Information processing system for classifying and/or tracking an object

• US 8,010,658 B2
• Filed: 02/07/2008
• Issued: 08/30/2011
• Est. Priority Date: 02/09/2007
• Status: Active Grant

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Guillochon, James, Khosla, Deepak, Choe, Howard C.
• Primary Examiner(s): Barqadle; Yasin
• Assistant Examiner(s): Fan; Hua
• Application Number:
  US12/027,588
• Publication Number:
  US 20080235318A1
• Time in Patent Office: 1,300 Days
• Field of Search: 709/224
• US Class Current: 709/224
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/62 relating to a temporal dimension, e.g. time-based feature extraction; Pattern tracking

Abstract

According to one embodiment, a computing system includes a computing node coupled to a number of sensors. The sensors are operable to generate records from received information and transmit these records to the computing node. The computing node is operable to bind the plurality of records in a plurality of classifications using a multiple level classifier such that each classification has a differing level of specificity.

Claims

1. A distributed computer system comprising:

a plurality of sensors operable to receive information about an object and generate records from the received information;

a first computing node coupled to the plurality of sensors and operable to;

receive a first plurality of records from the plurality of sensors;

apply the first plurality of records to a first one or more confusion matrices;

based on the first one or more confusion matrices, classify the object into a plurality of categories in each of a plurality of classifications, each of the plurality of classifications having a differing level of specificity, the plurality of classifications having a hierarchical classification structure; and

based on the first one or more confusion matrices, generate a belief state that includes, for each of the plurality of categories of each of the plurality of classifications, a likelihood that the object belongs in that category;

a second computing node coupled to the plurality of sensors and the first computing node, the second computing node operable to;

receive a second plurality of records generated by the plurality of sensors at a specified period of time after the first plurality of records were generated;

apply the second plurality of records to a second one or more confusion matrices; and

based on the second one or more confusion matrices, update the belief state; and

a third computing node coupled to the plurality of sensors and the first and second computing nodes, the third computing node operable to;

receive the first plurality of records and the second plurality of records, and

determine, using an information form of a Kalman filter, an estimated trajectory of the object according to the first plurality of records and the second plurality of records.

2. Code embodied in a non-transitory computer-readable storage media, when executed by a computer operable to perform at least the following:

receive, by a first computing node, a first plurality of records from a plurality of sensors coupled to the first computing node through a network, the plurality of sensors operable to generate the first plurality of records from information about an object;

apply the first plurality of records to a first one or more confusion matrices;

based on the first one or more confusion matrices, classify the object into a plurality of categories in each of a plurality of classifications, each of the plurality of classifications having a differing level of specificity, the plurality of classifications having a hierarchical classification structure;

based on the first one or more confusion matrices, generate a belief state that includes, for each of the plurality of categories of each of the plurality of classifications, a likelihood that the object belongs in that category;

receive, by a third computing node, the first plurality of records and a second plurality of records generated by the plurality of sensors at a specified period of time after the first plurality of records were generated, the third computing node communicating with the first computing node through the network;

determine an estimated trajectory of the object according to the first plurality of records and the second plurality of records; and

associate the first plurality of records with the second plurality of records using a scoring algorithm.

3. The code of claim 2, further operable to:

receive, by a second computing node, a second plurality of records generated by the plurality of sensors at a specified period of time after the first plurality of records were generated, the second computing node communicating with the first computing node through the network;

apply the second plurality of records to a second one or more confusion matrices; and

based on the second one or more confusion matrices, update the belief state.

4. The code of claim 2, further operable to determine the estimated trajectory of the object using an information form of the Kalman filter.

5. A computer system comprising:

a plurality of sensors operable to receive information about an object, the plurality of sensors operable to generate records from the received information;

a first computing node coupled to the plurality of sensors through a network and operable to;

receive a first plurality of records from the plurality of sensors;

apply the first plurality of records to a first one or more confusion matrices;

based on the first one or more confusion matrices, classify the object into a plurality of categories in each of a plurality of classifications, each of the plurality of classifications having a differing level of specificity, the plurality of classifications having a hierarchical classification structure; and

based on the first one or more confusion matrices, generate a belief state that includes, for each of the plurality of categories of each of the plurality of classifications, a likelihood that the object belongs in that category; and

a third computing node coupled to the first computing node and the plurality of sensors through the network, the third computing node operable to;

receive the first plurality of records and a second plurality of records generated by the plurality of sensors at a specified period of time after the first plurality of records were generated; and

determine an estimated trajectory of the object according to the first plurality of records and the second plurality of records, wherein the third computing node is further operable to associate the first plurality of records with the second plurality of records using a scoring algorithm.

6. The computer system of claim 5, further comprising a second computing node coupled to the first computing node and the plurality of sensors through the network, the second computing node operable to:

receive a second plurality of records generated by the plurality of sensors at a specified period of time after the first plurality of records were generated;

apply the second plurality of records to a second one or more confusion matrices; and

based on the second one or more confusion matrices, update the belief state.

7. The computer system of claim 5, wherein the third computing node is further operable to determine the estimated trajectory of the object using an information form of the Kalman filter.

8. The computer system of claim 5, wherein the first computing node is further operable to classify the object into the plurality of categories in each of the plurality of classifications by recursively combining the first one or more confusion matrices.

9. A method comprising:

receiving, by a first computing node, a first plurality of records from a plurality of sensors coupled to the first computing node through a network, the plurality of sensors operable to generate the first plurality of records from information about an object;

applying, by the first computer node, the first plurality of records to a first one or more confusion matrices;

based on the first one or more confusion matrices, classifying, by the first computer node, the object into a plurality of categories in each of a plurality of classifications, each of the plurality of classifications having a differing level of specificity, the plurality of classifications having a hierarchical classification structure;

based on the first one or more confusion matrices, generating, by the first computer node, a belief state that includes, for each of the plurality of categories of each of the plurality of classifications, a likelihood that the object belongs in that category;

receiving, by a third computing node the first plurality of records and a second plurality of records generated by the plurality of sensors at a specified period of time after the first plurality of records were generated, the third computing node communicating with the first computing node through the network;

determining, by the third computing node, an estimated trajectory of the object according to the first plurality of records and the second plurality of records; and

associating, by the third computing node, the first plurality of records with the second plurality of records using a scoring algorithm.

10. The method of claim 9, further comprising:

receiving, by a second computing node a second plurality of records generated by the plurality of sensors at a specified period of time after the first plurality of records were generated, the second computing node communicating with the first computing node through the network;

applying, by the second computing node, the second plurality of records to a second one or more confusion matrices; and

based on the second one or more confusion matrices, updating, by the second computing node, the belief state.

11. The method of claim 9, further comprising determining, by the third computing node, the estimated trajectory of the object using an information form of the Kalman filter.