.

[0059] Based on the given initial values, the error term corresponding to the first magnetic field data is... Error terms corresponding to the second magnetic field data This is simultaneously incorporated into the objective function to form the overall residual function. For example, the residual of the small-angle oscillation geometric constraint term is: The residual of the consistency term for a spherical surface rotating at all angles is: .

[0060] The residual function on the parameter vector It is continuously differentiable and can be solved using nonlinear least squares optimization algorithms known in the field, such as the Gauss-Newton method, the Levenberg-Marquardt method, or other equivalent iterative least squares optimization algorithms.

[0061] In each iteration, the residuals of the two types of error terms are calculated based on the current parameter estimates, and the parameters are updated until any of the following convergence conditions are met: the objective function decreases by less than a first preset threshold; the parameter update amount is less than a second preset threshold; or the maximum preset number of iterations is reached. The first preset threshold, the second preset threshold, and the maximum preset number of iterations are preset based on experimental data.

[0062] During the optimization process, the consistency of the offset matrix is ​​ensured by selecting the parameterization form or performing orthogonalization after iterative updates. Always satisfy orthogonal constraints: Geomagnetic field modulus Through full-angle rotation of the spherical consistency term Constrained to be constant, its value converges automatically during the optimization process. Weight coefficients and To balance the influence of small-angle oscillation data and full-angle rotation data in joint optimization, the parameters can be set according to the experimental noise level or the proportion of data volume. After optimization convergence, the final estimated values ​​of the parameters are obtained: The result simultaneously satisfies: local geometric consistency constraints under small-angle oscillation conditions; and spherical consistency constraints under full-angle rotation conditions.

[0063] In this embodiment, the objective function integrates a small-angle oscillation geometric constraint term and a full-angle rotation spherical consistency term corresponding to the first and second magnetic field data, respectively. The geometric constraint term is established based on the geometric law that the direction of change of the magnetic field during small-amplitude rotation is orthogonal to the magnetic field vector, enhancing the ability to identify installation direction deviations. The spherical consistency term is constructed based on the constant characteristic of the geomagnetic field modulus, constraining amplitude-type errors. The joint participation of these two types of constraints in the optimization, with complementary information, can reduce the risk of multiple solutions in parameter solving, improve the convergence effect of nonlinear optimization, and enhance the accuracy of error parameter identification.

[0064] After obtaining the orientation consistency offset matrix, heading error matrix, zero bias, and geomagnetic field modulus, step S5 is executed. The solved parameters are used to compensate for the new magnetic field measurement data from the magnetic field sensor, and the compensation effect is verified using the new magnetic field measurement data. After solving for the error parameters, multi-dimensional error compensation is further performed on the magnetic field measurement data, and the compensation effect is verified using independently measured new magnetic field measurement data, which can improve the heading measurement accuracy of the magnetic field sensor.

[0065] In this embodiment, the steps of compensating the new magnetic field measurement data of the magnetic field sensor using the solved parameters and verifying the compensation effect using the new magnetic field measurement data include: compensating the new magnetic field measurement data using the solved direction consistency offset matrix, heading error matrix, zero bias and geomagnetic field modulus to obtain the magnetic field vector estimate in the reference coordinate system; calculating the heading angle using the horizontal component of the compensated magnetic field vector, and verifying the magnetic field modulus consistency and heading consistency respectively.

[0066] Specifically, after obtaining calibration parameters such as the orientation consistency offset matrix, heading error matrix, zero bias, and geomagnetic field modulus, the magnetic sensor is installed in the same way, or reassembled according to the three installation attitudes, and a new set of independent verification magnetic field datasets is collected under the same or similar environmental conditions: Each set of verification magnetic field data includes: the first magnetic field data for the corresponding installation method; and the second magnetic field data for the corresponding installation method.

[0067] In the Under this installation method, the triaxial raw measurement output of the magnetic sensor at any given time is recorded as: ,in, The data format is consistent with that of the calibration phase, originating from the magnetic sensor output under the corresponding installation method. Based on the established unified measurement model: The following calibration parameters have been estimated through joint optimization: The estimated value of the orientation consistency offset matrix, used to represent the fixed orientation deviation of the sensor coordinate system relative to the reference coordinate system, satisfies: ; : The estimated value of the heading error matrix, which represents the systematic error of the magnetic sensor in the heading-related direction, is assumed to be an invertible matrix; : Triaxial zero bias estimate; The estimated value of the geomagnetic field modulus is considered constant within the same experiment or the same local area.

[0068] For new measurement data First, perform zero bias compensation: Then, using the directional consistency matrix and heading error matrix obtained from the calibration, the magnetic field vector is reverse-mapped to obtain the estimated value of the magnetic field vector in the reference coordinate system: ,in, : The magnetic field vector expressed in the reference coordinate system after compensation; subscript Indicates the reference coordinate system; because If a matrix is ​​orthogonal, its inverse is equivalent to its transpose, i.e., .

[0069] For any original magnetic field measurement value in the verification data Using the above compensation model, zero bias removal and error inversion are performed to obtain the estimated value of the magnetic field vector in the reference coordinate system: Take the horizontal component of the compensated magnetic field vector: ,in: , , : These are the reference coordinate systems. , Magnetic field components along the z-axis; in the reference coordinate system, – The plane is defined as the horizontal plane. The heading angle is calculated based on the horizontal component. : ,in, This is the arctangent function in the four quadrants, and its output range can be mapped to... or .

[0070] When verifying the full-angle rotation data, in any installation method Taking the independent retest data of full-angle rotation as an example, the verification process is as follows: Select the full-angle rotation verification dataset under this installation method. ,in, These are discrete rotation angles recorded manually; for each angle point, a heading error compensation model is applied to obtain... Based on the compensated magnetic field vector, the magnetic field magnitude consistency index and the heading consistency index are calculated.

[0071] The consistency of magnetic field modulus is verified as follows: Define the error in the magnitude of the magnetic field after compensation: Calculate its root mean square error: The verification passed under the following condition: relative error condition: Magnetic field modulus consistency threshold A threshold of 0.1% is selected, meaning that the relative root mean square error of the compensated magnetic field modulus does not exceed this threshold, at which point the magnetic field modulus consistency verification is considered successful. Threshold The settings can be configured based on the noise characteristics of the magnetic sensor, the stability of the ambient magnetic field, and the accuracy requirements of the specific task.

[0072] The heading consistency verification is as follows: When the independently measured full-angle rotation data is taken about the vertical axis of the platform, the compensated heading angle is... With turntable angle The relationship should be approximately linear. In the full-angle rotation verification, the difference between the compensated heading angle and the turntable angle is only a constant zero-position offset. Therefore, the difference between the heading angle and the turntable angle can be considered as a linear observation with respect to this constant offset and estimated using the least squares method. Specifically, after normalizing the difference between the heading angle and the turntable angle at all sampling points, the average value is taken to obtain the estimated value of the constant heading offset. The constant heading offset is estimated using the least squares method. And define the heading residual: ,in, The heading is zero biased by least squares estimation; This is an angle normalization operator used to avoid the effects of angle jumps by mapping the angle difference to a continuous interval, thus preventing... abrupt changes, for example: Instead Calculate the root mean square value of the heading residual: The verification passed under the following conditions: Heading error threshold A threshold of 0.1 degrees is selected. If the root mean square error of the compensated heading angle residual does not exceed this threshold, the heading consistency verification is considered successful. Threshold The heading accuracy can be adjusted according to the accuracy level of the magnetic sensor, the mechanical stability of the platform, and the application scenario.

[0073] When verifying small-angle swing data, any installation method can be used. Taking the independently measured small-angle swing data as an example, in the swing angle The mean magnetic field measurement was obtained by sampling from both sides at rest: After performing heading error compensation on the above data, we obtained: The directional derivative after central difference approximation compensation: Based on the platform's principal axis unit vector Construct a geometric consistency index for small-angle oscillation: The verification passed under the following conditions: Threshold The settings can be adjusted based on the noise level of the magnetic sensor and the mechanical stability of the platform. In one embodiment, Selectable as Magnitude; in another implementation, the relative improvement ratio can also be used as the criterion, that is, when the compensation... When the descent is not less than the preset ratio compared to before compensation, the geometric consistency verification of the small-angle swing is deemed successful.

[0074] In this embodiment, the following comprehensive judgment rules are used as the conditions for passing the heading error compensation verification: Under at least one installation method, the magnetic field modulus consistency verification for full-angle rotation is passed; when the rotation axis is a vertical axis, the heading consistency verification is passed; and the geometric consistency verification for small-angle oscillation is passed. If any of the above conditions are not met, it is preferable to perform any of the following processing steps: re-execute the independent re-measurement of magnetic field data acquisition; adjust the weighting coefficients. Then, re-execute the unified optimization solution; check the platform leveling and north compatibility status, and re-collect verification data after meeting the requirements.

[0075] By using multi-dimensional error parameters obtained from calibration to jointly compensate for independently measured magnetic field data, assembly orientation deviation, heading system error and sensor zero bias are avoided, and the magnetic field vector in the reference coordinate system is obtained. At the same time, the heading angle is calculated based on the horizontal magnetic field component, and the compensation effect is evaluated through a dual verification scheme of magnetic field magnitude consistency and heading consistency. This can identify calibration anomalies and improve the reliability of the magnetic field sensor heading error calibration results.

[0076] As described above, the magnetic field sensor heading error calibration method of this invention utilizes a three-axis non-magnetic rotation platform to establish a reproducible unified reference coordinate system, avoiding systematic errors introduced by inconsistent benchmarks. Simultaneously, magnetic field data is acquired through three orthogonal installation postures. For each posture, first magnetic field data with small-angle oscillation and second magnetic field data with full-angle rotation are simultaneously obtained. By fusing local geometric constraints of the magnetic field and spherical amplitude constraints of the geomagnetic field, the first and second magnetic field data form a complementary constraint relationship, which can suppress parameter divergence caused by noise and improve the stability of parameter identification. Furthermore, by establishing a measurement model including a direction consistency offset matrix, heading error matrix, zero bias, and geomagnetic field modulus, and using nonlinear optimization to achieve joint solution of multiple error parameters, sensor assembly deviations and systematic heading errors can be avoided, improving the accuracy of heading error identification.

[0077] It should be noted that the above are only preferred embodiments of the present invention, but the design concept of the invention is not limited thereto. Any non-substantial modifications made to the present invention using this concept also fall within the protection scope of the present invention.

Claims

1. A three-axis non-magnetic rotary platform, characterized in that: Includes platform base and rotating module; The platform base is equipped with a horizontal adjustment mechanism and a reference indicator module. The horizontal adjustment mechanism is used to adjust the top surface of the platform to a horizontal state, and the reference indicator module is used to indicate the geomagnetic north direction and the horizontal state. The rotating module includes a rotating base and a rotating disk. The rotating disk is rotatably mounted on the rotating base and rotates about the axis of rotation of the rotating disk. A detachable connection structure is provided between the rotating module and the platform base, which allows the rotating module to be installed relative to the platform base in one of three mutually orthogonal installation postures. In the three mutually orthogonal mounting postures, the axis of rotation of the rotating disk is aligned with the three mutually orthogonal directions respectively.

2. The three-axis non-magnetic rotary platform according to claim 1, characterized in that: The leveling mechanism includes a platform top plate, a platform bottom plate, and leveling bolts. The platform top plate is mounted on the platform bottom plate via the leveling bolts. The leveling bolts are used to adjust the relative height between the platform top plate and the platform bottom plate to achieve leveling of the platform top surface.

3. The three-axis non-magnetic rotary platform according to claim 2, characterized in that: The reference indicator module includes a level and a compass, both of which are mounted on the platform base.

4. The three-axis non-magnetic rotary platform according to claim 2, characterized in that: The detachable connection structure includes a slot provided on the top plate of the platform and a plug-in part provided on the rotating base, the plug-in part being configured to cooperate with the slot.

5. The three-axis non-magnetic rotary platform according to any one of claims 1 to 4, characterized in that: The rotation module also includes a pointer, which is mounted on the rotation base; The rotating disk is provided with circumferential angle scales, and the pointer is configured in conjunction with the angle scales.

6. A method for calibrating the heading error of a magnetic field sensor, applied to the three-axis non-magnetic rotary platform according to any one of claims 1 to 5, characterized in that: The method includes: The platform top surface is adjusted to a horizontal state using the horizontal adjustment mechanism, and the three-axis non-magnetic rotating platform is aligned with the geomagnetic north direction using the reference indicator module to establish a unified reference coordinate system; The magnetic field sensor to be calibrated is fixed on the rotating module; The rotating module is installed on the platform base in three orthogonal installation postures. In each installation posture, the rotating turntable swings at a small angle near a preset reference angle to collect the first magnetic field data of the magnetic field sensor at at least two swing angle positions. The rotating turntable is then rotated around the current axis of rotation to collect the second magnetic field data at multiple discrete angle positions. A measurement model is established, including the orientation consistency offset matrix, heading error matrix, zero bias, and geomagnetic field modulus of the magnetic field sensor. An optimization objective function is established based on multiple sets of first magnetic field data and second magnetic field data. The optimization objective function is solved by a preset nonlinear optimization method to obtain the orientation consistency offset matrix, the heading error matrix, the zero bias, and the geomagnetic field modulus.

7. The method for calibrating the heading error of a magnetic field sensor according to claim 6, characterized in that: The optimization objective function includes a small-angle swing geometric constraint term and a full-angle rotation spherical consistency term, wherein the small-angle swing geometric constraint term and the full-angle rotation spherical consistency term correspond to the first magnetic field data and the second magnetic field data, respectively. The small-angle swing geometric constraint term is constructed based on the geometric constraint that the direction of change of the magnetic field vector during small-angle rotation is perpendicular to the magnetic field vector itself; The full-angle rotating spherical consistency term is constructed based on the spherical constraint that the geomagnetic field modulus remains constant during spatial rotation.

8. The method for calibrating the heading error of a magnetic field sensor according to claim 7, characterized in that: The optimization objective function is expressed as: ； ； ； The constraints are: ； in, This is the direction consistency offset matrix. The heading error matrix is... For the zero bias, The magnitude of the geomagnetic field is given. and As a weighting factor, It is the small-angle swing geometric constraint term. It is the consistency term of the spherical surface under full rotation. For the first The platform principal axis unit vector corresponding to the axis of rotation under a certain installation posture. This is the local directional derivative obtained from the first magnetic field data through a central difference approximation. For the first Under various installation postures at discrete angles The second magnetic field data collected at the location, This represents the number of discrete sampling points during full-angle rotation. 3 A 3-dimensional identity matrix.

9. The method for calibrating the heading error of a magnetic field sensor according to any one of claims 6 to 8, characterized in that: After obtaining the direction consistency offset matrix, the heading error matrix, the zero offset, and the geomagnetic field modulus, the method further includes: The new magnetic field measurement data of the magnetic field sensor is compensated using the solved parameters, and the compensation effect is verified using the new magnetic field measurement data.

10. The method for calibrating the heading error of a magnetic field sensor according to claim 9, characterized in that: The steps of compensating the new magnetic field measurement data of the magnetic field sensor using the solved parameters and verifying the compensation effect using the new magnetic field measurement data include: The new magnetic field measurement data are compensated using the obtained direction consistency offset matrix, heading error matrix, zero offset and geomagnetic field modulus to obtain the magnetic field vector estimate in the reference coordinate system. The heading angle is calculated using the horizontal component of the compensated magnetic field vector, and the consistency of magnetic field magnitude and heading are verified respectively.

Images