Magnetometer bias and anomaly detector

• US 8,577,640 B2
• Filed: 08/10/2012
• Issued: 11/05/2013
• Est. Priority Date: 08/17/2011
• Status: Active Grant

• Current Assignee: Invensense Incorporated (TDK Corporation)
• Original Assignee: Invensense Incorporated (TDK Corporation)
• Inventors: Keal, William Kerry
• Primary Examiner(s): Kundu, Sujoy
• Assistant Examiner(s): PARK, HYUN D
• Application Number:
  US13/572,441
• Publication Number:
  US 20130046489A1
• Time in Patent Office: 452 Days
• Field of Search: 702/57, 702/104, 702/105, 702/189, 702/193, 702/194, 702/195, 702/196, 324/244, 324/202
• US Class Current: 702/104
• CPC Class Codes: G01R 33/02 Measuring direction or magnitude of magnetic fields or magnetic flux G01R33/20 takes precedence

Abstract

The computer implemented method, system or computer program product comprises collecting magnetometer data from the device; and calculating a center of a shape of the magnetometer data as a result of minimization. The minimization of calculating the center of the shape further comprises calculating a plurality of running sums of the magnetometer data; storing the plurality of running sums; storing a count of the number of terms in each of the running sums; and calculating the center of the shape and setting the estimated magnetometer bias to the center of the shape. The radius of the sphere is calculated to ensure accuracy in the estimator of the magnetometer bias.

Claims

1. A computer implemented method of estimating a magnetometer bias for a device comprising:

collecting magnetometer data from the device; and

calculating using a computer a center of a shape of the magnetometer data as a result of minimization;

wherein the magnetometer data is represented by data points, wherein the minimization is a derivative of a sum of weighted distances of pairs of data points, wherein each of the weighted distances is weighted based upon how far apart the pair of points are from each other;

wherein the minimization of calculating the center of the shape further comprises;

calculating a plurality of running sums of the magnetometer data;

storing the plurality of running sums;

storing a count of the number of terms in each of the running sums;

and calculating the center of the shape and setting the estimated magnetometer bias to the center of the shape.

2. The computer implemented method of claim 1, wherein the shape comprises a sphere.

3. The computer implemented method of claim 1, wherein the shape comprises an ellipsoid.

4. The computer implemented method of claim 1, wherein the calculating comprises calculating a plurality of running sums of the magnetometer data, a plurality of running sums of the square of the magnetometer data, a plurality of running sums of the cube of the magnetometer data and a plurality of running sums of the cross products of the magnetometer data.

5. The computer implemented method of claim 1, which includes evaluating if the estimated bias is accurate.

6. The computer implemented method of claim 1, further comprises calculating a radius of the sphere utilizing the plurality of running sums.

7. The computer implemented method of claim 6, wherein the calculated plurality of running sums are normalized utilizing the minimization to establish a threshold.

8. The computer implemented method of claim 7, wherein the threshold can be utilized to determine if the estimated magnetometer bias has converged.

9. The computer implemented method of claim 6, wherein it is determined if a particular magnetometer data point is on the sphere based upon the computed minimization.

10. The computer implemented method of claim 1, wherein a data range can be utilized to track convergence of the magnetometer bias.

11. The computer implemented method of claim 1, wherein a minimum eigenvalue of a matrix is utilized to track convergence of the bias.

12. The computer implemented method of claim 1, which includes determining if the magnetometer bias remains within a predetermined range for a predetermined time period to track convergence of the bias.

13. The computer implemented method of claim 1, which excludes magnetometer data being used as part of the magnetometer bias calculation, based upon the absolute value of the average normalized magnetometer data being below a threshold.

14. The computer implemented method of claim 1, wherein the sum of the weighted distances comprises the sum of the weighted distances squared.

15. A device comprising:

a processor;

a magnetometer for providing magnetometer data;

and a memory coupled to the processor;

wherein the memory receives the magnetometer data;

wherein the memory includes an algorithm for estimating the bias of the magnetometer data;

wherein the processor executes the algorithm to perform the following operations;

calculating a center of a shape of the magnetometer data as a result of minimization;

wherein the magnetometer data is represented by data points, wherein the minimization is a derivative of a sum of weighted distances of pairs of data points, wherein each of the weighted distances is weighted based upon how far apart the pairs of data points are from each other;

wherein the minimization of calculating the center of the shape further comprises calculating a plurality of running sums of the magnetometer data;

storing the plurality of running sums;

storing a count of the number of terms in each of the running sums; and

calculating the center of the shape and setting the estimated magnetometer bias to the center of the shape.

16. The device of claim 15, wherein the shape comprises a sphere.

17. The device of claim 16, wherein it is determined if a particular magnetometer data point is on the sphere based upon the computed minimization.

18. The device of claim 15, wherein the shape comprises an ellipsoid.

19. The device of claim 15, wherein calculating a plurality of running sums of the magnetometer data, a plurality of running sums of the square of the magnetometer data, a plurality of running sums of the cube of the magnetometer data and a plurality of running sums of the cross products of the magnetometer data.

20. The device of claim 15, wherein the minimization is computed utilizing the plurality of running sums.

21. The device of claim 20, further comprises calculating a radius of the sphere utilizing the plurality of running sums.

22. The device of claim 15, wherein the calculated plurality of running sums are normalized utilizing the minimization to establish a threshold.

23. The device of claim 22, wherein the threshold can be utilized to determine if the estimated magnetometer bias has converged.

24. The device of claim 15, wherein a data range can be utilized to track convergence of the magnetometer bias.

25. The device of claim 15, wherein a minimum eigenvalue of a matrix is utilized to track convergence of the bias.

26. The device of claim 15, which includes determining if the magnetometer bias remains within a predetermined range for a predetermined time period to track convergence of the bias.

27. The device of claim 15, wherein the device comprises any of a portable phone, laptop, tablet, notebook, and smartphone.

28. The device of claim 15, wherein the sum of the weighted distances comprises the sum of the weighted distances squared.

29. A non-transitory computer program product including a computer readable medium to be executed on a computer, the computer readable medium for performing the following operations comprising:

collecting magnetometer data from the device; and

calculating a center of a shape of the magnetometer data as a result of minimization;

wherein the magnetometer data is represented by data points, wherein the minimization is a derivative of a sum of weighted distances of pairs of data points, wherein each of the weighted distances is weighted based upon how far apart the pairs of data points are from each other;

wherein the minimization of calculating of a center of an the shape further comprises;

calculating a plurality of running sums of the magnetometer data;

storing the plurality of running sums;

storing a count of the number of terms of running sums; and

calculating the bias from the stored running sums when the count of the plurality of running sums is less than a predetermined threshold.

30. The computer implemented method of claim 29, wherein the calculating comprises calculating a plurality of running sums of the magnetometer data, a plurality of running sums of the square of the magnetometer data, a plurality of running sums of the cube of the magnetometer data and a plurality of running sums of the cross products of the magnetometer data.