Passive vehicle classification using low frequency electro-magnetic emanations

• US 6,014,447 A
• Filed: 03/20/1997
• Issued: 01/11/2000
• Est. Priority Date: 03/20/1997
• Status: Expired due to Term

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Eaton, Wilbur W., Kohnen, Kirk
• Primary Examiner(s): Isen, Forester W.
• Assistant Examiner(s): Mei, Xu
• Application Number:
  US08/821,749
• Time in Patent Office: 1,027 Days
• Field of Search: 381/71.1, 381/71.2, 381/71.4, 381/71.14, 381/86, 381/94.2, 381/94.3, 381/FOR 123, 381/FOR 124, 381/56, 340/825.71, 340/825.75, 340/933, 340/941, 340/825.31, 340/825.34, 701/29, 701/32
• US Class Current: 381/86
• CPC Class Codes: G01R 29/0814 Field measurements related to measuring influence on or from apparatus, components or humans EMC, EMI and similar testing in general G01R31/001, e.g. in ESD, EMI, EMC, EMP testing, measuring radiation leakage; detecting presence of micro- or radiowave emitters; dosimetry; testing shielding; measurements related to lightning

Abstract

A system for passively sensing low frequency (5 Hz to 50 kHz) EM emanations from vehicles and classifying them on that basis. A temporal frequency response is computed for the sensed EM emanations. Harmonic, non-harmonic and temporal features are extracted from the response and used to classify the vehicle. In the preferred embodiment, the temporal frequency response is displayed as a time, frequency and intensity plot, from which a technician visually extracts these features. The features can then be used either to determine a "signature" for classifying the vehicle or to determine specific characteristics which, in turn, can be used to classify the vehicle. Alternately, the feature extraction and classification can be performed by an automated classifier.

Claims

1. A method for classifying vehicles that emanate low frequency electro-magnetic (EM) signals, comprising:

a) passively sensing low frequency EM signals emanated from a vehicle;

b) computing the temporal frequency response of the EM signals;

c) extracting features from the temporal frequency response;

d) using the extracted features to determine a set of vehicle characteristics; and

e) using the set of vehicle characteristics to classify the vehicle;

wherein the vehicles comprise a 4-cycle combustion engine having a rotating crankshaft and a plurality of cylinders that fire at ignition frequencies F_I, equal to one-half their crankshaft frequency F_C times the number of cylinders N in that vehicle, said engine emanating relatively weak EM signals at the harmonic frequencies of the crankshaft frequency F_C and relatively strong intensity levels at the harmonic frequencies of the ignition frequency F_I, said harmonic features being extracted by identifying a first harmonic interrelationship between the relatively weak signal emanations and a second harmonic interrelationship between relatively strong signal emanations, said first and second harmonic interrelationships being used to compute the crankshaft and ignition frequencies, respectively, from which the number of cylinders N in the vehicle'"'"'s engine is computed and used to classify the vehicle.

2. The method of claim 1, wherein non-harmonic features are extracted by using a priori information to first select a frequency range for a particular vehicle characteristic and then identifying a relatively strong intensity level in the temporal frequency response that occurs at a non-harmonic frequency in that frequency range.

3. The method of claim 2, wherein the vehicles comprise a 4-cycle combustion engine having a crankshaft that rotates at a crankshaft frequency F_C and an alternator with a d-pole rotor that rotates inside the alternator at an alternator frequency F_A, is equal to d×

F_C ×

r_P where r_P is the engine'"'"'s pulley ratio and known to lie in a given range, said alternator emanating a relatively strong non-harmonic signal at the alternator frequency, said non-harmonic features being extracted by first selecting said frequency range using the pulley ratio'"'"'s given range, selecting the relatively strong intensity level in said frequency range, and assigning its frequency as the alternator frequency F_A, from which the pulley ratio is computed.

4. A method for classifying vehicles that emanate low frequency electro-magnetic (EM) signals, comprising:

a) passively sensing low frequency EM signals emanated from a vehicle;

b) computing the temporal frequency response of the EM signals;

c) extracting features from the temporal frequency response;

d) using the extracted features to determine a set of vehicle characteristics; and

e) using the set of vehicle characteristics to classify the vehicle;

wherein harmonic features are extracted by identifying a plurality of intensity levels in said temporal frequency response that are approximately uniformly spaced apart in frequency;

wherein temporal features are extracted by monitoring the dynamic properties of the extracted harmonic features;

and wherein said vehicle comprises a 4-cycle engine that fires a plurality of cylinders at an ignition frequency F_I, to rotate a crankshaft that is engaged through a transmission to power the vehicle, said engine emanating strong low frequency EM signals at the ignition frequency F_I and its harmonics, said temporal features being extracted by detecting momentary dips in the ignition and harmonic frequencies when the vehicle is accelerating and determining a rate of change for the frequencies, said temporal features being compared against the characteristics of a manual transmission, which include momentary dips in frequency when the clutch is engaged and a rate of change that is proportional to the change in the vehicle'"'"'s speed, and the characteristics of an automatic transmission, which include a smooth frequency response and a rate of change that exceed the change in the vehicle'"'"'s speed, to characterize the vehicle'"'"'s transmission as manual or automatic.

5. A method for classifying vehicles that emanate low frequency electro-magnetic (Em) signals, comprising:

a) passively sensing low frequency EM signals emanated from a vehicle between 5 Hz and 50 kHz;

b) computing the temporal frequency response of the EM signals;

c) displaying the temporal frequency response as a time, frequency, intensity plot on a display;

d) visually extracting features from the time, frequency, intensity plot including harmonic, non-harmonic, and temporal properties;

e) using the extracted features to determine a set of vehicle characteristics;

f) comparing the set of vehicle characteristics to the known characteristics for a plurality of candidate vehicles to classify the vehicle; and

wherein the vehicle comprise a 4-cycle combustion engine that fires a plurality of cylinders at an ignition frequency F_I to rotate a crankshaft that is engaged through a transmission to power the vehicle, said ignition frequency F_I being equal to one-half the crankshaft frequency F_c times the number of cylinders N in the vehicle, said engine emanating relatively weak EM signals at the harmonic frequencies of the crankshaft frequency F_c and relatively strong intensity levels at the harmonic frequencies of the ignition frequency F_I, said harmonic features being extracted by identifying a first harmonic interrelationship between the relative weak signal emanations and a second harmonic interrelationship between relatively strong signal emanations, said first and second harmonic interrelationships being used to compute the crankshaft and ignition frequencies, respectively, from which the number of cylinders N in the vehicle'"'"'s engine is computed and used to classify the vehicle.

6. The method of claim 5, wherein the set of vehicle characteristics corresponds to a signature for the vehicle, said signature being compared to the known signatures for a plurality of candidate vehicles to classify the vehicle.

7. The method of claim 5, wherein the vehicle characteristics in said set each correspond to a specific system in the vehicle that together characterize the vehicle.

8. The method of claim 5, wherein the vehicles further comprise an alternator with a d-pole rotor that rotates inside the alternator at an alternator frequency F_A, is equal to d×

F_C ×

r_P where r_P is the engine'"'"'s pulley ratio and known to lie in a given range, said alternator emanating a relatively strong non-harmonic signal at the alternator frequency, said non-harmonic features being extracted by first selecting said frequency range using the pulley ratio'"'"'s given range, selecting the relatively strong intensity level in said frequency range, and assigning its frequency as the alternator frequency F_A, from which the pulley ratio is computed.

9. The method of claim 5, wherein said temporal features are extracted by detecting momentary dips in the ignition and harmonic frequencies when the vehicle is accelerating and determining a rate of change for the frequencies, said temporal features being compared against the characteristics of a manual transmission, which include momentary dips in frequency when the clutch is engaged and a rate of change that is proportional to the change in the vehicle'"'"'s speed, and the characteristics of an automatic transmission, which include a smooth frequency response and a rate of change that exceed the change in the vehicle'"'"'s speed, to characterize the vehicle'"'"'s transmission as manual or automatic.

10. A system for classifying electro-magnetic (EM) vehicles, comprising:

a vehicle having a 4-cycle combustion engine that emanates low frequency EM signals that have distinctive harmonic interrelationships, strong non-harmonic components, and distinctive temporal properties;

a sensor for passively sensing the low frequency EM signals from the vehicle;

a processor for computing the temporal frequency response of the EM signals and representing them as a time, frequency, intensity plot; and

a display that displays the time, frequency, intensity plot from which a person can observe harmonic interrelationships, strong non-harmonic components and temporal properties as features of the vehicle and use them to classify the vehicle.

11. The system of claim 10, wherein said sensor senses EM signals in a frequency range from approximately 5 Hz to 50 kHz.

12. The system of claim 10, wherein said vehicle'"'"'s 4-cycle combustion engine has a rotating crankshaft and a plurality of cylinders that fire at an ignition frequency F_I equal to one-half the crankshaft frequency F_C times the number of cylinders N, said engine emanating relatively weak EM signals at the harmonic frequencies of the crankshaft frequency F_C and relatively strong EM signals at the harmonic frequencies of the ignition frequency F_I, said display displaying relatively weak intensity levels at the harmonic frequencies of the crankshaft frequency F_C and relatively strong intensity levels at the harmonic frequencies of the ignition frequency F_I from which the person can compute the crankshaft and ignition frequencies and use them to identify the number of cylinders N in the vehicle'"'"'s engine.

13. The system of claim 12, wherein said combustion engine includes an alternator with a d-pole rotor that rotates inside the alternator at an alternator frequency F_A, which lies in a known frequency range with respect to the crankshaft frequency, and is equal to d×

F_C ×

r_P where r_P is the engine'"'"'s pulley ratio, said alternator emanating a relatively strong non-harmonic signal at the alternator frequency, said display displaying a strong non-harmonic intensity level that lies in the known frequency range, from which the person can determine the alternator frequency F_A and compute the pulley ratio r_P.

14. The system of claim 12, wherein said vehicle further comprises a transmission that exhibits different temporal ignition frequency characteristics when accelerating depending on whether it is an automatic or a manual transmission, said display displaying the temporal characteristics of the relatively strong intensity levels at the harmonic frequencies of the ignition frequency F_I, from which the person can detect momentary dips in the harmonic frequencies, which is indicative of a manual transmission, and can determine a rate of change of the harmonic frequencies that exceeds the vehicle'"'"'s acceleration, which is indicative of an automatic transmission.

15. The system of claim 10, wherein said temporal frequency response is resolved into a plurality of frequency bins, further comprising:

a transient noise filter that low pass filters the EM signal in each said frequency bin to reduce the noise associated with transient EM signals.

16. The system of claim 15, further comprising a noise reduction filter that pre-whitens the sensed EM signals so that its power spectral density is approximately flat when the EM sensor is sensing background EM signals in the absence of the vehicle.

17. A system for classifying electro-magnetic (EM) vehicles, comprising:

a sensor for passively sensing low frequency EM signals emanated from a vehicle;

a processor for computing the temporal frequency response of the EM signal; and

a classifier that a) extracts features from the temporal frequency response including harmonic interrelationships of the EM signals, strong non-harmonic EM signals, and temporal properties of strong EM signals, b) uses the features to identify known characteristics of possible vehicles, and c) classifies the vehicle based upon its identified characteristics.

18. The system of claim 17, wherein said EM vehicle emits EM signals in a frequency range from approximately 5 Hz to 50 kHz.

19. The system of claim 17, wherein said vehicle comprises a 4-cycle combustion engine having a rotating crankshaft and a plurality of cylinders that fire at an ignition frequency F_I equal to one-half the crankshaft frequency F_C times the number of cylinders N, said engine emanating relatively weak EM signals at the harmonic frequencies of the crankshaft frequency F_C and relatively strong intensity levels at the harmonic frequencies of the ignition frequency F_I,

said classifier a) extracting a first harmonic interrelationship between the relatively weak signal emanations and a second harmonic interrelationship between relatively strong signal emanations, b) computing the crankshaft frequency F_C from the first harmonic interrelationship, c) computing the ignition frequency F_I from the second harmonic interrelationship, d) using the values of F_C and F_I to identify the number of cylinders N in the vehicle'"'"'s engine, and e) using the number of cylinders N to classify the vehicle.

20. The system of claim 19, wherein said combustion engine includes an alternator with a d-pole rotor that rotates inside the alternator at an alternator frequency F_A, which lies in a known frequency range with respect to the crankshaft frequency, and is equal to d×

F_C ×

r_P where r_P is the engine'"'"'s pulley ratio, said alternator emanating a relatively strong non-harmonic signal at the alternator frequency,

said classifier determining the known frequency range in which the alternator'"'"'s non-harmonic signal can occur for the computed crankshaft frequency, extracting a relatively strong component from the temporal frequency response that occurs at a non-harmonic frequency in said known frequency range and assigning the component'"'"'s frequency as the alternator frequency F_A, computing the pulley ratio r_P, and using it to classify the vehicle.

21. The system of claim 19, wherein said vehicle further comprises a transmission that exhibits different temporal ignition frequency characteristics when accelerating depending on whether it is an automatic or a manual transmission,

said classifier monitoring the temporal response of the ignition frequency F_I and its harmonics to detect momentary dips in those frequencies and to determine a rate of change for the harmonic frequencies, said classifier comparing these characteristics against the known temporal ignition frequency characteristics for automatic and manual transmission vehicles to classify the vehicle'"'"'s transmission as being automatic or manual.

22. The system of claim 17, wherein said temporal frequency response is resolved into a plurality of frequency bins, further comprising:

a transient noise filter that low pass filters the EM signal in each said frequency bin to reduce the noise associated with transient EM signals.

23. The system of claim 22, further comprising a noise reduction filter that pre-whitens the sensed EM signals so that its power spectral density is approximately flat when the EM sensor is sensing background EM signals in the absence of the vehicle.

24. The system of claim 17, further comprising:

a display that displays the temporal frequency response on a time, frequency, intensity plot that displays the harmonic interrelationships of the EM signals, the strong non-harmonic EM signals, and the temporal properties of the strong signals, from which a person can observe these features, identify known characteristics of possible vehicles, and classify the vehicle based upon its identified characteristics to confirm or reject the classification made by the classifier.