Biometric radar system and method for identifying persons and positional states of persons

• US 8,026,840 B2
• Filed: 10/28/2005
• Issued: 09/27/2011
• Est. Priority Date: 10/28/2005
• Status: Active Grant

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Dwelly, Wesley H., Adams, Vinh N.
• Primary Examiner(s): Gregory; Bernarr E
• Application Number:
  US11/261,875
• Publication Number:
  US 20100321229A1
• Time in Patent Office: 2,160 Days
• Field of Search: 342/21, 342/22, 342/27, 342/28, 342/59, 342 82-103, 342/175, 342192-197, 382/100, 382/103, 382115-127, 382/181, 382209-224
• US Class Current: 342/28
• CPC Class Codes: G01S 13/56 for presence detection presence detection using near field arrangements G01V3/00, e.g. G01V3/08, G01V3/12; burglar, theft or intruder alarms with electrical actuation G08B13/22 - G08B13/26
• CPC Class Codes: G01S 13/88 Radar or analogous systems specially adapted for specific applications electromagnetic prospecting or detecting of objects, e.g. near-field detection, G01V3/00
• CPC Class Codes: G01S 7/415 Identification of targets based on measurements of movement associated with the target
• CPC Class Codes: G01S 7/417 involving the use of neural networks

Abstract

A biometric radar system and method for identifying a person'"'"'s positional state are generally described herein. The biometric radar may phase adjust a sequence of radar return signals received through two or more receive antennas to remove at least some phase noise due to the stationary objects. The biometric radar may also segment the phase adjusted radar return signals into a plurality of multi-resolutional Doppler components. Each multi-resolutional Doppler component may be associated with one of a plurality of biometric features. The biometric radar system may also combine and weight the segmented radar returns for each biometric feature to generate weighted classifications for a feature extraction process.

Claims

1. A biometric radar comprising:

phase adjusters to phase adjust a sequence of radar return signals received through two or more receive antennas, the phase-adjusting removing at least some phase noise due to stationary objects;

a signal processor to segment the phase adjusted radar return signals into a plurality of multi-resolutional Doppler components, each multi-resolutional Doppler component being associated with one of a plurality of biometric features; and

a non-linear convolutional neural network to combine and weight the segmented radar returns for each biometric feature to generate weighted classifications for a feature extraction process.

2. The radar system of claim 1 wherein each multi-resolutional Doppler component has a frequency structure associated with one of the biometric features,

wherein the biometric features associated with humans and comprise heartbeat, breathing, postural sway and gait.

3. The radar system of claim 2 wherein the neural network processes the segmented and phase-adjusted radar return signals to identify a person from a non-person.

4. The radar system of claim 2 wherein the weighted classifications correspond to one or more positional states of a person including a person walking, a person talking, a person standing, a person sitting, a person sleeping, multiple persons, and a person with an object.

5. The radar system of claim 4 wherein the signal processor further segments the phase-adjusted radar return signals into non-biometric features including static objects and clutter.

6. The radar system of claim 5 further comprising phase processors to lock a radar range gate on static objects and place other radar range gates on target objects to help force coherency.

7. The radar system of claim 6 wherein the signal processor integrates the radar return signals over differing coherency times for the different of the multi-resolutional Doppler components of the plurality, and

wherein some of the multi-resolutional Doppler components have longer coherency times than other multi-resolutional Doppler components.

8. The radar system of claim 7 wherein the signal processor integrates the multi-resolutional Doppler components associated with the biometric features of heartbeat and breathing for longer coherency times, and

wherein the signal processor integrates the multi-resolutional Doppler components associated with the biometric features of postural sway and gait for shorter coherency times.

9. The radar system of claim 7 wherein the non-linear convolutional neural network is trained with a back-propagation learning process to set weights of the biometric features associated with the positional states.

10. The radar system of claim 9 wherein the neural network comprises a combination of a plurality of non-linear convolutional networks, each of the non-linear convolutional networks being associated with one of the weighted classifications,

wherein the signal processor directs the multi-resolution Doppler components associated with particular biometric features to one or more of the convolutional networks of the plurality based on the classifications, and

wherein the combination of the convolutional networks performs a multi-layer non-linear convolutional process on the directed multi-resolutional Doppler components using non-linear filters to generate weighted positional states for a human target.

11. The radar system of claim 10 wherein the plurality of non-linear convolutional networks are configured to operate in parallel and are weighted to the quantity and quality of information in the multi-resolutional Doppler components.

12. The radar system of claim 10 wherein the neural network is further trained with the back-propagation learning process to set weights of biometric features associated with a particular individual.

13. The radar system of claim 10 further comprising an antenna structure having a transmit antenna and two or more receive antennas, the two or more receive antennas to receive the sequence of radar return signals in azimuth and elevation.

14. The radar system of claim 13 further comprising radar transmitter circuitry to transmit a series of short pulses stepped in frequency and corresponding to the received sequence of radar return signals.

15. The radar system of claim 4 further comprising a feature extractor to perform the feature extraction process, the feature extraction process to generate an output selected from one of the positional states.

16. A method for identifying a person and a person'"'"'s positional state comprising:

phase adjusting a sequence of radar return signals received through two or more receive antennas, the phase-adjusting removing at least some phase noise due to stationary objects;

segmenting the phase adjusted radar return signals into a plurality of multi-resolutional Doppler components, each multi-resolutional Doppler component being associated with one of a plurality of biometric features; and

combining and weighting the segmented radar returns for each biometric feature in a non-linear convolutional neural network to generate weighted classifications for a feature extraction process.

17. The method of claim 16 wherein each multi-resolutional Doppler component has a frequency structure associated with one of the biometric features,

wherein the biometric features associated with humans and comprise heartbeat, breathing, postural sway and gait.

18. The method of claim 17 wherein the neural network processes the segmented and phase-adjusted radar return signals to identify a person from a non-person.

19. The method of claim 17 wherein the weighted classifications correspond to one or more positional states of a person including a person walking, a person talking, a person standing, a person sitting, a person sleeping, multiple persons, and a person with an object.

20. The method of claim 19 further comprising segmenting the phase-adjusted radar return signals into non-biometric features including static objects and clutter.

21. The method of claim 20 further comprising locking a radar range gate on static objects and placing other radar range gates on target objects to help force coherency.

22. The method of claim 21 further comprising integrating the radar return signals over differing coherency times for the different of the multi-resolutional Doppler components of the plurality, and

wherein some of the multi-resolutional Doppler components are integrated over longer coherency times than other multi-resolutional Doppler components.

23. The method of claim 22 wherein the multi-resolutional Doppler components associated with the biometric features of heartbeat and breathing are integrated for longer coherency times, and

wherein the multi-resolutional Doppler components associated with the biometric features of postural sway and gait are integrated for shorter coherency times.

24. The method of claim 22 further comprising training the non-linear convolutional neural network with a back-propagation learning process to set weights of the biometric features associated with the positional states.

25. The method of claim 24 wherein the neural network comprises a combination of a plurality of non-linear convolutional networks, each of the non-linear convolutional networks being associated with one of the weighted classifications,

wherein the method further comprises:

directing the multi-resolution Doppler components associated with particular biometric features to one or more of the convolutional networks of the plurality based on the classifications; and

performing, by the combination of the convolutional networks, a multi-layer non-linear convolutional process on the directed multi-resolutional Doppler components using non-linear filters to generate weighted positional states for a human target.

26. The method of claim 25 further comprising:

configuring the plurality of non-linear convolutional networks to operate in parallel; and

weighting the networks to the quantity and quality of information in the multi-resolutional Doppler components.

27. The method of claim 25 further comprising training the neural network with the back-propagation learning process to set weights of biometric features associated with a particular individual.

28. The method of claim 25 further comprising:

transmitting a series of short pulses stepped in frequency and corresponding to the received sequence of radar return signals; and

receiving, with two or more receive antennas, the sequence of radar return signals in azimuth and elevation.

29. The method of claim 19 further comprising performing the feature extraction process to generate an output selected from one of the positional states.

30. A radar system comprising:

transmitter circuitry to transmit a series of short pulses stepped in frequency;

phase adjusters to phase adjust a sequence of radar return signals received through two or more receive antennas, the phase-adjusting removing at least some phase noise due to stationary objects, the transmitted series of short pulses corresponding to the received sequence of radar return signals;

a signal processor to segment the phase adjusted radar return signals into a plurality of multi-resolutional Doppler components, each multi-resolutional Doppler component being associated with one of a plurality of biometric features; and

a non-linear convolutional neural network to combine and weight the segmented radar returns for each biometric feature to generate weighted classifications for a feature extraction process,

wherein each multi-resolutional Doppler component has a frequency structure associated with one of the biometric features, and

wherein the biometric features associated with humans and comprise heartbeat, breathing, postural sway and gait.

31. The radar system of claim 30 wherein the weighted classifications correspond to one or more positional states of a person including a person walking, a person talking, a person standing, a person sitting, a person sleeping, multiple persons, and a person with an object, and

wherein the signal processor further segments the phase-adjusted radar return signals into non-biometric features including static objects and clutter.

32. The radar system of claim 31 further comprising:

phase processors to lock a radar range gate on static objects and place other radar range gates on target objects to help force coherency; and

a feature extractor to perform the feature extraction process, the feature extraction process to generate an output selected from one of the positional states,

wherein the signal processor integrates the radar return signals over differing coherency times for the different of the multi-resolutional Doppler components of the plurality, and

wherein some of the multi-resolutional Doppler components have longer coherency times than other multi-resolutional Doppler components.

33. The radar system of claim 32 wherein the neural network comprises a combination of a plurality of non-linear convolutional networks, each of the non-linear convolutional networks being associated with one of the weighted classifications,

wherein the signal processor directs the multi-resolution Doppler components associated with particular biometric features to one or more of the convolutional networks of the plurality based on the classifications, and

wherein the combination of the convolutional networks performs a multi-layer non-linear convolutional process on the directed multi-resolutional Doppler components using non-linear filters to generate weighted positional states for a human target.