Method and system for classifying vehicle tracks

• US 9,208,447 B1
• Filed: 09/14/2012
• Issued: 12/08/2015
• Est. Priority Date: 09/14/2012
• Status: Active Grant

• Current Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Original Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Inventors: Pedersen, Richard N., Barbustiak, Mark E.
• Primary Examiner(s): Vincent, David
• Assistant Examiner(s): Buss, Benjamin
• Application Number:
  US13/617,879
• Time in Patent Office: 1,180 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: F41G 1/00 Sighting devices for indirect laying of fire F41G3/16; bombsights F41G3/24 ; structurally associated with laser telemeters F41G3/065; mounting tubular or beam shaped aiming devices on firearms F41G11/001
• CPC Class Codes: F41G 3/22 for vehicle-borne armament, e.g. on aircraft
• CPC Class Codes: F41G 7/007 Preparatory measures taken before the launching of the guided missiles
• CPC Class Codes: G01S 13/726 Multiple target tracking
• CPC Class Codes: G01S 15/66 Sonar tracking systems
• CPC Class Codes: G01S 17/66 Tracking systems using electromagnetic waves other than radio waves
• CPC Class Codes: G01S 7/41 using analysis of echo signal for target characterisation; Target signature; Target cross-section
• CPC Class Codes: G01S 7/412 based on a comparison between measured values and known or stored values
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

A method for classifying unclassified vehicle tracks using real-time sensor data comprises acquiring a computer database of vehicle track characteristics data for known vehicle tracks; defining vehicle track signatures based on the vehicle track characteristics data; and generating a graph based on the vehicle track signatures, the graph having state nodes representative of distinguishing vehicle track characteristics defined in the vehicle track signatures and links between the state nodes representative of relationships between distinguishing vehicle track characteristics defined in the vehicle track signatures, said graph also having reporting nodes for classifying unknown vehicle tracks when sufficient distinguishing vehicle track characteristics have been observed. The method also comprises processing with the graph, using a processor, real-time sensor data corresponding to a first unclassified vehicle track until the first unclassified vehicle track is classified.

Claims

1. A method for classifying unclassified vehicle tracks using real-time sensor data comprising:

acquiring a computer database of vehicle track characteristics data for known vehicle tracks;

defining vehicle track signatures based on the vehicle track characteristics data including defining at least one relationship between a first vehicle track and a second vehicle track in which the first vehicle track and the second vehicle track are related;

generating a graph based on the vehicle track signatures, the graph having state nodes representative of distinguishing vehicle track characteristics defined in the vehicle track signatures and links between the state nodes representative of relationships between the distinguishing vehicle track characteristics defined in the vehicle track signatures, said graph also having reporting nodes for classifying unknown vehicle tracks when sufficient distinguishing vehicle track characteristics have been observed;

processing with the graph, using a processor, first real-time sensor data corresponding to a first unclassified vehicle track and second real-time sensor data corresponding to a second unclassified vehicle track by;

determining if the first unclassified vehicle track and the second unclassified vehicle track are related by kinematic analysis of the first real-time sensor data and the second real-time sensor data;

responsive to a determination that the first unclassified vehicle track and the second unclassified vehicle track are not related;

performing run-time processing to classify the first vehicle track by passing the first real-time sensor data through the graph until the first unclassified vehicle track is classified; and

performing run-time processing to classify the second vehicle track by passing the second real-time sensor data through the graph until the second unclassified vehicle track is classified;

responsive to a determination that the first unclassified vehicle track and the second unclassified vehicle track are related;

performing run-time processing to classify the first vehicle track by passing the first real-time sensor data and the second real-time sensor data through the graph until the first unclassified vehicle track is classified;

performing run-time processing to classify the second vehicle track by passing the first real-time sensor data and the second real-time sensor data through the graph until the second unclassified vehicle track is classified.

2. The method for classifying vehicle tracks of claim 1 wherein defining vehicle track signatures further comprises:

identifying in the computer database the distinguishing vehicle track characteristics for at least some of the known vehicle tracks;

prioritizing the distinguishing vehicle track characteristics for at least some of the known vehicle tracks, wherein said prioritizing comprises defining relationships between the distinguishing vehicle track characteristics.

3. The method for classifying vehicle tracks of claim 2, wherein prioritizing the distinguishing vehicle track characteristics further comprises prioritizing the distinguishing vehicle track characteristics and defining the reporting nodes according to a probability analysis.

4. The method for classifying vehicle tracks of claim 2, wherein defining relationships between the distinguishing vehicle track characteristics comprises defining a relationship between the distinguishing vehicle track characteristics representative of the vehicle performing a known tactical maneuver.

5. The method for classifying vehicle tracks of claim 2, wherein defining at least one relationship between a first vehicle track and a second vehicle track in which the first vehicle track and the second vehicle track are related comprises defining at least one relationship between a first vehicle track and a second vehicle track in which one of the vehicle tracks is for a vehicle capable of launching an attack weapon and the other of the vehicle tracks is for a launched attack weapon.

6. The method for classifying vehicle tracks of claim 1, wherein the vehicle track characteristics data comprises at least one of speed, size, position, IFF status, heading, acceleration, airborne state, in motion state, operating altitude above or depth below sea level, likelihood of operating within a region of interest, acoustic response characteristics, electromagnetic emanation characteristics, kinematic maneuver characteristics, and electro-optic characteristics.

7. The method for classifying vehicle tracks of claim 1, wherein passing the first real-time sensor data through the graph until a reporting node is reached and the first unclassified vehicle track is classified comprises:

receiving first initial real-time sensor data for a first unclassified vehicle track and passing the first initial real-time sensor data through an initial portion of the graph to determine a first current track state for the first unclassified vehicle track;

determining whether one of said reporting nodes has been reached; and

responsive to one of said reporting nodes being reached, classifying the first unclassified vehicle track;

or

responsive to one of said reporting nodes not being reached, receiving first updated real-time sensor data and passing the first updated real-time sensor data through an intermediate portion of the graph linked to a current track state to determine a first updated current track state, until one of said reporting nodes is reached and the first unclassified vehicle track is classified.

8. The method for classifying vehicle tracks of claim 7, wherein prioritizing the distinguishing vehicle track characteristics further comprises prioritizing the distinguishing vehicle track characteristics and defining the reporting nodes according to a probability analysis.

9. The method for classifying vehicle tracks of claim 7, wherein passing the second real-time sensor data through the graph until a reporting node is reached and the second unclassified vehicle track is classified comprises:

receiving second initial real-time sensor data for a second unclassified vehicle track and passing the second initial real-time sensor data through an initial portion of the graph to determine a second current track state for the second unclassified vehicle track; and

receiving second updated real-time sensor data and passing the second updated real-time sensor data through an intermediate portion of the graph linked to the second current track state to determine a second updated current track state, until the second unclassified vehicle track is classified.

10. The method for classifying vehicle tracks of claim 9, wherein determining if the first unclassified vehicle track and the second unclassified vehicle track are related by kinematic analysis of the first real-time sensor data and the second real-time sensor data comprises determining if the first unclassified vehicle track and the second unclassified vehicle track are related by kinematic analysis of one or more of the first initial real-time sensor data, first updated real-time sensor data, second initial real-time sensor data, or second updated real-time sensor data.

11. The method for classifying vehicle tracks of claim 10, wherein, when the first unclassified vehicle track and the second unclassified vehicle track are related,

passing the first real-time sensor data through the graph until the first unclassified vehicle track is classified comprises passing the first updated real-time sensor data in addition to the second updated real-time sensor data through the intermediate portion of the graph linked to the second current track state to determine a second updated current track state, until the second unclassified vehicle track is classified;

and

passing the second real-time sensor data through the graph until the second unclassified vehicle track is classified comprises passing the second updated real-time sensor data in addition to the first updated real-time data sensor to the intermediate portion of the graph linked to the first current track state to determine a first updated current track state, until the first unclassified vehicle track is classified.

12. The method for classifying vehicle tracks of claim 10, further comprising, when the first unclassified vehicle track and the second unclassified vehicle track are related,

passing, when the first unclassified vehicle track has been classified, a first vehicle track classification in addition to the second updated real-time sensor data through the intermediate portion of the graph linked to the second current track state to determine a second updated current track state, until the second unclassified vehicle track is classified;

and

passing, when the second unclassified vehicle track has been classified, a second vehicle track classification in addition to the first updated real-time sensor data to the intermediate portion of the graph linked to the first current track state to determine a first updated current track state, until the first unclassified vehicle track is classified.

13. A method for pre-processing vehicle track characteristics data to define a graph for classifying unclassified vehicle tracks comprising:

acquiring a computer database of vehicle track characteristics data for known vehicle tracks;

identifying distinguishing characteristics for at least some of the known vehicle tracks;

prioritizing the identified distinguishing characteristics for at least some of the known vehicle tracks; and

generating a graph having state nodes representative of the identified distinguishing characteristics and links between the state nodes representative of the prioritizing of the identified distinguishing characteristics, said graph also having reporting nodes for classifying unclassified vehicle tracks when a sufficient number of distinguishing characteristics have been observed; and

prioritizing the distinguishing characteristics for at least some of the known vehicle tracks, wherein said prioritizing comprises defining at least one relationship between a first vehicle track and a second vehicle track in which one of the vehicle tracks is for a vehicle capable of launching an attack weapon and the other of the vehicle tracks is for the launched attack weapon.

14. The method for pre-processing vehicle track characteristics data of claim 13, wherein the vehicle track characteristics data comprises at least one of speed, size, position, IFF status, heading, acceleration, airborne state, in motion state, altitude above or depth below sea level, likelihood of operating within a region of interest, acoustic response characteristics, electromagnetic emanation characteristics, kinematic maneuver characteristics, and electro-optic characteristics.

15. The method for pre-processing vehicle track characteristics data of claim 13, wherein prioritizing the identified distinguishing characteristics further comprises prioritizing the identified distinguishing characteristics and defining the reporting nodes according to a probability analysis.

16. The method for pre-processing vehicle track characteristics data of claim 13, wherein defining relationships between identified distinguishing characteristics comprises defining a relationship between identified distinguishing characteristics representative of a vehicle performing a known tactical maneuver.

17. A computer system comprising:

a database of defined vehicle track signatures for known vehicle tracks;

a first computer processor for executing program instructions;

a first memory, coupled to the first computer processor, storing first program instructions which, when executed by the first computer processor, cause the first computer processor to;

identify distinguishing characteristics for at least some of the known vehicle tracks;

prioritize the identified distinguishing characteristics for at least some of the known vehicle tracks by defining at least one relationship between a first vehicle track and a second vehicle track in which the first vehicle track and the second vehicle track are related; and

generate a graph from the database of defined vehicle track signatures by defining state nodes corresponding to characteristics of the vehicle track signatures and defining links between defined state nodes, the links between the defined state nodes corresponding to defined vehicle track signatures;

a second computer processor coupled to the generated graph;

a second memory, coupled to the second computer processor, storing second program instructions which, when executed by the second computer processor, cause the second computer processor to;

determine if the first unclassified vehicle track and the second unclassified vehicle track are related, by kinematic analysis of the first real-time sensor data and the second real-time sensor data;

responsive to a determination that the first unclassified vehicle track and the second unclassified vehicle track are not related;

perform run-time processing to classify the first vehicle track by passing the first real-time sensor data through the graph until the first unclassified vehicle track is classified; and

perform run-time processing to classify the second vehicle track by passing the second real-time sensor data through the graph until the second unclassified vehicle track is classified;

responsive to a determination that the first unclassified vehicle track and the second unclassified vehicle track are related;

perform run-time processing to classify the first vehicle track by passing the first real-time sensor data and the second real-time sensor data through the graph until the first unclassified vehicle track is classified;

perform run-time processing to classify the second vehicle track by passing the first real-time sensor data and the second real-time sensor data through the graph until the second unclassified vehicle track is classified.

18. The computer system of claim 17, further comprising a third computer processor and third memory storing third program instructions which cause the third computer processor to define the vehicle track signatures for the known vehicle tracks using a probability analysis.

19. The computer system of claim 17, wherein the first memory stores program instructions which cause the first computer processor to define at least one relationship between a first vehicle track and a second vehicle track in which one of the vehicle tracks is for a vehicle capable of launching an attack weapon and the other of the vehicle tracks is for the launched attack weapon.