Frequency domain processing of Doppler signals in a traffic monitoring system

• US 5,912,822 A
• Filed: 10/15/1996
• Issued: 06/15/1999
• Est. Priority Date: 06/01/1994
• Status: Expired due to Term

• Current Assignee: Amtech Systems LLC (Roper Technologies, Inc.)
• Original Assignee: American Traffic Systems, Inc. (Roper Technologies, Inc.)
• Inventors: Mee, Gary L., Davis, Clint A.
• Primary Examiner(s): Cosimano, Edward R.
• Application Number:
  US08/730,736
• Time in Patent Office: 973 Days
• Field of Search: 340/936, 340/937, 342/66, 342/104, 346/330, 364/438, 364/565
• US Class Current: 702/143
• CPC Class Codes: G01S 13/581 using transmission of interrupted pulse modulated waves and based upon the Doppler effect resulting from movement of targets
• CPC Class Codes: G01S 13/92 for velocity measurement
• CPC Class Codes: G01S 7/415 Identification of targets based on measurements of movement associated with the target

Abstract

Frequency domain processing is used to determine rapidly whether false signals are present, and to improve the accuracy of speed measurements by rejecting speed measurements that could be inaccurate due to the presence of false signals. A Doppler signal responsive to speed of a moving vehicle is generated by a Doppler radar transceiver. A digital computer transforms the Doppler signal into a frequency domain signal such as an energy spectrum, and the speed of the moving vehicle is computed from the weighted arithmetic mean of the energy spectrum. The mean value, however, is rejected as an indication of speed if the variance of the energy spectrum from the mean exceeds a threshold, or if a differential of the spectrum with respect to frequency exceeds a threshold. The frequency domain processing of the present invention can also be used to reject certain bands of frequencies in the Doppler signal, such as bands of frequencies that differ from the mean value by more than a threshold, and the mean value can then be recomputed to provide an even better indication of the speed of the moving vehicle. Moreover, speed samples computed from successive frames of spectrum can be averaged, and the average value can be rejected as an indication of speed if the variance of the speed samples about the average value exceeds a threshold, or if the difference between successive speed samples exceeds a threshold.

Claims

1. A method of determining a speed of a moving vehicle, said method comprising the steps of:

(a) operating a Doppler radar transceiver to generate a Doppler signal responsive to said speed of said moving vehicle;

(b) operating a digital computer to transform said Doppler signal into a frequency domain signal, and to determine said speed of said moving vehicle by computing a statistic of said frequency domain signal, wherein the frequency domain signal indicates a distribution of energy of the Doppler signal as a function of frequency.

2. The method as claimed in claim 1, wherein said digital computer transforms said Doppler signal into a frequency domain signal by computing a fast Fourier transform of the Doppler signal.

3. The method as claimed in claim 1, further comprising the step of computing a differential of said frequency domain signal with respect to frequency, and comparing said differential to a threshold, and rejecting said mean value as a correct indication of said speed of said vehicle when said differential exceeds said threshold.

4. The method as claimed in claim 1, wherein said frequency domain signal indicates a distribution of energy of said Doppler signal as a function of frequency, and said statistic is a moment of said distribution of energy of said Doppler signal as a function of frequency.

5. The method as claimed in claim 4, wherein said moment is a weighted arithmetic mean of said distribution of energy of said Doppler signal as a function of frequency.

6. The method as claimed in claim 4, wherein said moment is a variance about a mean of said distribution of energy of said Doppler signal as a function of frequency.

7. The method as claimed in claim 1, wherein said statistic is a mean value of said frequency domain signal, and said mean value indicates said speed of said vehicle.

8. The method as claimed in claim 7, further comprising the steps of computing a variance of said frequency domain signal about said mean value, comparing said variance to a threshold, and rejecting said mean value as a correct indication of said speed of said vehicle when said variance exceeds said threshold.

9. The method as claimed in claim 7 further comprising the steps of identifying portions of said frequency-domain signal having frequencies that differ from said mean value by more than a threshold, and recomputing said mean value while excluding the identified portions of said frequency-domain signal, so that the recomputed mean value is an even better indication of the speed of said moving vehicle.

10. A method of monitoring moving vehicles, said method comprising the steps of:

(a) operating a Doppler radar transceiver to generate a Doppler signal responsive to speeds of said moving vehicles; and

(b) operating a digital computer to transform said Doppler signal into a frequency domain signal, wherein the frequency domain signal indicates a distribution of energy of the Doppler signal as a function of frequency, and wherein the frequency domain signal has successive portions, and for each portion,

(i) to compute a mean value of said each portion of said frequency domain signal,

(ii) to compute a variance of said each portion of said frequency domain signal from said mean value, and

(iii) to compare said variance to a threshhold, and when said variance is greater than said threshhold, to reject said each portion of said frequency domain signal as an indication of a speed of any of said vehicles, and when said variance is less than said threshhold, to compute from said each portion of said frequency domain signal a measurement of speed of one of said signals.

11. The method as claimed in claim 10, wherein said step of transforming is performed by converting said Doppler signal to digital samples, and performing a discrete Fourier transform on successive groups of said digital samples, each of said successive groups of digital samples being transformed to produce a corresponding one of said portions of said frequency domain signal.

12. The method as claimed in claim 11, wherein each portion of said frequency domain signal comprises a series of energy values for respective frequency intervals.

13. The method as claimed in claim 10, which includes the step of averaging a plurality of speed samples from a plurality of said successive portions of said Doppler signal in order to compute said measurement of speed of one of said vehicles.

14. The method as claimed in claim 13, further including the step of computing a variance of said speed samples from said measurement of speed of one of said vehicles, and comparing said variance of said speed samples to a threshold value in order to validate said measurement of speed of one of said vehicles.

15. The method as claimed in claim 13, further including the step of computing differences between successive speed samples, and comparing said differences to a threshold in order to validate said measurement of speed of one of said vehicles.

16. A traffic monitoring system comprising:

(a) a Doppler radar transceiver for emitting a radar beam and generating a Doppler signal responsive to speed of a vehicle when said vehicle passes through said radar beam;

(b) an analog-to-digital converter connected to said Doppler radar transceiver for converting said Doppler signal into a series of digital samples which are frequency domain signals that indicate a distribution of energy of the Doppler signl as a function of frequency; and

(c) digital computer means connected to said analog-to-digital converter for computing a discrete Fourier transform of successive groups of said series of digital samples to produce a series of frames of spectrum, and from said series of frames of spectrum, to detect the presence of said vehicle in said radar beam and to compute te speed of said vehicle.

17. The traffic monitoring system as claimed in claim 16, wherein said digital computer means includes means for computing a mean frequency of said spectrum, means for computing a variance of said spectrum about said mean frequency, means for comparing said variance of said spectrum to a threshold, and means responsive to said variance being less than said threshold for verifying that said vehicle has a speed indicated by said mean frequency.

18. The traffic monitoring system as claimed in claim 16, wherein said digital computer means includes a first central processor connected to said analog-to-digital converter for computing said discrete Fourier transform, and a second central processor connected to said first central processor for detecting the presence of said vehicle in said radar beam and for computing the speed of said vehicle from said series of frames of spectrum.

19. The traffic monitoring system as claimed in claim 16, wherein said Doppler radar transceiver includes a pair of detector diodes for generating a pair of channels for said Doppler signal, and wherein said digital computer means includes means for computing discrete Fourier transforms for each channel and combining the discrete Fourier transforms for the channels to produce said series of frames of spectrum.

20. The traffic monitoring system as claimed in claim 16, wherein said digital computer means includes means for computing a speed sample for each frame of spectrum, means for computing an average of a series of said speed samples, means for computing a variance of said series of said speed samples about said average, and means for comparing said variance of said series of said speed samples to a threshold for verifying that said average represents the speed of said moving vehicle.