Parameterization of non-linear/non-Gaussian data distributions for efficient information sharing in distributed sensor networks

• US 7,698,108 B2
• Filed: 10/10/2007
• Issued: 04/13/2010
• Est. Priority Date: 10/10/2006
• Status: Expired due to Fees

• Current Assignee: Frantorf Investments GmbH LLC (Intellectual Ventures LLC)
• Original Assignee: Frantorf Investments GmbH LLC (Intellectual Ventures LLC)
• Inventors: Edelson, Geoffrey S., Haney, Philip J., Fiore, Paul D.
• Primary Examiner(s): Feliciano; Eliseo Ramos
• Assistant Examiner(s): Huynh; Phuong
• Application Number:
  US11/973,693
• Publication Number:
  US 20080215298A1
• Time in Patent Office: 916 Days
• Field of Search: 702179-190, 702121-123
• US Class Current: 702/189
• CPC Class Codes: G06F 18/251 of input or preprocessed data
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters

Abstract

A method for parameterization of data distributions for efficient information sharing in distributed sensor networks including a plurality of sensors, comprising the steps of performing Bayesian multi-source data fusion and sharing probalistic data information.

Claims

1. A method for updating a target track, the method comprising:

receiving a measurement associated with a target track at a first sensor node;

estimating a local posterior density based on the received measurement at the first sensor node;

receiving parameterized data from a second sensor at the first sensor node;

receiving third parameterized data from a third sensor at the first sensor node;

reconstructing a second local posterior density by a processor at the first sensor node based on the received parameterized data;

reconstructing a third local posterior density by the processor at the first sensor node based on the received third parameterized data;

estimating a global posterior density using the estimated local posterior density the reconstructed second local posterior density at the first sensor node, and the reconstructed third local posterior density;

updating the target track using the estimated global posterior density by the processor at the first sensor node.

2. The method of claim 1, further comprising, at the first sensor node, parameterizing the estimated local posterior density;

and sending the parameterized local posterior density to the second sensor.

3. The method of claim 2, wherein the estimated local posterior density is parameterized using a generalized lambda distribution.

4. The method of claim 1, wherein estimating the global posterior density comprises multiplying joint probability densities of the estimated local posterior density and the reconstructed second local posterior density.

5. The method of claim 1, wherein the local posterior density is estimated using a particle filter.

6. The method of claim 1, wherein the second local posterior density is reconstructed using an inverse transformation method.

7. A node comprising:

a sensor configured to generate a measurement associated with a target track;

a receiver configured to receive parameterized data from a second sensor node and to receive third parameterized data from a third sensor node; and

a sensor manager configured

to estimate a local posterior density based on the generated measurement;

to reconstruct a second local posterior density based on the received parameterized data;

to reconstruct a third local posterior density based on the received third parameterized data;

to estimate a global posterior density using the estimated local posterior density the reconstructed second local posterior density; and

the reconstructed third local posterior density; and

to update the target track using the estimated global posterior density.

8. The node of claim 7, wherein the sensor manager is further configured to parameterize the estimated local posterior density;

and to send the

parameterized local posterior density to the second node.

9. The node of claim 8, wherein the estimated local posterior density is parameterized using a generalized lambda distribution.

10. The node of claim 7, wherein estimating the global posterior density comprises multiplying joint probability densities of the estimated local posterior density and the reconstructed second local posterior density.

11. The node of claim 7, wherein the local posterior density is estimated using a particle filter.

12. The node of claim 7, wherein the second local posterior density is reconstructed using an inverse transformation method.

13. A system comprising:

a plurality of sensing nodes, wherein each sensing node of the plurality of sensing nodes comprises

a sensor configured to generate a measurement associated with a target track;

a receiver configured to receive parameterized data from the plurality of sensing nodes; and

a sensor manager configured

to estimate a local posterior density based on the generated measurement;

to reconstruct a plurality of local posterior densities based on the received parameterized data;

to estimate a global posterior density using the estimated local posterior density and the reconstructed plurality of local posterior densities; and

to update the target track using the estimated global posterior density.

14. The system of claim 13, wherein estimating the global posterior density comprises multiplying joint probability densities of the estimated local posterior density and the reconstructed plurality of local posterior densities.

15. The system of claim 13, wherein the local posterior density is estimated using a particle filter.

16. The system of claim 13, wherein the sensor manager is further configured to parameterize the estimated local posterior density;

and to send the parameterized local posterior density to the plurality of sensing nodes.

17. The system of claim 16, wherein the estimated local posterior density is parameterized using a generalized lambda distribution.