Method and apparatus for data fusion of a three-axis magnetometer and three axis accelerometer

• US 9,915,550 B2
• Filed: 08/01/2013
• Issued: 03/13/2018
• Est. Priority Date: 08/02/2012
• Status: Active Grant

• Current Assignee: Memsic Semiconductor (Wuxi) Co., Ltd. (MZ Investment Holdings Ltd.)
• Original Assignee: Memsic Incorporated (MZ Investment Holdings Ltd.)
• Inventors: An, Dong, Zhao, Yang, Ding, Haipeng, Lee, Jong-Ki
• Primary Examiner(s): Henson, Mischita
• Application Number:
  US14/418,761
• Publication Number:
  US 20150177020A1
• Time in Patent Office: 1,685 Days
• CPC Class Codes: G01C 17/38 Testing, calibrating, or compensating of compasses
• CPC Class Codes: G01C 25/00 Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass testing, calibrating or compensating compasses G01C17/38
• CPC Class Codes: G01P 15/00 Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration
• CPC Class Codes: G01P 21/00 Testing or calibrating of apparatus or devices covered by the preceding groups
• CPC Class Codes: G01R 35/00 Testing or calibrating of apparatus covered by the other groups of this subclass
• CPC Class Codes: G06F 18/256 of results relating to different input data, e.g. multimodal recognition

Abstract

Method and apparatus of integrating a three-axis magnetometer and a three-axis accelerometer to provide attitude and heading, calibrated magnetometer and accelerometer data, and angular rate, while removing sensor error sources over time and temperature and to compensate for hard and soft iron distortions of the Earth magnetic field. Filtered accelerometer data are corrected to account for various error sources. The magnetic heading is calculated from a horizontal magnetic field vector transformed from three dimensional Earth'"'"'s magnetic field vector by using quasi-static roll and pitch angles from the filtered accelerometer data. A first Kalman filter estimates the state vector, based on the principle that the magnitude of local Earth'"'"'s magnetic field vector is constant, to form hard and soft iron correction matrices. A second Kalman filter estimates a correction matrix of coupled remaining soft iron and the misalignment of the magnetometer and the accelerometer, as the dot product of local Earth'"'"'s magnetic field vector and a corrected gravitational acceleration vector, at a quasi-static position, is constant. The three dimensional Earth'"'"'s magnetic field vector is received by removing the hard and soft irons through the soft iron and hard iron correction matrixes.

Claims

1. A computer-implemented method of fusing data from a three-axis magnetometer and a three-axis accelerometer to provide a magnetic heading of a device, the method comprising:

receiving, as input, unprocessed magnetic field measurements from the magnetometer associated with the device and unprocessed acceleration measurements from the accelerometer associated with the device;

filtering the unprocessed magnetic field measurements to remove outlying data points and noise and obtaining filtered magnetic field data;

applying a first Kalman filter to the filtered magnetic field data to generate first soft iron and hard iron matrices, wherein the first Kalman filter is based on the principle that the magnitude of local Earth'"'"'s magnetic field vector is constant;

applying a second Kalman filter to the filtered magnetic field data to generate a coupled soft iron/misalignment correction matrix, wherein the second Kalman filter is based on the principle that an angle between the local Earth'"'"'s magnetic field vector and the gravitational acceleration vector is constant;

multiplying the coupled soft iron/misalignment correction matrix by the first soft iron matrix to generate a soft-iron correction matrix;

filtering the unprocessed acceleration measurements to remove outlying data points and noise and obtaining filtered acceleration measurements;

correcting the filtered acceleration measurements as a function of one or more error sources and obtaining corrected acceleration measurements;

correcting the filtered magnetic field data for hard-iron and soft-iron distortions and obtaining corrected magnetic field measurements;

computing quasi-static roll and pitch angles of the device as a function of the corrected acceleration measurements;

forming a transformation matrix from a body frame to a level coordinate frame as a function of the computed quasi-static roll and pitch angles;

transforming the corrected magnetic field measurements from the body frame to the level frame and generating a horizontal magnetic field vector;

determining a magnetic heading of the device as a function of the generated horizontal magnetic field vector and the generated soft-iron correction matrix;

providing, to an application operating on the device, the magnetic heading of the device;

deriving, by the application, angular rates of the device; and

performing auto-calibration of at least one of;

the magnetometer and the accelerometer, based on the corrected acceleration measurements and the derived angular rates.

2. The computer-implemented method of claim 1, further comprising:

determining when the device is in a quasi-static condition,

wherein applying the second Kalman filter to the filtered magnetic field data is performed in response to determining that the device is in a quasi-static condition.

3. The computer-implemented method of claim 1, further comprising:

determining if a magnetic field disturbance is present.

4. The computer-implemented method of claim 1, further comprising:

determining if the magnetic field is stable.

5. The computer-implemented method of claim 1, wherein performing auto-calibration of the accelerometer comprises estimating the one or more error sources.

6. The computer-implemented method of claim 1, wherein providing, as output, the magnetic heading of the device comprises providing quasi-static roll, pitch, and magnetic heading information as part of an e-compass application.

7. The computer-implemented method of claim 1, wherein providing, as output, the magnetic heading of the device comprises providing a pseudo nine-axis solution as part of a soft gyroscope function.

8. The computer-implemented method of claim 7, wherein the pseudo nine-axis solution comprises at least one of:

dynamic roll angles;

dynamic pitch angles;

dynamic heading angles;

dynamic angular rates; and

acceleration determinations.

9. A computer-implemented method of fusing data from a three-axis magnetometer and a three-axis accelerometer to provide a magnetic heading of a device, the method comprising:

receiving unprocessed magnetic field measurements from the magnetometer associated with the device;

receiving unprocessed acceleration measurements from the accelerometer associated with the device;

filtering the unprocessed magnetic field measurements to remove outlying data points and noise and obtaining filtered magnetic field data;

filtering the unprocessed acceleration measurements to remove outlying data points and noise and obtaining filtered acceleration measurements;

correcting the filtered acceleration measurements as a function of one or more error sources and obtaining corrected acceleration measurements;

correcting the filtered magnetic field data for hard-iron and soft-iron distortions and obtaining corrected magnetic field measurements, wherein correcting the filtered magnetic field data for hard-iron and soft-iron distortions comprises applying a first Kalman filter to the filtered magnetic field data and generating first soft iron and hard iron matrices, wherein the first Kalman filter is based on the principle that the magnitude of local Earth'"'"'s magnetic field vector is constant;

computing quasi-static roll and pitch angles of the device as a function of the corrected acceleration measurements;

forming a transformation matrix from a body frame to a level coordinate frame as a function of the computed quasi-static roll and pitch angles;

transforming the corrected magnetic field measurements from the body frame to the level frame and generating a horizontal magnetic field vector; and

determining a magnetic heading of the device as a function of the generated horizontal magnetic field vector;

providing to an application operating on the device the magnetic heading of the device,

deriving, by the application, angular rates of the device, and

performing auto-calibration of at least one of;

the magnetometer and the accelerometer, based on the corrected acceleration measurements and the derived angular rates

the method further comprising;

determining when the device is in a quasi-static condition and, if so determined, then;

applying a second Kalman filter to the filtered magnetic field data and generating a coupled soft iron/misalignment correction matrix, wherein the second Kalman filter is based on the principle that an angle between the local Earth'"'"'s magnetic field vector and the gravitational acceleration vector is constant; and

multiplying the coupled soft iron/misalignment correction matrix by the first soft iron matrix and generating a soft-iron correction matrix,

wherein the magnetic heading of the device is determined as a function of the generated soft-iron correction matrix.

10. A computer system including a storage device storing computer instructions that, when executed by a computer, implements a method of fusing data from a three-axis magnetometer and a three-axis accelerometer to provide a magnetic heading of a corresponding device, the method comprising:

receiving, as input, unprocessed magnetic field measurements from the magnetometer associated with the device and unprocessed acceleration measurements from the accelerometer associated with the device;

filtering the unprocessed magnetic field measurements to remove outlying data points and noise and obtaining filtered magnetic field data;

applying a first Kalman filter to the filtered magnetic field data to generate first soft iron and hard iron matrices, wherein the first Kalman filter is based on the principle that the magnitude of local Earth'"'"'s magnetic field vector is constant;

applying a second Kalman filter to the filtered magnetic field data to generate a coupled soft iron/misalignment correction matrix, wherein the second Kalman filter is based on the principle that an angle between the local Earth'"'"'s magnetic field vector and the gravitational acceleration vector is constant;

multiplying the coupled soft iron/misalignment correction matrix by the first soft iron matrix to generate a soft-iron correction matrix;

filtering the unprocessed acceleration measurements to remove outlying data points and noise and obtaining filtered acceleration measurements;

correcting the filtered acceleration measurements as a function of one or more error sources and obtaining corrected acceleration measurements;

correcting the filtered magnetic field data for hard-iron and soft-iron distortions and obtaining corrected magnetic field measurements;

computing quasi-static roll and pitch angles of the device as a function of the corrected acceleration measurements;

forming a transformation matrix from a body frame to a level coordinate frame as a function of the computed quasi-static roll and pitch angles;

transforming the corrected magnetic field measurements from the body frame to the level frame and generating a horizontal magnetic field vector;

determining a magnetic heading of the device as a function of the generated horizontal magnetic field vector and the generated soft-iron correction matrix; and

providing, to an application operating on the device, the magnetic heading of the device;

deriving, by the application, angular rates of the device; and

performing auto-calibration of at least one of;

the magnetometer and the accelerometer, based on the corrected acceleration measurements and the derived angular rates.

11. The computer system of claim 10, wherein the method further comprises:

determining when the device is in a quasi-static condition,

wherein applying the second Kalman filter to the filtered magnetic field data is performed in response to determining that the device is in a quasi-static condition.

12. The computer system of claim 10, wherein the method further comprises:

determining if a magnetic field disturbance is present.

13. The computer system of claim 10, wherein the method further comprises:

determining if the magnetic field is stable.

14. The computer system of claim 10, wherein performing auto-calibration of the accelerometer comprises estimating the one or more error sources.

15. The computer system of claim 10, wherein providing, as output, the magnetic heading of the device comprises providing quasi-static roll, pitch, and magnetic heading information as part of an e-compass application.

16. The computer system of claim 10, wherein providing, as output, the magnetic heading of the device comprises providing a pseudo nine-axis solution as part of a soft gyroscope function.

17. The computer system of claim 16, wherein the pseudo nine-axis solution comprises at least one of:

dynamic roll angles;

dynamic pitch angles;

dynamic heading angles;

dynamic angular rates; and

acceleration determinations.