Alignment Method for an Inertial Unit

• US 20160290808A1
• Filed: 11/24/2014
• Published: 10/06/2016
• Est. Priority Date: 11/22/2013
• Status: Active Grant

• Current Assignee: Association pour la Recherche et le Développement des méthodes et processus Industriels
• Original Assignee: Association Pour La Recherche et Le Developpement Des Methodes et Processus Industriels Armines, Sagem Defense Securite (Safran SA)
• Inventors: BARRAU, Axel, BONNABEL, Silvere
• Granted Patent: US 10,859,380 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G01C 25/005 initial alignment, calibration or starting-up of inertial devices

Abstract

The invention relates to an alignment method based on a simplified mode allowing processing using an invariant Kalman filter, in which each speed involved in the navigation equations is expressed in a work reference frame (Rt) translated with respect to an inertial reference frame (Ri) and for which the origin moves along a reference inertial trajectory, the carrier of which is assumed to be close (geographic origin for alignment with the sun at known position, GPS trajectory for alignment in-motion, etc.). This simplified mode comprises the repetition of the following steps to estimate a mobile carrier state (P): propagation (PROP) determining an estimated current state from a preceding estimated state, inertial sensor measurements and theoretical information on the carrier trajectory (P) updating (MAJ) the estimated state using theoretical information on the carrier trajectory (P). The deterministic uncertainties of the sensors (bias/drift/scale factors, etc.) are only estimated during the propagation and update steps. A perturtion step (PERT) then allows inclusion of the values ignored in the simplified mode and estimation of these values by an optimisation method.

Claims

1. An method for aligning a carrier device (P) moving in an inertial reference frame (Ri) and comprising an inertial measurement unit (IN), the inertial measurement unit comprising a plurality of inertial sensors (RV, RO) and at least one invariant Kalman filter, the method comprising repeating the following steps for estimating a state of the carrier device (P) representative of its speed and its orientation:

a propagation step (PROP) determining an estimated current state from a preceding estimated state and measurements of the inertial sensors,

a update step updating (MAJ) the estimated state by means of the information a priori on the trajectory of the carrier device (P),

the method being characterized in that, to allow processing by at least one invariant Kalman filter, a simplified mode is implemented in which bias of the sensors are not estimated during propagation and updating steps, each speed occurring in the updating equations being expressed in a work reference frame (Rt) translated relative to the inertial reference frame (Ri) and whereof the origin moves along an inertial reference trajectory whereof the carrier device is deemed close.

2. The method according to the preceding claim, characterized in that the propagation equations produce a gravity variable which depends only on time.

3. The method according to one of the preceding claims, characterized in that the inertial reference trajectory is constructed from information a priori which is not measured by the inertial sensors.

4. The method according to one of the preceding claims, characterized in that in the simplified mode, each position occurring in the updating equations is expressed in a work reference frame (Rt) translated relative to the inertial reference frame (Ri).

5. The method according to one of the preceding claims, characterized in that processing is also carried out in parallel in which additional quantitiesquantities influencing the estimation are estimated in the following way:

at each instant, the differential relative to the additional quantitiesquantities of the estimation provided by the simplified mode is calculated recursively,

the additional quantitiesquantities are calculated by optimization of a quality criterion of the estimation.

6. The method according to claim 5, according to which calculation of the additional quantitiesquantities by optimization of a quality criterion is performed on completion of said alignment method.

7. The method according to claim 5, according to which calculation of the additional quantitiesquantities by optimization of a quality criterion is conducted permanently or at one or more instants of the alignment method.

8. An alignment method according to one of the preceding claims, wherein a series of gains is calculated prior to repeated execution of the propagation and updating steps (MAJ) in simplified mode, each gain occurring in an iteration of the updating step.

9. The method according to one of the preceding claims, characterized in that the origin of the work reference frame (Rt) is the origin of the geographical reference frame.

10. The method according to one of claims 1 to 8, characterized in that the origin of the work reference frame follows a GPS trajectory.

11. An inertial measurement unit (IN) movable in an inertial reference frame (Ri) and comprising:

a plurality of inertial sensors (CV, CO),

at least one invariant Kalman filter,

an embedded computer of said unit performing repetition of the following steps for estimating a state of the carrier device (P) representative of its speed and its orientation;

a propagation step (PROP) determining an estimated current state from a preceding estimated state and measurements of the inertial sensors,

an update step updating (MAJ) the estimated state by means of the

information a priori on the trajectory of the carrier device (P), characterized in that, to allow processing by the invariant Kalman filter, the computer is configured to execute an alignment method according to one of claims 1 to 10.

12. A computer program product, characterized in that it comprises program code instructions for execution of the steps of the method according to one of claims 1 to 10, when this program executed by a computer embedded in an inertial measurement unit can also be implemented.