System and method for constructing filters for detecting signals whose frequency content varies with time

• US 5,574,639 A
• Filed: 10/12/1994
• Issued: 11/12/1996
• Est. Priority Date: 10/12/1994
• Status: Expired due to Term

• Current Assignee: National Instruments Corporation (Emerson Electric Company), Los Alamos National Security LLC (Government of the United States of America)
• Original Assignee: National Instruments Corporation (Emerson Electric Company)
• Inventors: Qian, Shie, Dunham, Mark E.
• Primary Examiner(s): Malzahn, David H.
• Application Number:
  US08/322,053
• Time in Patent Office: 762 Days
• Field of Search: 364/724.01, 364/724.07, 364/724.08, 364/724.09, 364/724.11, 364/724.16
• US Class Current: 708/300
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06F 2218/12 Classification; Matching

Abstract

A system and method for constructing a bank of filters which detect the presence of signals whose frequency content varies with time. The present invention includes a novel system and method for developing one or more time templates designed to match the received signals of interest and the bank of matched filters use the one or more time templates to detect the received signals. Each matched filter compares the received signal x(t) with a respective, unique time template that has been designed to approximate a form of the signals of interest. The robust time domain template is assumed to be of the order of w(t)=A(t)cos{2πφ(t)} and the present invention uses the trajectory of a joint time-frequency representation of x(t) as an approximation of the instantaneous frequency function {φ'"'"'(t). First, numerous data samples of the received signal x(t) are collected. A joint time frequency representation is then applied to represent the signal, preferably using the time frequency distribution series (also known as the Gabor spectrogram). The joint time-frequency transformation represents the analyzed signal energy at time t and frequency ƒ, P(t,f), which is a three-dimensional plot of time vs. frequency vs. signal energy. Then P(t,f) is reduced to a multivalued function f(t), a two dimensional plot of time vs. frequency, using a thresholding process. Curve fitting steps are then performed on the time/frequency plot, preferably using Levenberg-Marquardt curve fitting techniques, to derive a general instantaneous frequency function φ'"'"'(t) which best fits the multivalued function f(t), a trajectory of the joint time-frequency domain representation of x(t). Integrating φ'"'"'(t) along t yields φ(t), which is then inserted into the form of the time template equation. A suitable amplitude A(t) is also preferably determined. Once the time template has been determined, one or more filters are developed which each use a version or form of the time template.

Claims

1. A method for constructing one or more filters to detect signals having a frequency content which varies with time, comprising the steps of:

collecting data regarding the signals;

computing a joint time-frequency representation of the signals;

mapping the joint time-frequency representation to a multi-valued function;

curve-fitting the multi-valued function to derive an instantaneous frequency function φ

'"'"'(t);

integrating the instantaneous frequency function φ

'"'"'(t) to obtain φ

(t);

determining a time template equation using said function φ

(t);

constructing one or more filters which receive the signals and compare the signals with one or more terms of said time template equation, wherein said one or more filters detect the signals.

2. The method of claim 1, wherein said step of computing a joint time-frequency representation of the signals comprises computing a joint time-frequency distribution series of the signals.

3. The method of claim 1, wherein said step of computing a joint time-frequency representation of the signals comprises computing a Short Time Fourier Transform (STFT) of the signals.

4. The method of claim 1, wherein said step of computing a joint time-frequency representation of the signals comprises computing a 3-dimensional representation of the signals of time vs. frequency vs. signal energy.

5. The method of claim 4, wherein said step of mapping the joint time-frequency representation to a multi-valued function comprises mapping the joint time-frequency to a 2-dimensional multi-valued function of time vs. frequency.

6. The method of claim 1, wherein said step of mapping comprises applying a thresholding process to said joint time-frequency representation of the signals to map said joint time-frequency representation of the signals to the multi-valued function.

7. The method of claim 1, wherein said step of mapping comprises determining the relative energy of each sample of said joint time-frequency representation above a threshold value;

wherein said step of curve-fitting uses said relative energy of each sample to set the appropriate weights in a non-linear curve fitting algorithm; and

wherein said step of curve-fitting uses said non-linear curve fitting algorithm.

8. The method of claim 1, wherein said step of curve-fitting uses Levenberg-Marquardt curve-fitting techniques.

9. The method of claim 1, wherein said step of curve-fitting uses one or more techniques from the group consisting of Simplex, Levenberg-Marquardt, and Conjugate Gradient.

10. The method of claim 1, wherein said time template equation is:

space="preserve" listing-type="equation">A(t) cos {2π

φ

(t)}.

11. The method of claim 1, further comprising the steps of:

determining an amplitude A(t); and

incorporating said amplitude A(t) into said time template equation.

12. The method of claim 11, wherein said amplitude A(t) comprises a unit step function.

13. The method of claim 1, wherein said step of constructing one or more filters comprises constructing a bank comprising a plurality N of filters, wherein each of said plurality N of filters compares the signals with a respective form of said time template equation to detect the signals.

14. The method of claim 13, wherein said step of constructing one or more filters comprises constructing a bank comprising a plurality N of filters, wherein each of said plurality N of filters compares the signals with a unique form of said time template equation that varies according to one or more parameters;

wherein said plurality N of filters each compare the signals with one of a corresponding plurality N of different forms of said time template equation.

15. The method of claim 14, wherein said plurality N of different forms of said time template equation are designed to match various possible forms of the signals.

16. The method of claim 1, wherein said step of constructing one or more filters comprises constructing a bank of transversal filters.

17. The method of claim 1, wherein said step of constructing one or more filters comprises constructing a bank of matched filters.

18. A method for constructing one or more filters to detect signals having a frequency content which varies with time, comprising the steps of:

collecting data regarding the signals;

computing a joint time-frequency representation of the signals;

calculating an instantaneous frequency function φ

'"'"'(t) using said joint time-frequency representation of the signals;

integrating the instantaneous frequency function φ

'"'"'(t) to obtain φ

(t);

determining a time template equation using said function φ

(t);

constructing one or more filters which receive the signals and compare the signals with one or more forms of said time template equation, wherein said one or more filters detect the signals.

19. The method of claim 18, wherein said step of computing a joint time-frequency representation of the signals comprises computing a joint time-frequency distribution series of the signals.

20. The method of claim 18, wherein said step of computing a joint time-frequency representation of the signals comprises computing a Short Time Fourier Transform (STFT) of the signals.

21. The method of claim 18, wherein said step of computing a joint time-frequency representation of the signals comprises computing a 3-dimensional representation of the signals of time vs. frequency vs. signal energy.

22. The method of claim 18, wherein said step of calculating comprises determining the frequency mean values at each time sample of said joint time-frequency representation with all energy values below a certain minimal value set to zero to calculate said instantaneous frequency function φ

'"'"'(t).

23. The method of claim 18,

wherein said joint time-frequency representation includes many frequency values at each time sample with significant signal energy;

and

wherein said step of calculating comprises finding the point of maximum energy change with respect to frequency on either side of the frequency value at each time sample having maximum energy, and then assigning the mean of these two values as the instantaneous frequency value at this location.

24. The method of claim 18, wherein said step of calculating an instantaneous frequency function comprises:

mapping the joint time-frequency representation to a multi-valued function; and

curve-fitting the multi-valued function to derive an instantaneous frequency function φ

'"'"'(t).

25. The method of claim 24, wherein said step of mapping the joint time-frequency representation to a multi-valued function comprises mapping the joint time-frequency to a 2-dimensional multi-valued function of time vs. frequency.

26. The method of claim 24, wherein said step of mapping comprises applying a thresholding process to said joint time-frequency representation of the signals to map said joint time-frequency representation of the signals to the multi-valued function.

27. The method of claim 24, wherein said step of curve-fitting uses Levenberg-Marquardt curve-fitting techniques.

28. The method of claim 18, wherein said time template equation is:

space="preserve" listing-type="equation">A(t) cos {2π

φ

(t)}.

29. The method of claim 18, further comprising the steps of:

determining an amplitude A(t); and

incorporating said amplitude A(t) into said time template equation.

30. The method of claim 29, wherein said amplitude A(t) comprises a unit step function.

31. The method of claim 18, wherein said step of constructing one or more filters comprises constructing a bank comprising a plurality N of filters, wherein each of said plurality N of filters compares the signals with a respective form of said time template equation to detect the signals.

32. The method of claim 31, wherein said step of constructing one or more filters comprises constructing a bank comprising a plurality N of filters, wherein each of said plurality N of filters compares the signals with a unique form of said time template equation that varies according to one or more parameters;

wherein said plurality N of filters, each compare the signals with one of a corresponding plurality N of different forms of said time template equation.

33. The method of claim 18, wherein said plurality N of different forms of said time template equation are designed to match various possible forms of the signals.

34. The method of claim 18, wherein said step of constructing one or more filters comprises constructing a bank of transversal filters.

35. The method of claim 18, wherein said step of constructing one or more filters comprises constructing a bank of matched filters.