Virtual sensor for data and sensor fusion

• US 7,047,161 B1
• Filed: 04/15/2004
• Issued: 05/16/2006
• Est. Priority Date: 06/10/2003
• Status: Active Grant

• Current Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Original Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Inventors: Friesel, Mark A.
• Primary Examiner(s): Bui, Bryan
• Application Number:
  US10/826,241
• Time in Patent Office: 761 Days
• Field of Search: 702/104, 702/182, 702/188, 702/189, 702/190, 702/191, 455/226.3, 356/73.1, 356/106
• US Class Current: 702/189
• CPC Class Codes: G06F 18/25 Fusion techniques

Abstract

A plurality of sensors observe an object, and the raw sensor data is processed to produce evidence signals representative of characteristics which may be used to classify the object as to type. The evidence from the plurality of sensors is fused to generate fused or combined evidence. Thus, the fused evidence is equivalent to signals produced by a virtual sensor. The fused evidence is applied to a taxonomic classifier to determine the object type.

Claims

1. A method for fusing information from plural sources, said method comprising the steps of:

observing an object with plural sensors including at least first and second sensors, each of which (a) evaluates evidence or information and (b) based on said evidence or information, assigns a taxonomic (type) classification to its observations of said object;

fusing said evidence from at least said first and second sensors to produce compound evidence; and

assigning a classification based on said compound evidence.

2. A method according to claim 1, wherein said classification based on compound evidence is a taxonomic classification.

3. A method according to claim 1, wherein said sensors are spatially separated.