High speed gyrocompass alignment via multiple Kalman filter based hypothesis testing

• US 7,512,493 B2
• Filed: 05/31/2006
• Issued: 03/31/2009
• Est. Priority Date: 05/31/2006
• Status: Active Grant

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: Morgan, Kenneth S., Thompson, I. Clay Jr.
• Primary Examiner(s): ARTHUR JEANGLAUDE, GERTRUDE
• Application Number:
  US11/421,337
• Publication Number:
  US 20070282528A1
• Time in Patent Office: 1,035 Days
• Field of Search: 701/200, 701/207, 701/214, 701/220, 701/221, 701/224, 701/225, 342/357.04, 342/357.14
• US Class Current: 701/214
• CPC Class Codes: G01C 21/165 combined with non-inertial navigation instruments
• CPC Class Codes: G01C 21/188 for accumulated errors, e.g. by coupling inertial systems with absolute positioning systems

Abstract

A method of aligning a gyro-compass comprising operating at least two Kalman filters in a set of Kalman filters to generate an error correction to at least a single navigation solution in a set of navigation solutions in order to provide coarse alignment azimuth convergence. The method further comprises selecting at least one selected Kalman filter from the set of Kalman filters and at least one selected navigation solution from the set of navigation solutions, and operating the at least one selected Kalman filter and the at least one selected navigation solution to provide fine alignment convergence to a correct azimuth. The selecting is based at least in part on the generated error correction, each navigation solution includes an azimuth different from the azimuth of each other navigation solution and each navigation solution azimuth is separated by no more than two times a small angle error assumption.

Claims

1. A method of aligning a gyro-compass, the method including:

operating at least two Kalman filters in a set of Kalman filters to generate an error correction to at least a single navigation solution in a set of navigation solutions to provide coarse alignment azimuth convergence;

selecting at least one Kalman filter from the set of Kalman filters and at least one navigation solution from the set of navigation solutions, wherein the selecting is based at least in part on the generated error correction; and

operating the at least one selected Kalman filter and the at least one selected navigation solution to provide fine alignment convergence to a correct azimuth,

wherein each navigation solution includes an azimuth different from the azimuth of each other navigation solution; and

wherein each navigation solution azimuth is separated by no more than two times a small angle error assumption.

2. The method of claim 1, the method further comprising:

initializing the Kalman filters in the set of Kalman filters;

operating each Kalman filter from the set of Kalman filters on an associated navigation solution from the set of navigation solutions;

receiving the error corrections from the Kalman filters; and

selecting the Kalman filter and the associated navigation solution which generated the error correction within a pre-selected range of error corrections.

3. The method of claim 2, wherein the set of navigation solutions comprises a single navigation solution and wherein each Kalman filter from the set of Kalman filters operates on the single navigation solution.

4. The method of claim 1, where the selected navigation solution and the selected Kalman filter are formed from a linear combination of at least two navigation solutions and at least two Kalman filters, respectively.

5. The method of claim 1, further comprising operating each Kalman filter in the set of Kalman filters to provide coarse alignment convergence to within the small angle error assumption.

6. The method of claim 1, wherein initial azimuth error assumptions for the Kalman filters in the set of Kalman filters are uniformly distributed within 360°

.

7. The method of claim 1, wherein an initial azimuth error assumption for each Kalman filter in the set of Kalman filters is one of:

separated by no more than two times the small angle error assumption;

initialized for less than about one half of an angle separating the Kalman filters; and

separated by no more than two times the small angle error assumption and initialized for less than about one half of an angle separating the Kalman filters.

8. The method of claim 1, wherein the error correction comprises a measurement residual and an estimated azimuth error and wherein selecting at least one selected Kalman filter comprises:

selecting at least one Kalman filter that outputs at least one of a smaller measurement residual and a smaller estimated-azimuth error.

9. The method of claim 1, wherein selecting one of the set of Kalman filters comprises:

determining a subset of Kalman filters taken from the set of Kalman filters wherein the Kalman filters in the subset output smaller error corrections than the Kalman filters not in the subset;

operating the subset of Kalman filters as the selected Kalman filter; and

terminating the operation of the Kalman filters not in the subset of Kalman filters.

10. An inertial navigation unit comprising:

at least one sensor to provide input data;

at least two Kalman filters communicatively coupled to the at least one sensor;

wherein the inertial navigation system generates at least a single navigation solution based at least in part on the input data using the at least two Kalman filters;

wherein the inertial navigation system terminates the operation of one or more Kalman filter based on at least in part measurement residuals and estimated-convergent-parameter errors received from each Kalman filter, wherein the inertial navigation system operates at least one of a selected Kalman filter and a selected navigation solution to provide fine alignment convergence to a convergent parameter.

11. The inertial navigation unit of claim 10, wherein each Kalman filter includes an azimuth error assumption different from the azimuth error assumption of each of the other Kalman filters.

12. The inertial navigation unit of claim 10, wherein each navigation solution includes an azimuth error initialization different from the azimuth error initialization of each of the other navigation solutions.

13. The inertial navigation unit of claim 10, wherein the convergent parameter is an azimuth parameter and the estimated-convergent-parameter error is an estimated-azimuth error.

14. The inertial navigation unit of claim 10, wherein the at least one sensor comprises:

at least one inertial sensor to provide inertial data; and

at least one data sensor to provide observation data, wherein the inertial data and the observation data comprise the input data.

15. The inertial navigation unit of claim 10, wherein the at least two Kalman filters comprise at least thirty-six Kalman filters, wherein the at least one navigation solution comprises at least thirty-six navigation solutions, wherein the number of Kalman filter equals the number of navigation solutions, and wherein the at least one of the selected Kalman filter and the selected navigation solution provide azimuth convergence to the required accuracy level, wherein a completion time for a gyrocompass alignment to the required accuracy level is reduced.

16. The inertial navigation unit of claim 15, wherein each Kalman filter includes an azimuth error assumption different from the azimuth error assumption of each of the other Kalman filters.

17. The inertial navigation unit of claim 15, wherein each navigation solution includes an initial azimuth different from the initial azimuth of each of the other navigation solutions.

18. The inertial navigation unit of claim 10, wherein the at least two Kalman filters comprise at least thirty-six Kalman filters, wherein each Kalman filter includes an azimuth error assumption different from the azimuth error assumption of each of the other Kalman filters, wherein the at least one navigation solution comprises one navigation solution, wherein the one navigation solution is the selected navigation solution, and wherein the selected Kalman filter and the selected navigation solution provide azimuth convergence to the required accuracy level, wherein a completion time for a gyrocompass alignment to the required accuracy level is reduced.

19. A program product comprising program instructions, embodied on a storage medium, that are operable to cause a programmable processor to:

operate at least two Kalman filters in a set of Kalman filters to generate an error correction to at least a single navigation solution in a set of navigation solutions to provide azimuth convergence;

select at least one selected Kalman filter from the set of Kalman filters and at least one selected navigation solution from the set of navigation solutions, wherein the selecting is based at least in part on the generated error correction; and

operate the at least one selected Kalman filter and the at least one selected navigation solution to provide azimuth convergence to the required accuracy level, wherein a completion time for a gyrocompass alignment to the required accuracy level is reduced.

20. The program product of claim 19, further comprising instructions operable to cause the programmable processor to:

initialize the Kalman filters in the set of Kalman filters;

operate each Kalman filter from the set of Kalman filters on an associated navigation solution from the set of navigation solutions;

receive the error corrections from the Kalman filters; and

select the Kalman filter and the associated navigation solution which generated the error correction within a pre-selected range of error corrections.