Constant false probability data fusion system

• US 5,661,666 A
• Filed: 11/06/1992
• Issued: 08/26/1997
• Est. Priority Date: 11/06/1992
• Status: Expired due to Fees

• Current Assignee: the united states of america as represented by the secretary of the navy
• Original Assignee: the united states of america as represented by the secretary of the navy
• Inventors: Pawlak, Robert J.
• Primary Examiner(s): Trammell, James P.
• Assistant Examiner(s): Pipala, Edward J.
• Application Number:
  US07/972,339
• Time in Patent Office: 1,754 Days
• Field of Search: 364/516, 364/571.01, 364/571.07, 364/715.06
• US Class Current: 702/182
• CPC Class Codes: G01S 13/87 Combinations of radar systems, e.g. primary radar and secondary radar
• CPC Class Codes: G01S 7/2923 based on data belonging to a number of consecutive radar periods
• CPC Class Codes: G06F 18/256 of results relating to different input data, e.g. multimodal recognition
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes

Abstract

A system for determining whether a given phenomenon has occurred based on ltiple sensor decisions is provided. Each sensor samples input data and attempts to decide if the given phenomenon exists. These sensor decisions are provided to the data fusion processor. The data fusion processor uses a sum of the sensor decisions multiplied by a logarithmic gain indicating the relative reliability of each sensor to generate a test existence metric. The test existence metric is compared to two threshold limits. The results of this comparison are used to provide a final decision indicating the existence of the given phenomenon. An optimization is used to determine the threshold values used in the threshold table to guarantee that the false alarm rate and the data fusion processor is constant, even in cases where data from some sensor is missing.

Claims

1. In a system for determining presence of a test phenomenon within an environment being monitored including:

a plurality of sensors receiving input data to produce outputs representative of said environment;

data fusion processor means for generating a weighted decision value from the input data based on relative reliability of said sensors;

means for generating a look-up table from which thresholds are selected based on said weighted decision value;

means for multiplying said outputs of the sensors to determine said relative reliability of the sensors;

processing means for summing the outputs in sequence with said multiplying thereof; and

comparator means rendered operative in response to said selected thresholds for generating a final decision reflecting said presence of the test phenomenon from the outputs of the sensors after said multiplying and summing thereof in sequence.

2. A system for determining the presence of the test phenomenon as in claim 1 wherein said plurality of sensors comprise an array of radar sensors.

3. A system for determining the existence of the test phenomenon as in claim 1 wherein said means for generating the lookup table is a combinatorial optimization algorithm with the following constrants:

probability of false alarm for said system is held constant;

probability of detection for said system is maximized;

probability of requesting additional measurements based on noise for said system is held constant; and

probability of detection based on the additional requested measurements for said system is maximized.

4. A system for determining the existence of the test phenomenon as in claim 1 wherein said means for generating a threshold lookup table is performed one time only thereby providing increased speed for data fusion.

5. A system for determining the existence of the test phenomenon as in claim 1 wherein said means for generating a threshold lookup table is performed periodically thereby providing adaptivity to variations in sensor performance.

6. A system for determining the existence of the test phenomenon as in claim 1 wherein said means for generating the threshold values from said weighted decision values is given by using a vector U to compute an index.

7. In a system for determining presence of a test phenomenon within an environment being monitored including:

a plurality of sensors receiving input data to produce outputs representative of said environment;

weighting means for generating thresholds based on relative reliability of the sensors;

means for multiplying said outputs of the sensors to determine said relative reliability of the sensors;

processing means for summing the outputs in sequence with said multiplying thereof;

comparator means rendered operative in response to said thresholds for generating a final decision reflecting said presence of the test phenomenon from the outputs of the sensors after said multiplying and summing thereof in sequence.

8. In a system for determining presence of a test phenomenon within an environment being monitored by a plurality of sensors receiving input data to produce outputs representative of said environment, a method including the steps of:

computing a look-up table from a collection of the outputs from the sensors;

limiting said collection of the outputs to probabilities of false alarm and remeasurement because of noise maintained at constant rates;

further limiting said collection of the outputs to maximize detection of the test phenomenon and said remeasurement;

computing thresholds from the look-up table selected from two values, one of which is equal to or greater than the other;

multiplying said outputs of the sensors to determine relative reliability of the sensors;

summing the multiplied outputs in sequence; and

utilizing the computed thresholds to generate a final decision reflecting said presence of the test phenomenon.