Risk management for object identification

• US 8,595,177 B1
• Filed: 03/08/2011
• Issued: 11/26/2013
• Est. Priority Date: 03/08/2011
• Status: Active Grant

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Nelson, Kenric P., Scannell, Brian J., Landau, Herbert
• Primary Examiner(s): Gaffin, Jeffrey A
• Assistant Examiner(s): Smith, Paulinho E
• Application Number:
  US13/042,699
• Time in Patent Office: 994 Days
• Field of Search: None
• US Class Current: 706/52
• CPC Class Codes: G01S 13/723 by using numerical data
• CPC Class Codes: G01S 7/41 using analysis of echo signal for target characterisation; Target signature; Target cross-section

Abstract

The technology described herein includes a system and/or a method of risk management for object identification. The method includes selecting at least two fusion functions from a plurality of fusion functions. Each of the fusion functions is associated with a different risk bias. The method further includes fusing a received set of data based on each fusion function to form at least two sets of fused data. The method further includes generating a probability of identification of the classification of the object based on each set of the fused data. The method further includes determining a risk bias difference between the probability of identification of each set of the fused data.

Claims

1. A system for risk management for object identification, the system comprising:

one or more processors;

a fusion function selection module, using the one or more processors, configured to select at least two fusion functions from a plurality of fusion functions, each of the fusion functions associated with a different risk bias;

a data fusion module, using the one or more processors, configured to fuse a received set of data based on each fusion function to form at least two sets of fused data;

a probability of identification module, using the one or more processors, configured to generate a probability of identification of a classification of the object based on each set of the fused data; and

a risk bias determination module, using the one or more processors, configured to determine a risk bias difference between the probability of identification of each set of the fused data, wherein the risk bias difference is a confidence interval which accounts for different levels of risk corresponding to the probability of identification of the classification of the object.

2. The system of claim 1, further comprising a fusion function generation module, using the one or more processors, configured to generate the plurality of fusion functions based on a base fusion function and one or more parameters.

3. The system of claim 2, wherein the one or more parameters comprises at least two sets of alpha and beta parameter pairs, and

the system further comprising a parameter module configured to determine the least two sets of alpha and beta parameter pairs, each alpha parameter associated with a fusion function, each beta parameter associated with a degree of expected independence of a received set of data, and the received set of data comprising information associated with the classification of the object.

4. The system of claim 1, wherein a first fusion function is associated with a decisive or robust risk bias and a second fusion function is associated with a neutral risk bias.

5. The system of claim 4, wherein the fusion function selection module, using the one or more processors, is further configured to determine the decisive risk bias, the neutral risk bias, and the robust risk bias based on probabilities of one or more training samples.

6. The system of claim 5, wherein the parameter module, using the one or more processors, is further configured to define the associations between classifications of the object based on a generalized mean between the probabilities of the one or more training samples.

7. The system of claim 5, wherein the decisive risk bias is associated with a higher confidence of identification of the classification of the object, the neutral risk bias is associated with a non-biased confidence of identification of the classification of the object, and the robust risk bias is associated with a lower confidence of identification of the classification of the object.

8. A method for risk management for object identification, the method comprising:

selecting at least two fusion functions from a plurality of fusion functions, each of the fusion functions associated with a different risk bias;

fusing a received set of data based on each fusion function to form at least two sets of fused data;

generating a probability of identification of the classification of the object based on each set of the fused data; and

determining a risk bias difference between the probability of identification of each set of the fused data, wherein the risk bias difference is a confidence interval which accounts for different levels of risk corresponding to the probability of identification of the classification of the object.

9. The method of claim 8, further comprising generating the plurality of fusion functions based on a base fusion function and one or more parameters.

10. The method of claim 8, wherein each of the fusion functions in the plurality of fusion functions is different.

11. The method of claim 8, further comprising optimizing the risk bias associated with each of the fusion functions based on a risk profile associated with one or more training samples.

12. The method of claim 8, wherein the probability of identification for each set of fused data is indicative of the identified object being a target object.

13. The method of claim 8, wherein the received set of data comprising radar signal data.

14. A non-transitory computer program product, tangibly embodied in an information carrier, the computer program product including instructions being operable to cause a data processing apparatus to:

select at least two fusion functions from a plurality of fusion functions, each of the fusion functions associated with a different risk bias;

fuse a received set of data based on each fusion function to form at least two sets of fused data;

generate a probability of identification of the classification of the object based on each set of the fused data; and

determine a risk bias difference between the probability of identification of each set of the fused data, wherein the risk bias difference is a confidence interval which accounts for different levels of risk corresponding to the probability of identification of the classification of the object.