Enhanced inertial system performance
An inertial system is provided. The system includes at least one inertial sensor, a processing unit and a plurality of Kalman filters implemented in the processing unit. The Kalman filters receive information from the at least one inertial sensor, and at most one of the plurality of Kalman filters has processed zero velocity updates on the last cycle. The plurality of Kalman filters is used to optimize system response and performance during periods of intermittent motion.
1. A method of generating data in an inertial system, the method comprising:
processing multiple Kalman filters at a processing unit;
selecting one of the multiple Kalman filters that did not process zero velocity updates when motion was occurring; and
copying an error state vector and a covariance of the selected Kalman filter that did not process zero velocity updates when motion was occurring to the other filters.
2. The method of claim 1 , wherein processing multiple Kalman filters comprises processing measurements not associated with zero velocity updates.
3. The method of claim 1 , wherein processing multiple Kalman filters comprises updating the Kalman filters.
4. The method of claim 1 , wherein determining when motion was occurring comprises using a solution separation algorithm.
5. The method of claim 1 , wherein determining when motion was occurring comprises taking observations and determining if they exceed a threshold.
6. A method for generating data in an inertial system, the method comprising:
forming non-zero velocity update observations;
updating the Kalman filters at a processing unit;
detecting whether there has been motion;
forming zero velocity update observations if no motion is detected;
copying an error state vector and a covariance of the other Kalman filters that have processed one more cycle of zero velocity updates to the Kalman filters that have processed one less cycle of zero velocity updates if no motion is detected;
applying zero velocity updates to the Kalman filter that processes zero velocity updates on all cycles, when no motion is detected; and
copying an error state vector and a covariance of the Kalman filter that did not process zero velocity updates when motion was occurring to the Kalman filters when motion is detected.
7. The method of claim 6 , wherein copying the error state vector and the covariance of the Kalman filters that have processed one more cycle of zero velocity updates to the Kalman filters that have processed one less cycle of zero velocity updates includes copying the error state vector and the covariance of the Kalman filter that has processed zero velocity updates on the last cycle to the Kalman filter that has not processed zero velocity updates on the last cycle.
8. The method of claim 6 , wherein copying the error state vector and the covariance of the Kalman filter that did not process zero velocity updates when motion was occurring to the other Kalman filters comprises copying the error state vector and the covariance of the Kalman filter that did not process zero velocity updates on the last cycle to the Kalman filter that processed zero velocity updates on the last cycle.
9. The method of claim 6 , wherein copying the error state vector and the covariance of the Kalman filter that did not process zero velocity updates when motion was occurring to the other Kalman filters comprises copying the error state vector and the covariance of the Kalman filter that did not process zero velocity updates on the last two cycles to the Kalman filter that processed zero velocity updates on the last cycle.
10. The method of claim 6 , wherein detecting whether there has been motion comprises using a solution separation algorithm.
11. The method of claim 6 , wherein detecting whether there has been motion comprises taking the observations and determining if they exceed a threshold.
12. A non-transitory machine-readable medium having instructions embodied thereon for a method for generating inertial measurement data, the method comprising:
forming non-zero velocity update observations;
updating the Kalman filters;
detecting whether there has been motion;
forming zero velocity update observations if no motion is detected;
copying an error state vector and a covariance of the Kalman filters that have processed one more cycle of zero velocity updates to the Kalman filters that have processed one less cycle of zero velocity updates if no motion is detected;
applying zero velocity updates to the Kalman filter that processes zero velocity updates on all cycles, when no motion is detected; and
copying an error state vector and a covariance of the Kalman filter that did not process zero velocity updates when motion was occurring to the other Kalman filters when motion is detected.
13. The non-transitory machine-readable medium of claim 12 , wherein detecting whether there has been motion comprises using a solution separation algorithm.
14. The non-transitory machine-readable medium of claim 12 , wherein detecting whether there has been motion comprises taking the observations and determining if they exceed a threshold.