Correlation match filter for passive signal detection

• US 4,905,209 A
• Filed: 05/27/1988
• Issued: 02/27/1990
• Est. Priority Date: 05/27/1988
• Status: Expired due to Term

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Honeywell Incorporated (Honeywell International Inc.)
• Inventors: Huang, Stanley S.
• Primary Examiner(s): Tarcza, Thomas H.
• Assistant Examiner(s): Swann, Tod
• Application Number:
  US07/199,507
• Time in Patent Office: 641 Days
• Field of Search: 367/7, 367/100, 367/135, 367/124-126, 367/901, 342/145, 342/194, 364/724.11, 364/724.12, 364/728.01, 364/728.05
• US Class Current: 367/125
• CPC Class Codes: G01V 1/001 Acoustic presence detection

Abstract

A passive detection technique is based on correlation characteristics of split beam inputs with given relative rates between a detection sensor platform and a target source. The correlation coefficient (10, 12) between two phase offset beams is first computed and, at the same time, the relative angular position between the sensor line-of-sight and the target source are estimated (14) to establish the pattern matching function. The correlation operations are carried out for both real and imaginary channels with a four quadrant multiplier (20, 22, 24, and 26). The outputs of the multiplier are summed (28, 30) to obtain real and imaginary channel outputs which are smoothed over a period of time in integrators (32, 34) to eliminate undesirable correlation patterns. The outputs of the integrators are independently thresholded (36, 40) and square summed (38, 42 and 44) and then threshold (46) to provide an indication of target detection.

Claims

1. A passive target detector employing a correlation match filter comprising:

beam forming means for forming two passive phase offset beams and dynamically sweeping said beats across a target to generate first and second phase offset received signals;

complex correlation means connected to said beam forming means for generating a complex cross-correlation coefficient of said first and second phase offset signals, said coefficient having real and imaginary parts;

replica function generating means for generating a replica signal in the form of complementary functions of a relative angle between a sensor line-of-sight and a source of signal to be detected;

four quadrant multiplier means for multiplying said real and imaginary parts of said coefficient by said complementary functions to generate a plurality of partial products;

summing and integrating means for accumulating over time the partial products from said four quadrant multiplier means to provide real and imaginary filter outputs; and

output means connected to said summing and integrating means for generating a detection output if a target has been detected.

2. The correlation match filter recited in claim 1 wherein said output means includes threshold means connected to said summing and integrating means for providing detection outputs when either said real or imaginary filter outputs exceeds a predetermined value.

3. The correlation match filter recited in claim 2 wherein said comparing means further includes:

square summing means for squaring said real and imaginary filter outputs and summing the squared outputs; and

second threshold means connected to said square summing means for providing a detection output when the summed squared outputs exceeds a predetermined value.

4. A method for detecting an acoustic source in a correlated or uncorrelated noise background comprising the steps of:

forming two passive phase offset beams which are dynamically swept across a target to generate first and second received signals;

continuously cross-correlating said first and second signals to produce a complex correlation coefficient having real and imaginary parts;

generating complementary functions of a relative angle between a sensor line-of-sight and a source of signal to be detected;

multiplying real and imaginary parts of said complex correlation coefficient by said complementary functions to produce real and imaginary product signals; and

accumulating over time said real and imaginary product signals to produce first and second accumulated signals to determine if a target has been detected.

5. The method recited in claim 4 further comprising the step of thresholding said first and second accumulated signals to produce a detection output if either of said first or second detection signals exceed a predetermined level.

6. The method recited in claim 5 further comprising the steps of:

squaring said first and second accumulated signals to produce squared outputs;

summing said squared outputs; and

thresholding the summed squared outputs to produce a detection output if the summed squared outputs exceeds a predetermined level.