Non-spatial speech detection system and method of using same

• US 8,935,164 B2
• Filed: 05/02/2012
• Issued: 01/13/2015
• Est. Priority Date: 05/02/2012
• Status: Active Grant

• Current Assignee: HL Klemove Corporation
• Original Assignee: Gentex Corporation
• Inventors: Turnbull, Robert R., Bryson, Michael A.
• Primary Examiner(s): GUERRA-ERAZO, EDGAR X
• Application Number:
  US13/462,506
• Publication Number:
  US 20130297305A1
• Time in Patent Office: 986 Days
• Field of Search: 704/216, 704/217, 704/218, 704/226, 704/227, 704/228, 704/233, 704/275
• US Class Current: 704/233
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 2021/02166 Microphone arrays; Beamforming
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: H04R 1/406 microphones
• CPC Class Codes: H04R 2430/20 Processing of the output signals of the acoustic transducers of an array for obtaining a desired directivity characteristic H04R2203/12 takes precedence
• CPC Class Codes: H04R 2499/13 Acoustic transducers and sound field adaptation in vehicles

Abstract

A non-spatial speech detection system includes a plurality of microphones whose output is supplied to a fixed beamformer. An adaptive beamformer is used for receiving the output of the plurality of microphones and one or more processors are used for processing an output from the fixed beamformer and identifying speech from noise though the use of an algorithm utilizing a covariance matrix.

Claims

1. A non-spatial speech detection system comprising:

a plurality of microphones;

a fixed beamformer for receiving the output of the plurality of microphones;

an adaptive beamformer for receiving the output of the plurality of microphones; and

at least one processor for processing an output from the fixed beamformer and identifying speech from noise using algorithm utilizing a covariance matrix.

2. A non-spatial speech detection system as in claim 1, wherein an output of the adaptive beamformer is provided as feedback for adjusting the weight of at least one frequency band that is an output from the adaptive beamformer.

3. A non-spatial speech detection system as in claim 2, wherein the weight is adjusted using a least mean squares (LMS) algorithm.

4. A non-spatial speech detection system as in claim 1, wherein the plurality of microphones receive speech emanating from a driver or front seat passenger in a vehicle.

5. A non-spatial speech detection system as in claim 1, wherein the plurality of microphones are located in a rear view housing.

6. A non-spatial speech detection system as in claim 5, wherein the rear view housing includes a mirror element.

7. A non-spatial detection system as in claim 1, wherein the covariance matrix is a Gram matrix.

22. A non-spatial detection system as in claim 1, wherein the fixed beamformer corresponds to a guide beam aimed in a fixed direction.

23. A non-spatial detection system as in claim 1, wherein the output from the fixed beamformer is utilized in the covariance matrix to generate a substantially noise free signal for identifying the speech.

8. A non-spatial speech detection system comprising:

a plurality of microphones; and

at least one processor for processing at least one real audio signal received by the plurality of microphones so that the real audio signal is processed by formatting each frequency band of a plurality of complex frequency bands into a covariance matrix having a plurality of covariance matrix elements and calculating a determinant for each of the covariance matrix elements to provide a speech presence metric to utilize the speech presence metric for identifying speech within the at least one real audio signal, wherein the covariance matrix is generated from an output of a fixed beamformer.

9. A non-spatial speech detection system as in claim 8, wherein the covariance matrix is a Gram matrix.

10. A non-spatial speech detection system as in claim 8, wherein the speech is identified in the presence of noise.

11. A non-spatial speech detection system as in claim 8, wherein the plurality of microphones are located in a vehicle.

12. A non-spatial speech detection system as in claim 8, wherein the at least one real audio signal is speech emanating from the driver or front seat passenger in a vehicle.

13. A non-spatial speech detection system as in claim 8, wherein the plurality of microphones are located in a rear view housing.

14. A non-spatial speech detection system as in claim 13, wherein the rear view housing includes a mirror element.

15. A non-spatial speech detection system comprising:

a plurality of microphones located in a vehicle; and

at least one processor for processing at least one real audio signal received by the plurality of microphones wherein the at least one real audio signal is processed using the steps of;

transforming each frame of the real audio signal into a plurality of complex bands using a fast Fourier transform;

formatting each band of the plurality of complex bands into a covariance matrix having a plurality of matrix elements;

calculating a determinant from the plurality of covariance matrix elements to provide a speech presence metric;

transforming the speech presence metric into a gain function;

calculating a fast average value and slow average value from the gain function; and

utilizing an input signal and the gain function for identifying speech within the at least one real audio signal.

16. A non-spatial speech detection system as in claim 15, wherein the covariance matrix is a Gram matrix.

17. A non-spatial speech detection system as in claim 15, wherein the at least one real audio signal is speech emanating from the driver or front seat passenger in the vehicle.

18. A non-spatial speech detection system as in claim 15, wherein the at least one processor uses no spatial information from the at least one real time audio signal for detecting speech.

19. A non-spatial speech detection system as in claim 15, further comprising at least one beamformer for receiving the at least one rear time audio signal from the plurality of microphones.

20. A non-spatial speech detection as in claim 15, wherein the plurality of microphones are located in a rear view housing.

21. A non-spatial speech detection system as in claim 15, wherein the rear view housing includes a mirror element.