Passive vehicle classification using low frequency electro-magnetic emanations
First Claim
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 FI, equal to one-half their crankshaft frequency FC times the number of cylinders N in that vehicle, said engine emanating relatively weak EM signals at the harmonic frequencies of the crankshaft frequency FC and relatively strong intensity levels at the harmonic frequencies of the ignition frequency FI, 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 Assignments
0 Petitions
Accused Products
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.
85 Citations
24 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 FI, equal to one-half their crankshaft frequency FC times the number of cylinders N in that vehicle, said engine emanating relatively weak EM signals at the harmonic frequencies of the crankshaft frequency FC and relatively strong intensity levels at the harmonic frequencies of the ignition frequency FI, 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. - View Dependent Claims (2, 3)
-
-
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 FI, 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 FI 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 FI to rotate a crankshaft that is engaged through a transmission to power the vehicle, said ignition frequency FI being equal to one-half the crankshaft frequency Fc times the number of cylinders N in the vehicle, said engine emanating relatively weak EM signals at the harmonic frequencies of the crankshaft frequency Fc and relatively strong intensity levels at the harmonic frequencies of the ignition frequency FI, 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. - View Dependent Claims (6, 7, 8, 9)
-
-
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. - View Dependent Claims (11, 12, 13, 14, 15, 16)
-
-
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. - View Dependent Claims (18, 19, 20, 21, 22, 23, 24)
-
Specification