ESTIMATING A STATE OF AT LEAST ONE TARGET USING A PLURALITY OF SENSORS

• US 20120089554A1
• Filed: 06/22/2010
• Published: 04/12/2012
• Est. Priority Date: 06/29/2009
• Status: Active Grant

• Current Assignee: BAE Systems Plc
• Original Assignee: BAE Systems Plc
• Inventors: Claxton, Christopher David, Nicholson, David, Reece, Steven, Roberts, Stephen John
• Granted Patent: US 8,706,676 B2
• Time in Patent Office: Days
• US Class Current: 706/52
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

The method includes receiving a plurality of target observations from a respective plurality of sensors and using the target observations to compute target state estimates. Each of the target state estimates is assessed to check if it suffers from one of a set of modelled possible fault types. The target state estimates are adjusted to compensate for any modelled fault type when it is assessed to be suffering from that fault type. A reliability value is computed for each of the target state estimates and the target state estimates are fused together based on the computed reliability values to produce a fused target state estimate.

Claims

1. A method of estimating a state of at least one target (101) using a plurality of sensors (102), the method including:

(1) receiving (202) a plurality of target observations (106) from a respective plurality of sensors;

(2) using (204) the target observations to compute target state estimates;

(3) assessing (206) whether each of the target state estimates suffers from one of a set of modelled possible fault types;

(4) adjusting (208) the target state estimates to compensate for a said modelled fault type if that target state estimate is assessed to suffer from that modelled fault type;

(5) computing (210) a reliability value for each of the target state estimates, and

(6) fusing (212) the target state estimates based on the computed reliability values to produce a fused target state estimate.

2. A method according to claim 1, wherein the steps (1), (2) and/or (3) involve a multi-hypothesis dual Kalman filter process.

3. A method according to any one of the preceding claims, further including estimating a time period during which each of the target state estimates suffered from the modelled possible fault type.

6. A method according to any one of the preceding claims, including calculating a failure hypothesis associated with each of the target state estimates, the failure hypothesis include a state covariance matrix;

predictions for the target state estimate that exclude at least some of the observations;

a probability value associated with the target state estimate suffering from the modelled fault type.

7. A method according to any one of the preceding claims, wherein the step of computing (210) a reliability value for each of the target state estimates is performed for the observation received from each sensor individually.

8. A method according to claim 7, wherein the step of computing (210) the reliability of the target state estimates includes computing a distance between the target state estimate as computed using the observation from a first one of the sensors and a target state estimate as computed using the observation(s) from at least one other of the sensors.

9. A method according to claim 8, wherein a said target state estimate is computed as being unreliable if the distance exceeds a predefined threshold value.

10. A method according to claim 9, wherein the step of fusing (210) the target state estimates includes fusing the target state estimates only if the target state estimates are computed as being reliable.

11. A method according to any one of the preceding claims, wherein the fusion step (210) includes Kalman filter (or even covariance intersection) and mixture reduction processes.

12. A computer program product comprising a computer readable medium, having thereon computer program code means, when the program code is loaded, to make the computer execute method of a estimating a state of at least one target using a plurality of sensors according to any one of the preceding claims.

13. A system configured to estimate a state of at least one target (101) using a plurality of sensors (102), the system including:

a plurality of sensors configured to receive target observations;

a device configured to receive (202) a plurality of target observations (106) from the plurality of sensors;

a device configured to use (204) the target observations to compute target state estimates;

a device configured to assess (206) whether each of the target state estimates suffers from one of a set of modelled possible fault types;

a device configured to adjust (208) the target state estimates to compensate for a said modelled fault type if that target state estimate is assessed to suffer from that modelled fault type;

a device configured to compute (210) a reliability value for each of the target state estimates, and

a device configured to fuse (212) the target state estimates based on the computed reliability values to produce a fused target state estimate.

14. A system according to claim 13, wherein the sensors (102) are configured to transfer their observations to a remote device (104) that includes:

the device configured to use the target observations to compute target state estimates;

the device configured to assess whether each of the target state estimates suffers from one of a set of modelled possible fault types, and the device configured to adjust the target state estimates to compensate for a said modelled fault type if that target state estimate is assessed to suffer from that modelled fault type.

15. A system according to claim 13, wherein each of the sensors (102) include the device configured to use the target observations to compute target state estimates;

the device configured to assess whether each of the target state estimates suffers from one of a set of modelled possible fault types, and the device configured to adjust the target state estimates to compensate for a said modelled fault type if that target state estimate is assessed to suffer from that modelled fault type.