WO2017063386A1 - 一种姿态测量系统的精度校准方法 - Google Patents

一种姿态测量系统的精度校准方法 Download PDF

Info

Publication number
WO2017063386A1
WO2017063386A1 PCT/CN2016/088303 CN2016088303W WO2017063386A1 WO 2017063386 A1 WO2017063386 A1 WO 2017063386A1 CN 2016088303 W CN2016088303 W CN 2016088303W WO 2017063386 A1 WO2017063386 A1 WO 2017063386A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
accelerometer
ellipsoid
matrix
calculated
Prior art date
Application number
PCT/CN2016/088303
Other languages
English (en)
French (fr)
Inventor
涂睿
沈雪峰
赵文龙
戴文鼎
岳强
Original Assignee
上海华测导航技术股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海华测导航技术股份有限公司 filed Critical 上海华测导航技术股份有限公司
Priority to RU2017125953A priority Critical patent/RU2662458C1/ru
Priority to KR1020177023778A priority patent/KR102008597B1/ko
Priority to US15/542,931 priority patent/US10605619B2/en
Priority to EP16854769.3A priority patent/EP3364151B1/en
Publication of WO2017063386A1 publication Critical patent/WO2017063386A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C17/00Compasses; Devices for ascertaining true or magnetic north for navigation or surveying purposes
    • G01C17/38Testing, calibrating, or compensating of compasses
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C25/00Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C25/00Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
    • G01C25/005Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass initial alignment, calibration or starting-up of inertial devices
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01PMEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
    • G01P21/00Testing or calibrating of apparatus or devices covered by the preceding groups
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01PMEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
    • G01P15/00Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration
    • G01P15/18Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration in two or more dimensions

Definitions

  • the present invention relates to the field of measurement technologies, and in particular, to an accuracy calibration method for an attitude measurement system.
  • the attitude measurement system is a series of devices that can measure the object's spatial attitude (pitch angle, roll angle, heading angle). It has been widely used in many fields of industry. In recent years, with the continuous decline of hardware costs, each Various types of attitude measurement systems have begun to enter the daily life of thousands of households. Taking mobile phones as an example, most smart phones have built-in accelerometers, gyroscopes and electronic compasses, which constitute a simple low-cost attitude measurement system. . However, in most of the current attitude measurement systems, due to cost constraints, the accuracy and stability of the built-in sensors of the system are not high, resulting in unsatisfactory measurement results of the entire attitude measurement system. Calibrating the sensor without changing the system hardware is a very practical way to improve the measurement accuracy of the entire system. Therefore, the calibration method for studying low-cost attitude measurement systems has a strong practical application value.
  • the present invention provides an attitude measurement system. Accurate calibration method, reliable calibration results, high precision, and less time-consuming calibration.
  • An accuracy calibration method for an attitude measurement system includes the following steps:
  • the ecliptic fitting model is used to calibrate the zero offset, the scale factor and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system;
  • the electronic compass is aligned by the ellipsoid fitting model according to the compensated accelerometer data
  • the attitude is solved based on the compensated accelerometer data and the electronic compass data.
  • the step of calibrating the zero offset, the scale factor and the non-orthogonal angle between the accelerometers of the attitude measuring system by the ellipsoid fitting model is performed before performing the following steps: performing horizontal calibration on the accelerometer to eliminate Accelerometer original zero offset.
  • the step of calibrating the zero offset, the scale factor and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • the step of calibrating the zero offset, the scale factor and the non-orthogonal angle between the accelerometers of the attitude measuring system by the ellipsoid fitting model is performed after performing the following steps: when new accelerometer data is acquired After that, the newly acquired data can be corrected with the ellipsoid parameters.
  • the step of quantifying the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data comprises the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation and overflow prevention processing, wherein the three-axis data is marked as [H X0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • the following steps are performed: after the new electronic compass data is acquired, the ellipsoid parameter pair may be used. The newly acquired data is corrected.
  • the electronic compass updates the ellipsoid parameters by using the measurement data during the measurement process, and improves the ellipsoid The accuracy of the ball parameters.
  • the electronic compass uses the measurement data to update the ellipsoid parameters during the measurement process, and improves the accuracy of the ellipsoid parameters.
  • the following steps are performed: the collected electronic compass data is eliminated by statistical rules. The calculation point where the jump occurs.
  • the CPU can only make a fitting operation based on a small number of sample points in order to ensure that the memory stack does not overflow during the operation, so the fitting precision is limited;
  • the matrix operation has been improved, and the calculation process is optimized to reduce the calculation of the ellipsoid fitting by more than 90%.
  • the calculation of the fitting algorithm takes 563 ms; relatively, the adoption The improved algorithm takes only 14ms; the strategy of calibrating with measurement data improves the fault tolerance during calibration, which reduces the initial calibration requirements and reduces the operational complexity of the user.
  • Embodiment 1 is a flowchart of a method of Embodiment 1 of an accuracy calibration method for an attitude measurement system according to the present invention
  • Embodiment 2 is a flowchart of a method of Embodiment 2 of an accuracy calibration method for an attitude measurement system according to the present invention
  • Embodiment 3 is a flowchart of a method of Embodiment 3 of an accuracy calibration method for an attitude measurement system according to the present invention
  • Embodiment 4 is a flowchart of a method of Embodiment 4 of an accuracy calibration method for an attitude measurement system according to the present invention
  • Embodiment 5 is a flowchart of a method of Embodiment 5 of an accuracy calibration method of an attitude measurement system according to the present invention
  • FIG. 6 is a flowchart of a method of Embodiment 6 of an accuracy calibration method of an attitude measurement system according to the present invention
  • FIG. 7 is a flowchart of a method of Embodiment 7 of an accuracy calibration method of an attitude measurement system according to the present invention.
  • an accuracy calibration method of an attitude measurement system includes the following steps:
  • Step S1 calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by using an ellipsoid fitting model
  • the step S1 the step of calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S2 Compensating the accelerometer raw data by using the calculated ellipsoid parameters.
  • Step S3 correcting the electronic compass by using the ellipsoid fitting model according to the compensated accelerometer data
  • step S3 performing the quasi-step of the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data comprises the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation overflow prevention processing, wherein the three-axis data is marked as [H X0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S4 Compensating the electronic compass raw data by using the calculated ellipsoid parameters.
  • Step S5 Calculating the posture according to the compensated accelerometer data and the electronic compass data.
  • an accuracy calibration method of an attitude measurement system includes the following steps:
  • Step S1 performing horizontal calibration on the accelerometer to eliminate the original zero offset of the accelerometer
  • Step S2 calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by using an ellipsoid fitting model
  • the step S2 the step of calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S3 compensating the accelerometer raw data by using the calculated ellipsoid parameters
  • Step S4 correcting the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data
  • the step S4: performing a quasi-step on the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data includes the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation overflow prevention processing, wherein the three-axis data is marked as [H X0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S5 compensating the original data of the electronic compass by using the calculated ellipsoid parameters
  • Step S6 Calculating the posture according to the compensated accelerometer data and the electronic compass data.
  • an accuracy calibration method of an attitude measurement system includes the following steps:
  • Step S1 performing horizontal calibration on the accelerometer to eliminate the original zero offset of the accelerometer
  • Step S3 calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by using an ellipsoid fitting model
  • the step S3 the step of calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S4 compensating the accelerometer raw data by using the calculated ellipsoid parameters
  • Step S5 correcting the electronic compass by using the ellipsoid fitting model according to the compensated accelerometer data
  • step S5 performing the quasi-step of the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data comprises the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation overflow prevention processing, wherein the three-axis data is marked as [H X0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S6 compensating the original data of the electronic compass by using the calculated ellipsoid parameters
  • Step S7 Solving the posture according to the compensated accelerometer data and the electronic compass data.
  • an accuracy calibration method of an attitude measurement system includes the following steps:
  • Step S1 performing horizontal calibration on the accelerometer to eliminate the original zero offset of the accelerometer
  • Step S3 calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by using an ellipsoid fitting model
  • the step S3 the step of calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S4 After the new accelerometer data is acquired, the newly acquired data can be corrected by the ellipsoid parameter.
  • Step S5 Compensating the accelerometer raw data by using the calculated ellipsoid parameters.
  • Step S6 correcting the electronic compass by using the ellipsoid fitting model according to the compensated accelerometer data
  • the step S6: performing a quasi-step on the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data includes the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation overflow prevention processing, wherein the three-axis data is marked as [H X0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S7 compensating the original data of the electronic compass by using the calculated ellipsoid parameters
  • Step S8 Solving the posture according to the compensated accelerometer data and the electronic compass data.
  • an accuracy calibration method of an attitude measurement system includes the following steps:
  • Step S1 performing horizontal calibration on the accelerometer to eliminate the original zero offset of the accelerometer
  • Step S3 Zero-biasing and engraving of the accelerometer of the attitude measuring system through the ellipsoid fitting model The degree coefficient is calibrated with the non-orthogonal angle between the axes;
  • the step S3 the step of calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S4 After the new accelerometer data is acquired, the newly acquired data can be corrected by the ellipsoid parameter.
  • Step S5 Compensating the accelerometer raw data by using the calculated ellipsoid parameters.
  • Step S6 correcting the electronic compass by using the ellipsoid fitting model according to the compensated accelerometer data
  • the step S6: performing a quasi-step on the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data includes the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation overflow prevention processing, wherein the three-axis data is marked as [H X0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S7 Compensating the electronic compass raw data by using the calculated ellipsoid parameters.
  • Step S8 After the new electronic compass data is acquired, the newly collected data can be corrected by using the ellipsoid parameter.
  • Step S9 Calculating the posture according to the compensated accelerometer data and the electronic compass data.
  • an accuracy calibration method of an attitude measurement system includes the following steps:
  • Step S1 performing horizontal calibration on the accelerometer to eliminate the original zero offset of the accelerometer
  • Step S3 calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by using an ellipsoid fitting model
  • the step S3 the step of calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S4 After the new accelerometer data is acquired, the newly acquired data can be corrected by the ellipsoid parameter.
  • Step S5 Compensating the accelerometer raw data by using the calculated ellipsoid parameters.
  • Step S6 correcting the electronic compass by using the ellipsoid fitting model according to the compensated accelerometer data
  • the step S6: performing a quasi-step on the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data includes the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation overflow prevention processing, wherein the three-axis data is marked as [H X0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S7 Compensating the electronic compass raw data by using the calculated ellipsoid parameters.
  • Step S8 After the new electronic compass data is acquired, the newly collected data can be corrected by using the ellipsoid parameter.
  • Step S9 solving the posture according to the compensated accelerometer data and the electronic compass data Count.
  • Step S10 The electronic compass uses the measurement data to update the ellipsoid parameters in the measurement process, and improves the accuracy of the ellipsoid parameters.
  • an accuracy calibration method of an attitude measurement system includes the following steps:
  • Step S1 performing horizontal calibration on the accelerometer to eliminate the original zero offset of the accelerometer
  • Step S3 calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by using an ellipsoid fitting model
  • the step S3 the step of calibrating the zero offset, the scale coefficient and the non-orthogonal angle between the axes of the accelerometer of the attitude measuring system by the ellipsoid fitting model comprises the following steps:
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S4 After the new accelerometer data is acquired, the newly acquired data may be corrected by using an ellipsoid fitting parameter.
  • Step S5 Compensating the accelerometer raw data by using the calculated ellipsoid parameters.
  • Step S6 correcting the electronic compass by using the ellipsoid fitting model according to the compensated accelerometer data
  • the step S6: performing a quasi-step on the electronic compass by the ellipsoid fitting model according to the compensated accelerometer data includes the following steps:
  • the three-axis data of the electronic compass is acquired for a period of time, and the collected data is subjected to operation overflow prevention processing, wherein the three-axis data is marked as [HX 0 H Y0 H Z0 ] T ;
  • the ellipsoidal constraint matrix C is introduced at the same time and the matrix S is divided into blocks, wherein
  • the ellipsoid fitting vector is obtained by the following formula:
  • ⁇ 2 -S 4 -1 S 2 T ⁇ 1 ;
  • ⁇ 1 is calculated
  • the three-axis scale coefficients k x , k y , k z , the inter-axis non-orthogonal angles ⁇ xy , ⁇ yz , ⁇ xz and the residual zero offset b x , b y , b z can pass the following formula Calculated:
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • ⁇ 1 is the eigenvector corresponding to the largest eigenvalue of the matrix C 1 -1 (S 1 -S 2 S 4 -1 S 2 T ), so in the actual calculation, all eigenvalues and eigenvectors are not calculated.
  • the power method or the inverse power method to calculate the largest eigenvector, taking the inverse power method as an example:
  • Step S7 Compensating the electronic compass raw data by using the calculated ellipsoid parameters.
  • Step S8 After the new electronic compass data is collected, the newly collected data can be corrected by using the ellipsoid.
  • Step S9 Calculating the posture according to the compensated accelerometer data and the electronic compass data.
  • Step S10 Excluding the calculation point of the jump in the collected electronic compass data by means of statistical rules.
  • Step S11 The electronic compass uses the measurement data to update the ellipsoid parameters in the measurement process, and improves the accuracy of the ellipsoid parameters.
  • the CPU can only make a fitting operation based on a small number of sample points in order to ensure that the memory stack does not overflow during the operation, so the fitting precision is limited;
  • the matrix operation has been improved, and the calculation process is optimized to reduce the calculation of the ellipsoid fitting by more than 90%.
  • the calculation of the fitting algorithm takes 563 ms; relatively, the adoption The improved algorithm takes only 14ms; the strategy of calibrating with measurement data improves the fault tolerance during calibration, which reduces the initial calibration requirements and reduces the operational complexity of the user.

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Manufacturing & Machinery (AREA)
  • Length Measuring Devices With Unspecified Measuring Means (AREA)
  • Navigation (AREA)

Abstract

一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准(S1);利用计算出的椭球参数对加速度计原始数据进行补偿(S2);根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行校准(S3);利用计算出的椭球参数对电子罗盘原始数据进行补偿(S4);根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算(S5),上述方法步骤校准结果可靠、精度高、校准耗时少。

Description

一种姿态测量系统的精度校准方法 技术领域
本发明涉及测量技术领域,尤其涉及一种姿态测量系统的精度校准方法。
背景技术
姿态测量系统是一种可以测量物体空间姿态(俯仰角、横滚角、航向角)的一系列设备,在工业上的许多领域有着广泛的应用,近年来,随着硬件成本的不断下降,各种类型的姿态测量系统开始走进千家万户的日常生活之中,以手机为例,目前大部分智能手机都内置了加速度计、陀螺仪和电子罗盘,构成了一个简易的低成本姿态测量系统。然而,目前大部分姿态测量系统中,由于成本的限制,系统内置传感器的精度和稳定性都不高,从而导致了整个姿态测量系统的测量结果不理想。在不改变系统硬件的情况下对传感器进行校准,无疑是提升整个系统测量精度的一种极为实用的方式。因此,研究低成本姿态测量系统的校准方法具有极强的现实应用价值。
传感器的校准与标定一直是传感器技术领域内的一个重要课题,对于姿态测量系统的校准方案,国内外学者也有大量的相关研究。孙伟、付心如等研究了利用速率转台进行MEMS惯导的多位置标定方法;马斌良等学者提出了一种在有角度基准条件下基于傅里叶变换的电子罗盘校准方案;秦伟等学者研究了基于神经网络和UKF(无迹卡尔曼滤波)的姿态测量系统在线标定技术。
目前的研究大多数集中在利用转台等校准器械对姿态测量系统进行校准。然而,校准专用器械价格昂贵,且其中大部分器械操作复杂,导致生产成本较高;另一方面,在实际应用中,随着周围环境的变化和传感器本身的老化,系统内传感器的特性会不断变化,仅仅依靠出厂前的校准在实际应用中难以起到很好的效果。另外,目前一部分关于非器械校准的研究中,其校准方案要么太过简易,对实际测量精度的提升帮助不大;要么过于复杂,对使用人员的相关技能有着较高的要求。
发明内容
鉴于目前姿态校准存在的上述不足,本发明提供一种姿态测量系 统的精度校准方法,校准结果可靠、精度高、校准耗时少。
为达到上述目的,本发明的实施例采用如下技术方案:
一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
利用计算出的椭球参数对加速度计原始数据进行补偿;
根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
利用计算出的椭球参数对电子罗盘原始数据进行补偿;
根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
依照本发明的一个方面,所述通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤执行前执行以下步骤:对加速度计进行水平校准,消除加速度计原始零偏。
依照本发明的一个方面,所述对加速度计进行水平校准,消除加速度计原始零偏步骤执行后执行以下步骤:采集加速度计在一段时间内的三轴数据,其中,三轴数据标记为[AX0、AY0、AZ0]T,零偏标记为B=[AX0、AY0、AZ0]T-[0 0 g]T。。
依照本发明的一个方面,所述通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000001
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000002
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000003
Figure PCTCN2016088303-appb-000004
Figure PCTCN2016088303-appb-000005
依照本发明的一个方面,所述通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤执行后执行以下步骤:当采集到新的加速度计数据后,可以用椭球参数对新采集到的数据进行修正。
依照本发明的一个方面,所述根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算 防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000006
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000007
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000008
Figure PCTCN2016088303-appb-000009
Figure PCTCN2016088303-appb-000010
依照本发明的一个方面,所述根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤执行后执行以下步骤:当采集到新的电子罗盘数据后,可以用椭球参数对新采集到的数据进行修正。
依照本发明的一个方面,所述根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算步骤执行后执行以下步骤:电子罗盘在测量过程中利用测量数据对椭球参数进行更新,提高椭球参数的准确性。
依照本发明的一个方面,所述电子罗盘在测量过程中利用测量数据对椭球参数进行更新,提高椭球参数的准确性步骤执行前执行以下步骤:通过统计规则的方式剔除采集的电子罗盘数据中出现跳变的计算点。
本发明实施的优点:通过采用了椭球拟合的方式对姿态测量系统中的电子罗盘和加速度进行校准,可有效地避免传感器偶然输出异常、跳数等问题,使得校准结果的可靠性得到了有效的保证;对椭球拟合算法进行了递推化处理,使得姿态测量系统内部CPU在处理数据时不需要存储传感器采集的过往所有点的数据,而仅需要存储之前递推得到的矩阵和当次的测量结果即可,这使得在相同的硬件环境下,该椭球拟合方法可以处理更多的数值,从而提高椭球拟合的精度。相对应的,采用传统拟合算法时,CPU为了确保运算过程中内存栈不会溢出,只能根据少量的样本点做拟合运算,因而拟合精度有限;对椭球拟合算法中大规模的矩阵运算进行了改进,优化了其中的计算流程,使得椭球拟合的计算量减少90%以上,以Matlab程序为例,按照传统算法,运算一次拟合算法耗时563ms;相对地,采用改进后的算法耗时仅为14ms;采用了利用测量数据进行校准的策略,提高了校准时的容错能力,从而降低了初始校准时的要求,减少了使用人员的操作复杂程度。
附图说明
为了更清楚地说明本发明实施例中的技术方案,下面将对实施例 中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1为本发明所述的一种姿态测量系统的精度校准方法的实施例1的方法流程图;
图2为本发明所述的一种姿态测量系统的精度校准方法的实施例2的方法流程图;
图3为本发明所述的一种姿态测量系统的精度校准方法的实施例3的方法流程图;
图4为本发明所述的一种姿态测量系统的精度校准方法的实施例4的方法流程图;
图5为本发明所述的一种姿态测量系统的精度校准方法的实施例5的方法流程图;
图6为本发明所述的一种姿态测量系统的精度校准方法的实施例6的方法流程图;
图7为本发明所述的一种姿态测量系统的精度校准方法的实施例7的方法流程图。
具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本发明一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。
实施例1:
如图1所示,一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
步骤S1:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
所述步骤S1:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000011
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000012
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000013
Figure PCTCN2016088303-appb-000014
Figure PCTCN2016088303-appb-000015
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S2:利用计算出的椭球参数对加速度计原始数据进行补偿。
步骤S3:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
所述步骤S3:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000016
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000017
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000018
Figure PCTCN2016088303-appb-000019
Figure PCTCN2016088303-appb-000020
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S4:利用计算出的椭球参数对电子罗盘原始数据进行补偿。
步骤S5:根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
实施例2:
如图2所示,一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
步骤S1:对加速度计进行水平校准,消除加速度计原始零偏;
将姿态测量设备通过水平气泡调整到水平位置,对加速度计进行水平校准,去除加速度计的零偏。
步骤S2:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
所述步骤S2:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为 D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000021
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000022
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000023
Figure PCTCN2016088303-appb-000024
Figure PCTCN2016088303-appb-000025
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S3:利用计算出的椭球参数对加速度计原始数据进行补偿;
步骤S4:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
所述步骤S4:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000026
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000027
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000028
Figure PCTCN2016088303-appb-000029
Figure PCTCN2016088303-appb-000030
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至 不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S5:利用计算出的椭球参数对电子罗盘原始数据进行补偿;
步骤S6:根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
实施例3:
如图3所示,一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
步骤S1:对加速度计进行水平校准,消除加速度计原始零偏;
将姿态测量设备通过水平气泡调整到水平位置,对加速度计进行水平校准,去除加速度计的零偏。
步骤S2:采集加速度计在一段时间内的三轴数据,其中,三轴数据标记为[AX0、AY0、AZ0]T,零偏标记为B=[AX0、AY0、AZ0]T-[0 0 g]T
步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
所述步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为 D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000031
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000032
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000033
Figure PCTCN2016088303-appb-000034
Figure PCTCN2016088303-appb-000035
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S4:利用计算出的椭球参数对加速度计原始数据进行补偿;
步骤S5:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
所述步骤S5:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000036
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000037
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000038
Figure PCTCN2016088303-appb-000039
Figure PCTCN2016088303-appb-000040
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至 不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S6:利用计算出的椭球参数对电子罗盘原始数据进行补偿;
步骤S7:根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
实施例4:
如图4所示,一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
步骤S1:对加速度计进行水平校准,消除加速度计原始零偏;
将姿态测量设备通过水平气泡调整到水平位置,对加速度计进行水平校准,去除加速度计的零偏。
步骤S2:采集加速度计在一段时间内的三轴数据,其中,三轴数据标记为[AX0、AY0、AZ0]T,零偏标记为B=[AX0、AY0、AZ0]T-[0 0 g]T
步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
所述步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、 AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000041
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000042
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000043
Figure PCTCN2016088303-appb-000044
Figure PCTCN2016088303-appb-000045
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S4:当采集到新的加速度计数据后,可以用椭球参数对新采集到的数据进行修正。
步骤S5:利用计算出的椭球参数对加速度计原始数据进行补偿。
步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
所述步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T 拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy 12 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000046
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000047
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000048
Figure PCTCN2016088303-appb-000049
Figure PCTCN2016088303-appb-000050
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S7:利用计算出的椭球参数对电子罗盘原始数据进行补偿;
步骤S8:根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
实施例5:
如图5所示,一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
步骤S1:对加速度计进行水平校准,消除加速度计原始零偏;
将姿态测量设备通过水平气泡调整到水平位置,对加速度计进行水平校准,去除加速度计的零偏。
步骤S2:采集加速度计在一段时间内的三轴数据,其中,三轴数据标记为[AX0、AY0、AZ0]T,零偏标记为B=[AX0、AY0、AZ0]T-[0 0 g]T
步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻 度系数与轴间不正交角进行校准;
所述步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000051
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000052
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000053
Figure PCTCN2016088303-appb-000054
Figure PCTCN2016088303-appb-000055
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S4:当采集到新的加速度计数据后,可以用椭球参数对新采集到的数据进行修正。
步骤S5:利用计算出的椭球参数对加速度计原始数据进行补偿。
步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
所述步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000056
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000057
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000058
Figure PCTCN2016088303-appb-000059
Figure PCTCN2016088303-appb-000060
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S7:利用计算出的椭球参数对电子罗盘原始数据进行补偿。
步骤S8:当采集到新的电子罗盘数据后,可以用椭球参数对新采集到的数据进行修正。
步骤S9:根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
实施例6:
如图6所示,一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
步骤S1:对加速度计进行水平校准,消除加速度计原始零偏;
将姿态测量设备通过水平气泡调整到水平位置,对加速度计进行水平校准,去除加速度计的零偏。
步骤S2:采集加速度计在一段时间内的三轴数据,其中,三轴数据标记为[AX0、AY0、AZ0]T,零偏标记为B=[AX0、AY0、AZ0]T-[0 0 g]T
步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
所述步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000061
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000062
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000063
Figure PCTCN2016088303-appb-000064
Figure PCTCN2016088303-appb-000065
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S4:当采集到新的加速度计数据后,可以用椭球参数对新采集到的数据进行修正。
步骤S5:利用计算出的椭球参数对加速度计原始数据进行补偿。
步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
所述步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000066
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000067
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000068
Figure PCTCN2016088303-appb-000069
Figure PCTCN2016088303-appb-000070
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1,继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S7:利用计算出的椭球参数对电子罗盘原始数据进行补偿。
步骤S8:当采集到新的电子罗盘数据后,可以用椭球参数对新采集到的数据进行修正。
步骤S9:根据补偿后的加速度计数据与电子罗盘数据对姿态进行解 算。
步骤S10:电子罗盘在测量过程中利用测量数据对椭球参数进行更新,提高椭球参数的准确性。
实施例7:
如图7所示,一种姿态测量系统的精度校准方法,所述姿态测量系统的精度校准方法包括如下步骤:
步骤S1:对加速度计进行水平校准,消除加速度计原始零偏;
将姿态测量设备通过水平气泡调整到水平位置,对加速度计进行水平校准,去除加速度计的零偏。
步骤S2:采集加速度计在一段时间内的三轴数据,其中,三轴数据标记为[AX0、AY0、AZ0]T,零偏标记为B=[AX0、AY0、AZ0]T-[0 0 g]T
步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
所述步骤S3:通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000071
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000072
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000073
Figure PCTCN2016088303-appb-000074
Figure PCTCN2016088303-appb-000075
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1, 继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S4:当采集到新的加速度计数据后,可以用椭球拟合参数对新采集到的数据进行修正。
步骤S5:利用计算出的椭球参数对加速度计原始数据进行补偿。
步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
所述步骤S6:根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
Figure PCTCN2016088303-appb-000076
将矩阵S分块:
Dk=[D1 D2]
D1=[Ax1 2 Ay1 2 Az1 2]
D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
Figure PCTCN2016088303-appb-000077
通过以下公式求得椭球拟合向量:
C1 -1(S1-S2S4 -1S2 T1=λα1
α2=-S4 -1S2 Tα1
计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxy,δyz,δxz及残余零偏bx,by,bz可以通过以下公式计算得出:
Figure PCTCN2016088303-appb-000078
Figure PCTCN2016088303-appb-000079
Figure PCTCN2016088303-appb-000080
上述公式中,α1为矩阵C1 -1(S1-S2S4 -1S2 T)最大特征值对应的特征向量,因此在实际计算中,不用将所有特征值和特征向量全部计算出来,甚至不用计算特征值,而只要利用幂法或者反幂法进行计算出最大的特征向量即可,以反幂法为例:
令A=(C1 -1(S1-S2S4 -1S2 T))-1,取一个初始迭代向量u=[1 1 1]T,设置迭代精度ε=1e-6
对A做LU分解,即A=LU;
解线性方程组:
Ly(k)=u(k-1),Uv(k)=y(k)
mk=max(v(k)),u(k)=v(k)/mk
若|mk-mk-1|<ε,则跳出迭代,停止计算,此时α1=u(k),否则k=k+1, 继续计算。
一般而言,对于ε=1e-6,5次以下的迭代计算即可达到此精度。
步骤S7:利用计算出的椭球参数对电子罗盘原始数据进行补偿。
步骤S8:当采集到新的电子罗盘数据后,可以用椭球拟数对新采集到的数据进行修正。
步骤S9:根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
步骤S10:通过统计规则的方式剔除采集的电子罗盘数据中出现跳变的计算点。
步骤S11:电子罗盘在测量过程中利用测量数据对椭球参数进行更新,提高椭球参数的准确性。
本发明实施的优点:通过采用了椭球拟合的方式对姿态测量系统中的电子罗盘和加速度进行校准,可有效地避免传感器偶然输出异常、跳数等问题,使得校准结果的可靠性得到了有效的保证;对椭球拟合算法进行了递推化处理,使得姿态测量系统内部CPU在处理数据时不需要存储传感器采集的过往所有点的数据,而仅需要存储之前递推得到的矩阵和当次的测量结果即可,这使得在相同的硬件环境下,该椭球拟合方法可以处理更多的数值,从而提高椭球拟合的精度。相对应的,采用传统拟合算法时,CPU为了确保运算过程中内存栈不会溢出,只能根据少量的样本点做拟合运算,因而拟合精度有限;对椭球拟合算法中大规模的矩阵运算进行了改进,优化了其中的计算流程,使得椭球拟合的计算量减少90%以上,以Matlab程序为例,按照传统算法,运算一次拟合算法耗时563ms;相对地,采用改进后的算法耗时仅为14ms;采用了利用测量数据进行校准的策略,提高了校准时的容错能力,从而降低了初始校准时的要求,减少了使用人员的操作复杂程度。
以上所述,仅为本发明的具体实施方式,但本发明的保护范围并不局限于此,任何熟悉本领域技术的技术人员在本发明公开的技术范围内,可轻易想到的变化或替换,都应涵盖在本发明的保护范围之内。因此,本发明的保护范围应以所述权利要求的保护范围为准。

Claims (9)

  1. 一种姿态测量系统的精度校准方法,其特征在于,所述姿态测量系统的精度校准方法包括如下步骤:
    通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准;
    利用计算出的椭球参数对加速度计原始数据进行补偿;
    根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准;
    利用计算出的椭球参数对电子罗盘原始数据进行补偿;
    根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算。
  2. 根据权利要求1所述的姿态测量系统的精度校准方法,其特征在于,所述通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤执行前执行以下步骤:对加速度计进行水平校准,消除加速度计原始零偏。
  3. 根据权利要求2所述的姿态测量系统的精度校准方法,其特征在于,所述对加速度计进行水平校准,消除加速度计原始零偏步骤执行后执行以下步骤:采集加速度计在一段时间内的三轴数据,其中,三轴数据标记为[AX0、AY0、AZ0]T,零偏标记为B=[AX0、AY0、AZ0]T–[0 0 g]T
  4. 根据权利要求3所述的姿态测量系统的精度校准方法,其特征在于,所述通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤包括以下步骤:
    设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
    设定校准过程迭代次数为n,将加速度采集的第一个值[AX0、AY0、AZ0]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Ax1 2 Ay1 2 Az1 2 Ax1Ay1 Ax1Az1 Ay1Az1 Ax1 Ay1 Az1 1]T,S1=D1TD1;
    若加速度计三轴数据的采集次数未达到n次,则用新采集的加速度计数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
    计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
    Figure PCTCN2016088303-appb-100001
    Figure PCTCN2016088303-appb-100002
    将矩阵S分块:
    Dk=[D1 D2]
    D1=[Ax1 2 Ay1 2 Az1 2]
    D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
    Figure PCTCN2016088303-appb-100003
    通过以下公式求得椭球拟合向量:
    Figure PCTCN2016088303-appb-100004
    计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
    再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxyyzxz及残余零偏bx,by,bz可以通过以下公式计算得出:
    Figure PCTCN2016088303-appb-100005
    Figure PCTCN2016088303-appb-100006
    Figure PCTCN2016088303-appb-100007
  5. 根据权利要求4所述的姿态测量系统的精度校准方法,其特征在于,所述通过椭球拟合模型对姿态测量系统的加速度计的零偏、刻度系数与轴间不正交角进行校准步骤执行后执行以下步骤:当采集到新的加速度计数据后,可以用椭球参数对新采集到的数据进行修正。
  6. 根据权利要求5所述的姿态测量系统的精度校准方法,其特征在于,所述根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤包括如下步骤:
    设定拟合椭球参数向量,标记为α=[a b c d e f k l m n]T
    采集电子罗盘在一段时间内的三轴数据,并对采集的数据进行运算防溢处理,其中,三轴数据标记为[HX0 HY0 HZ0]T
    设定校准过程迭代次数为n,将加速度采集的第一个值[HX1 HY1 HZ1]T拓展为10维列向量D1,并计算矩阵S1,其中,D1标记为D1=[Hx1 2 Hy1 2 Hz1 2 Hx1Hy1 Hx1Hz1 Hy1Hz1 Hx1 Hy1 Hz1 1]T,S1=D1TD1;
    若电子罗盘三轴数据的采集次数未达到n次,则用新采集的电子罗盘数据DK对矩阵S进行更新,其中,Sk=Sk-1+Dk TDk
    计算出矩阵S后,同时引入椭球限制矩阵C并将矩阵S分块,其中,
    Figure PCTCN2016088303-appb-100008
    Figure PCTCN2016088303-appb-100009
    将矩阵S分块:
    Dk=[D1 D2]
    D1=[Ax1 2 Ay1 2 Az1 2]
    D2=[Ax1Ay1 Ay1Az1 Ax1Az1 Ax1 Ay1 Az1 1]
    Figure PCTCN2016088303-appb-100010
    通过以下公式求得椭球拟合向量:
    计算出α1后,通过上述公式计算出α2,则椭球拟合向量的所有参数全部被计算出,其中,α=[α1 α2]=[a b c d e f k l m n]T
    再根据实际的物理对应关系,三轴的刻度系数kx,ky,kz、轴间不正交角δxyyzxz及残余零偏bx,by,bz可以通过以下公式计算得出:
    Figure PCTCN2016088303-appb-100012
    Figure PCTCN2016088303-appb-100013
    Figure PCTCN2016088303-appb-100014
  7. 根据权利要求6所述的姿态测量系统的精度校准方法,其特征在于,所述根据补偿后的加速度计数据通过椭球拟合模型对电子罗盘进行准步骤执行后执行以下步骤:当采集到新的电子罗盘数据后,可以用椭球参数对新采集到的数据进行修正。
  8. 根据权利要求7所述的姿态测量系统的精度校准方法,其特征在于,所述根据补偿后的加速度计数据与电子罗盘数据对姿态进行解算步骤执行后执行以下步骤:电子罗盘在测量过程中利用测量数据对椭球参数进行更新,提高椭球参数的准确性。
  9. 根据权利要求8所述的姿态测量系统的精度校准方法,其特征在于,所述电子罗盘在测量过程中利用测量数据对椭球参数进行更新,提高椭球参数的准确性步骤执行前执行以下步骤:通过统计规则的方式剔除采集的电子罗盘数据中出现跳变的计算点。
PCT/CN2016/088303 2015-10-13 2016-07-04 一种姿态测量系统的精度校准方法 WO2017063386A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
RU2017125953A RU2662458C1 (ru) 2015-10-13 2016-07-04 Способ прецизионной калибровки системы измерения пространственного положения
KR1020177023778A KR102008597B1 (ko) 2015-10-13 2016-07-04 자세 측정 시스템의 정밀 보정 방법
US15/542,931 US10605619B2 (en) 2015-10-13 2016-07-04 Precision calibration method of attitude measuring system
EP16854769.3A EP3364151B1 (en) 2015-10-13 2016-07-04 Precision calibration method for attitude measurement system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510665021.X 2015-10-13
CN201510665021.XA CN105352487B (zh) 2015-10-13 2015-10-13 一种姿态测量系统的精度校准方法

Publications (1)

Publication Number Publication Date
WO2017063386A1 true WO2017063386A1 (zh) 2017-04-20

Family

ID=55328497

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/088303 WO2017063386A1 (zh) 2015-10-13 2016-07-04 一种姿态测量系统的精度校准方法

Country Status (6)

Country Link
US (1) US10605619B2 (zh)
EP (1) EP3364151B1 (zh)
KR (1) KR102008597B1 (zh)
CN (1) CN105352487B (zh)
RU (1) RU2662458C1 (zh)
WO (1) WO2017063386A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102017220867A1 (de) * 2017-11-22 2019-05-23 Robert Bosch Gmbh Verfahren zum automatischen Kalibrieren und Verfahren zum Verwenden eines Beschleunigungssensors
CN114264997A (zh) * 2021-12-14 2022-04-01 武汉联影生命科学仪器有限公司 梯度灵敏度校准方法、装置及磁共振设备
CN116224468A (zh) * 2023-05-10 2023-06-06 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) 一种陆地重力仪标定方法及标定系数

Families Citing this family (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105352487B (zh) 2015-10-13 2018-06-15 上海华测导航技术股份有限公司 一种姿态测量系统的精度校准方法
CN106556383B (zh) * 2016-12-02 2019-05-07 上海华测导航技术股份有限公司 一种rtk倾斜补偿测量精度验证的方法
CN106979775B (zh) * 2017-04-28 2019-07-30 江苏号百信息服务有限公司 一种电子罗盘阶跃时平滑滤波的方法
CN107479079A (zh) * 2017-07-13 2017-12-15 临沂大学 一种基于pdr与led结合的方法
CN107656227B (zh) * 2017-09-21 2019-10-11 大连理工大学 基于Levenberg-Marquardt算法的磁力计校准方法
US20190346897A1 (en) * 2018-05-13 2019-11-14 Sean Joseph Rostami Introspective Power Method
CN109188422B (zh) * 2018-08-08 2023-01-06 中国航空工业集团公司雷华电子技术研究所 一种基于lu分解的卡尔曼滤波目标跟踪方法
CN109470277B (zh) * 2018-12-26 2022-09-13 湖南航天机电设备与特种材料研究所 非正交角度测量装置标定系数的测定方法及系统
CN112146678B (zh) * 2019-06-27 2022-10-11 华为技术有限公司 一种确定校准参数的方法及电子设备
CN111765879A (zh) * 2019-11-29 2020-10-13 深圳市瑞芬科技有限公司 一种三维电子罗盘装置及实用校准方法
CN112325901B (zh) * 2020-09-28 2022-09-16 中国船舶重工集团公司第七0七研究所 一种平台式惯导系泊状态下计算方位陀螺仪标度的方法
CN112284366B (zh) * 2020-10-26 2022-04-12 中北大学 一种基于tg-lstm神经网络的偏振光罗盘航向角误差校正方法
CN112284421B (zh) * 2020-11-02 2021-03-19 蘑菇车联信息科技有限公司 一种imu内参调整方法及相关装置
CN112596015A (zh) * 2020-12-28 2021-04-02 上海矽睿科技有限公司 三轴磁传感器的测试方法及系统
CN113124905B (zh) * 2021-04-27 2022-10-28 西安电子科技大学 一种用于多轴惯性姿态传感器精度评估的自动测量方法
CN113985494A (zh) * 2021-10-13 2022-01-28 哈尔滨工程大学 一种基于无迹卡尔曼算法海底地震计中电子罗盘误差补偿方法
CN114485728A (zh) * 2022-01-04 2022-05-13 中国煤炭科工集团太原研究院有限公司 一种捷联惯导系统现场快速自标定方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102252689A (zh) * 2010-05-19 2011-11-23 北京国浩传感器技术研究院(普通合伙) 一种基于磁传感器的电子罗盘校准方法
US20120203487A1 (en) * 2011-01-06 2012-08-09 The University Of Utah Systems, methods, and apparatus for calibration of and three-dimensional tracking of intermittent motion with an inertial measurement unit
CN103776451A (zh) * 2014-03-04 2014-05-07 哈尔滨工业大学 一种基于mems的高精度三维姿态惯性测量系统以及测量方法
CN105352487A (zh) * 2015-10-13 2016-02-24 上海华测导航技术股份有限公司 一种姿态测量系统的精度校准方法

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0625671B2 (ja) * 1987-08-28 1994-04-06 日本航空電子工業株式会社 慣性航法装置
RU2098764C1 (ru) * 1996-05-29 1997-12-10 Русланов Александр Семенович Способ определения местоположения подвижных объектов и устройство для его реализации
CA2253052A1 (en) * 1997-01-31 1998-10-29 Greenfield Enterprises, Inc. Navigation system and method
US7376507B1 (en) * 2004-05-27 2008-05-20 Sandia Corporation Geophysics-based method of locating a stationary earth object
US7275008B2 (en) * 2005-09-02 2007-09-25 Nokia Corporation Calibration of 3D field sensors
US7451549B1 (en) * 2006-08-09 2008-11-18 Pni Corporation Automatic calibration of a three-axis magnetic compass
US8005635B2 (en) * 2007-08-14 2011-08-23 Ching-Fang Lin Self-calibrated azimuth and attitude accuracy enhancing method and system (SAAAEMS)
DE102008042989A1 (de) * 2008-10-21 2010-04-22 Robert Bosch Gmbh Elektronischer Kompass
US8645093B2 (en) * 2009-11-04 2014-02-04 Qualcomm Incorporated Calibrating multi-dimensional sensor for offset, sensitivity, and non-orthogonality
CN103153790B (zh) * 2010-10-01 2016-06-08 希尔克瑞斯特实验室公司 使用运动传感器和附接至装置的磁力计的测量数据估计该装置在重力参照系中的偏航角的设备和方法
KR101209571B1 (ko) * 2010-10-20 2012-12-07 한국과학기술연구원 자동 교정 방법 및 장치
US9541393B2 (en) * 2011-06-30 2017-01-10 Qualcomm Incorporated Reducing power consumption or error of digital compass
US9683865B2 (en) * 2012-01-26 2017-06-20 Invensense, Inc. In-use automatic calibration methodology for sensors in mobile devices
KR101503046B1 (ko) * 2013-08-22 2015-03-24 한국과학기술연구원 다축 감지 장치 및 이의 교정 방법
CN104898681B (zh) * 2015-05-04 2017-07-28 浙江工业大学 一种采用三阶近似毕卡四元数的四旋翼飞行器姿态获取方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102252689A (zh) * 2010-05-19 2011-11-23 北京国浩传感器技术研究院(普通合伙) 一种基于磁传感器的电子罗盘校准方法
US20120203487A1 (en) * 2011-01-06 2012-08-09 The University Of Utah Systems, methods, and apparatus for calibration of and three-dimensional tracking of intermittent motion with an inertial measurement unit
CN103776451A (zh) * 2014-03-04 2014-05-07 哈尔滨工业大学 一种基于mems的高精度三维姿态惯性测量系统以及测量方法
CN105352487A (zh) * 2015-10-13 2016-02-24 上海华测导航技术股份有限公司 一种姿态测量系统的精度校准方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
LIU, YANXIA ET AL.: "Three-Axis Accelerometer Error Calibration and Compensation based on Ellisoid Hypothesis", TRANSDUCER AND MICROSYSTEM TECHNOLOGIES, vol. 33, no. 6, 31 December 2014 (2014-12-31), XP009509388, ISSN: 2096-2436 *
ZHANG, WEI ET AL.: "Automatic Error Compensation Method of Three-Axis Electronic Compass Based on the Fluxgate", CHINESE JOURNAL OF SENSORS AND ACTUATORS, vol. 25, no. 12, 31 December 2012 (2012-12-31), XP055375420, ISSN: 1004-1699 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102017220867A1 (de) * 2017-11-22 2019-05-23 Robert Bosch Gmbh Verfahren zum automatischen Kalibrieren und Verfahren zum Verwenden eines Beschleunigungssensors
CN114264997A (zh) * 2021-12-14 2022-04-01 武汉联影生命科学仪器有限公司 梯度灵敏度校准方法、装置及磁共振设备
CN114264997B (zh) * 2021-12-14 2024-03-22 武汉联影生命科学仪器有限公司 梯度灵敏度校准方法、装置及磁共振设备
CN116224468A (zh) * 2023-05-10 2023-06-06 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) 一种陆地重力仪标定方法及标定系数
CN116224468B (zh) * 2023-05-10 2023-08-22 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) 一种陆地重力仪标定方法及标定系数

Also Published As

Publication number Publication date
RU2662458C1 (ru) 2018-07-26
EP3364151B1 (en) 2020-11-25
EP3364151A4 (en) 2019-06-19
KR20170105619A (ko) 2017-09-19
US10605619B2 (en) 2020-03-31
CN105352487A (zh) 2016-02-24
EP3364151A1 (en) 2018-08-22
US20180010923A1 (en) 2018-01-11
KR102008597B1 (ko) 2019-08-07
CN105352487B (zh) 2018-06-15

Similar Documents

Publication Publication Date Title
WO2017063386A1 (zh) 一种姿态测量系统的精度校准方法
CN107024674B (zh) 一种基于递推最小二乘法的磁强计现场快速标定方法
CN108426571B (zh) 一种电子罗盘本地实时校准方法及装置
CN110146839A (zh) 一种移动平台磁梯度张量系统校正方法
CN110361683A (zh) 基于双目标函数粒子群优化的磁力计校正方法
CN107870001A (zh) 一种基于椭球拟合的磁力计校正方法
WO2022160391A1 (zh) 磁力计信息辅助的mems陀螺仪标定方法及标定系统
KR101698682B1 (ko) 지자기 센서의 출력값을 보정하는 방법 및 장치
CN111189474A (zh) 基于mems的marg传感器的自主校准方法
CN108645404A (zh) 一种小型多旋翼无人机姿态解算方法
US20110060543A1 (en) Method for self-adjustment of a triaxial acceleration sensor during operation, and sensor system having a three -dimentional acceleration sensor
CN112525144B (zh) 一种非线性姿态检测补偿方法及终端
CN108088431B (zh) 一种自校正电子罗盘及其校正方法
CN111982155B (zh) 磁传感器的标定方法、装置、电子设备和计算机存储介质
CN106092140B (zh) 一种陀螺仪零偏估计方法
Chen et al. High-precision geomagnetic directional technology based on sensor error correction and adaptive hybrid filter
Liu et al. Two-step calibration method for three-axis magnetic sensor error based on particle swarm optimization
CN110160530B (zh) 一种基于四元数的航天器姿态滤波方法
CN112632454A (zh) 一种基于自适应卡尔曼滤波算法的mems陀螺滤波方法
CN116753987A (zh) 一种三轴地磁传感器误差标定方法
CN110672127A (zh) 阵列式mems磁传感器实时标定方法
CN114089244B (zh) 一种捷联三轴磁强计两步标定方法
CN109737940A (zh) 一种电子罗盘实时自由校准方法及装置
CN112683265B (zh) 一种基于快速iss集员滤波的mimu/gps组合导航方法
CN109211271B (zh) 一种磁罗盘自校正方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16854769

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 15542931

Country of ref document: US

REEP Request for entry into the european phase

Ref document number: 2016854769

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2017125953

Country of ref document: RU

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 20177023778

Country of ref document: KR

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE