Always on compass calibration system and methods

• US 10,197,396 B2
• Filed: 01/26/2017
• Issued: 02/05/2019
• Est. Priority Date: 01/27/2016
• Status: Active Grant

• Current Assignee: Apple Inc.
• Original Assignee: Apple Inc.
• Inventors: Howell, Adam, Pham, Hung A., Tu, Xiaoyuan
• Primary Examiner(s): Nguyen, Tung X
• Assistant Examiner(s): Hawkins, Dominic E
• Application Number:
  US15/416,768
• Publication Number:
  US 20170211936A1
• Time in Patent Office: 740 Days
• CPC Class Codes: G01C 17/02 Magnetic compasses
• CPC Class Codes: G01C 17/38 Testing, calibrating, or compensating of compasses

Abstract

A method for performing continuous calibration of a magnetometer in a device includes during operation of a device, continually performing magnetometer measurements; continuously determining a state of the device; determining a magnetometer calibration model based on the magnetometer measurements and the state of the device; continually evaluating an accuracy of the magnetometer calibration model based the magnetometer measurements and the state of the device; and updating the magnetometer calibration model based on the evaluation of the accuracy magnetometer calibration model, the magnetometer measurements, and the state of the device.

Claims

1. A method for performing continuous calibration of a magnetometer in a device, the method comprising:

during operation of the device, continually performing magnetometer measurements;

continuously determining, by a processor circuit, a state of the device selected from a plurality of states of the device, where the plurality of states comprises;

a two dimensional state indicating the device is experiencing planar motion; and

a three dimensional state indicating the device is experiencing three dimensional motion;

determining, by the processor circuit, a magnetometer calibration model based on the magnetometer measurements and the state of the device;

continually evaluating, by the processor circuit, an accuracy of the magnetometer calibration model based on the magnetometer measurements and the state of the device;

updating, by the processor circuit, the magnetometer calibration model based on the evaluation of the accuracy of the magnetometer calibration model, the magnetometer measurements, and the state of the device; and

estimating an attitude or a heading of the device using the updated magnetometer calibration model.

2. The method of claim 1, where the plurality of states comprises:

an uncalibrated state indicating a calibration is not available.

3. The method of claim 1, where the device is configured to transition from one of the plurality of states to another of the plurality of states during operation and the processor circuit updates the magnetometer calibration model based on the state of the device.

4. The method of claim 1, where the processor circuit uses a fitting algorithm to determine the magnetometer calibration model.

5. The method of claim 4, where the fitting algorithm comprises a linear least squares algorithm.

6. The method of claim 4, where the fitting algorithm comprises a nonlinear least squares algorithm.

7. The method of claim 4, where the processor circuit determines a residual root mean square (“

RMS”

) error of the fitting algorithm to evaluate the accuracy of the magnetometer calibration model.

8. The method of claim 4, where the processor circuit determines a geometric diversity of the measurements used by the fitting algorithm as measured by a spherical area defined by the measurements data to evaluate the accuracy of the magnetometer calibration model.

9. The method of claim 1, where the processor circuit determines a relative difference between an estimated magnetic field strength from the magnetometer calibration model and a predicted local earth magnetic field strength to evaluate the accuracy of the magnetometer calibration model.

10. The method of claim 9, where the predicted local earth magnetic field strength is computed from the World Magnetic Model (WMM).

11. The method of claim 1 further comprising:

using the estimated heading of the device to allow for more accurate heading measurements in a direction application on the device.

12. A magnetometer calibration system, comprising:

a magnetometer;

a processor circuit coupled to the magnetometer and configured to execute instructions to perform continuous calibration of the magnetometer in a device, the calibration comprising;

during operation of the device, continually performing magnetometer measurements;

continuously determining a state of the device selected from a plurality of states of the device;

determining a magnetometer calibration model based on the magnetometer measurements and the state of the device, wherein determining the magnetometer calibration model comprises performing a batch model fit of an ellipsoid having a surface onto which the magnetometer measurements lie;

continually evaluating an accuracy of the magnetometer calibration model based on the magnetometer measurements and the state of the device; and

updating the magnetometer calibration model based on the evaluation of the accuracy magnetometer calibration model, the magnetometer measurements, and the state of the device; and

estimating an attitude or a heading of the device using the updated magnetometer calibration model.

13. The system of claim 12, where the state of the device is one of a plurality of states comprising:

an uncalibrated state indicating a calibration is not available;

a two dimensional state indicating the device is experiencing planar motion; and

a three dimensional state indicating the device is experiencing three dimensional motion.

14. The system of claim 13, where the device is able to transition from one state to another state during operation and the processor circuit updates the magnetometer calibration model based on the state of the device.

15. The system of claim 12, where the processor circuit uses a fitting algorithm to determine the magnetometer calibration model.

16. The system of claim 15, where the fitting algorithm comprises a linear least squares algorithm.

17. The system of claim 15, where the fitting algorithm comprises a nonlinear least squares algorithm.

18. The system of claim 15, where the processor circuit determines a residual root mean square (“

RMS”

) error of the fitting algorithm to evaluate the accuracy of the magnetometer calibration model.

19. The system of claim 15, the processor circuit determines a geometric diversity of the measurements used by the fitting algorithm as measured by a spherical area defined by the measurements data to evaluate the accuracy of the magnetometer calibration model.

20. The system of claim 12, where the processor circuit determines a relative difference between an estimated magnetic field strength from the calibration model and a predicted local earth magnetic field strength to evaluate the accuracy of the magnetometer calibration model.

21. The system of claim 20, where the predicted local earth magnetic field strength is computed from the World Magnetic Model (WMM).

22. The system of claim 12 wherein determining the magnetometer calibration model comprises:

determining the ellipsoid'"'"'s shape with at least a predetermined accuracy, and

inverting the ellipsoid to remove effects of distortion in the magnetometer measurements.