MOTOR-VEHICLE VOICE-CONTROL SYSTEM AND MICROPHONE-SELECTING METHOD THEREFOR

• US 20120221341A1
• Filed: 02/25/2012
• Published: 08/30/2012
• Est. Priority Date: 02/26/2011
• Status: Active Grant

• Current Assignee: Paragon AG
• Original Assignee: Paragon AG
• Inventors: RODEMER, Klaus
• Granted Patent: US 8,996,383 B2
• Time in Patent Office: Days
• US Class Current: 704/275
• 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 21/0264 characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques

Abstract

A voice-control system for motor vehicles has a plurality of spaced microphones emitting respective microphone signals, and an evaluation unit connected to the microphones. This unit serves for assembling correlation pairs from the signals of two of the microphones, calculating a correlation coefficient for each correlation pair, detecting an energy value for each microphone, detecting a respective delay time of a voice signal between a voice signal source and the each of the microphones, and selecting in dependence on current correlation coefficients of the correlation pairs, on the current energy values of the microphones, and on the current delay time of the voice signal to the microphones, that microphone whose signal is optimal as a basis for the operation of the voice-control system.

Claims

1. A voice-control system for motor vehicles, the system comprising:

a plurality of spaced microphones emitting respective microphone signals; and

an evaluation unit connected to the microphones for;

assembling correlation pairs from the signals of two of the microphones,

calculating a correlation coefficient for each correlation pair,

detecting an energy value for each microphone,

detecting a respective delay time of a voice signal between a voice signal source and the each of the microphones, and

selecting in dependence on current correlation coefficients of the correlation pairs, on the current energy values of the microphones, and on the current delay time of the voice signal to the microphones, that microphone whose signal is optimal as a basis for the operation of the voice-control system.

2. The voice-control system defined in claim 1, wherein the microphones are a first microphone, a second microphone, and a third microphone, the evaluation unit serving to assemble

a first correlation pair from the signal of the first microphone and the signal of the second microphone, and

a second correlation pair from the signal of the second microphone and the signal of the third microphone;

calculate a correlation coefficient for both correlation pairs,

detect an energy value for each of the three microphones and the delay time of the voice signal between the voice signal source and each of the microphones,

calculate a delay time difference between the delay time of the voice signal from the voice signal source to the second microphone and to the first microphone can be calculated for the first correlation pair and a delay time difference between the delay time of the voice signal from the voice signal source to the third microphone and to the second microphone for the second correlation pair, and,

selecting, depending on the correlation coefficients of the two correlation pairs, on the current energy values of the microphones, and on the current delay time differences (lag₁₂, lag₂₃), that microphone whose signal is currently optimal as a basis for the operation of the voice-control system.

3. The voice-control system defined in claim 2, wherein the evaluation unit further serves to:

determine the correlation coefficient of each correlation pair by processing the two signals of the microphones of each correlation pair with a Hann window function and with a fast Fourier transform for the two processed signals to determine a cross spectrum from the two transformed signals,

then subjecting the determined cross spectrum to a frequency weighting by calculating energy values from the two transformed signals and calculating mean values for the energy values the two signals and for the frequency-weighted cross spectrum,

then subjecting the frequency-weighted and averaged cross spectrum to an inverse fast Fourier transform and processing it with a covariance function such that a maximum can be determined from the covariance function, and

dividing the maximum determined from the covariance function by the root of the product from the averaged energy values of the two signals.

4. The voice-control system defined in claim 2, wherein the evaluation unit also determines whether:

he correlation coefficient of the first correlation pair is above a first threshold value and above a second higher s threshold value,

the correlation coefficient of the second correlation pair is above a third threshold value,

the delay time difference of the microphones of the first correlation pair is between 121 μ

s and 400 μ

s or between −

185 μ

s and 121 μ

s,

the delay time difference of the microphones of the second correlation pair is between 122 μ

s and 400 μ

s,

the energy value of the first microphone is lower than the energy value of the second microphone,

the energy value of the first microphone is higher than a predetermined energy threshold value,

the energy value of the second microphone is lower than the energy value of the third microphone,

the energy value of the second microphone is higher than the predetermined energy threshold value,

the energy value of the second microphone is lower than the energy value of the first microphone,

the energy value of the third microphone is higher than the predetermined energy threshold value. and

the energy value of the third microphone is lower than the energy value of the second microphone.

5. A method for selecting a microphone from several spaced-apart microphones provided in a motor vehicle of a voice-control system for operating same, the method comprising the steps of:

assembling correlation pairs from the signals of two microphones,

calculating a correlation coefficient for each correlation pair,

detecting an energy value for each microphone,

detecting a respective delay time of a voice signal between a voice signal source and each of the microphones, and

selecting, depending on the current correlation coefficients of the correlation pairs, on the current energy values of the microphones, and on the current delay times of the voice signal to the microphones that microphone whose signal is currently optimal as a basis for the operation of the voice-control system.

6. The method defined in claim 5, wherein the microphones are a first microphone, a second microphone, and a third microphone, the method further comprising the steps of:

assembling a first correlation pair from the signal of the first microphone and the signal of the second microphone and a second correlation pair from the signal of the second microphone and the signal of the third microphone,

calculating respective first and second correlation coefficient for the first correlation pair and for the second correlation pair,

detecting for each microphone an energy value,

detecting a delay time of the voice signal between the voice signal source and each of the microphones,

calculating a delay time difference between the delay time of the voice signal from the voice signal source to the second microphone and to the first microphone for the first correlation pair and a delay time difference between the delay time of the voice signal from the voice signal source to the third microphone and to the second microphone for the second correlation pair, and

selecting, depending on the correlation coefficients of the two correlation pairs, on the current energy values of the microphones, and on the current delay time differences that microphone whose signal is currently optimal as a basis for the operation of the voice-control system.

7. The method defined in claim 6, wherein the correlation coefficient of each correlation pair is determined by:

by processing the two signals of the microphones of each correlation pair with a Hann window function,

carrying out a fast Fourier transform for the two processed signals to determine a cross spectrum from the two transformed signals to a frequency weighting

calculating energy values from the two transformed signals and mean values for the energy values of the two signals and for the frequency-weighted cross spectrum,

subjecting the frequency-weighted and averaged cross spectrum to an inverse fast Fourier transform and processing it to a covariance function,

determining a maximum from the covariance function, and

dividing the maximum determined from the covariance function by the root of the product from the averaged energy values of the two signals.

8. The method defined in claim 6, further comprising the steps of determining whether:

the correlation coefficient of the first correlation pair is above a first threshold value and above a second higher threshold value,

the correlation coefficient of the second correlation pair is above a third threshold value,

a delay time difference of the microphones of the first correlation pair is between 121 μ

s and 400 μ

s or between −

185 μ

s and 121 μ

s,

a delay time difference of the microphones of the second correlation pair is between 122 μ

s and 400 μ

s,

the energy value of the first microphone is lower than the energy value of the second microphone,

the energy value of the first microphone is higher than a predetermined energy threshold value,

the energy value of the second microphone is lower than the energy value of the third microphone,

the energy value of the second microphone is higher than the predetermined energy threshold value,

the energy value of the second microphone is lower than the energy value of the first microphone,

the energy value of the third microphone is higher than the predetermined energy threshold value, and

the energy value of the third microphone is lower than the energy value of the second microphone.

9. The method defined in claim 8, further comprising the step of

selecting the third microphone for operation of the voice-control system if

the correlation coefficient of the second correlation pair is above the third threshold value,

the delay time difference of the microphones of the second correlation pair is between 122 μ

s and 400 μ

s,

the energy value of the second microphone is lower than the energy value of the third microphone, and

the energy value of the third microphone is higher than the predetermined energy threshold value.

10. The method defined in claim 8, further comprising the steps of:

selecting the second microphone for operation of the voice-control system if

the correlation coefficient of the second correlation pair is not above the third threshold value,

the delay time difference of the microphones of the second correlation pair is not between 122 μ

s and 400 μ

s, and

the energy value of the second microphone is not lower than the energy value of the third microphone or the energy value of the third microphone is not higher than the predetermined energy threshold value, and if

the correlation coefficient of the first correlation pair is above the second threshold value,

the delay time difference of the microphones of the first correlation pair is between 121 μ

s and 400 μ

s,

the energy value of the first microphone is lower than the energy value of the second microphone,

the energy value of the second microphone is higher than the predetermined energy threshold value, and

the energy value of the third microphone is lower than the energy value of the second microphone.

11. The method according to claim 8, further comprising the step of selecting the first microphone for operation of the voice-control system if

the correlation coefficient of the second correlation pair is not above the third threshold value,

the delay time difference of the microphones of the second correlation pair is not between 122 μ

s and 400 μ

s, and

the energy value of the second microphone is not lower than the energy value of the third microphone or the energy value of the third microphone is not higher than the predetermined energy threshold value, if

the correlation coefficient of the first correlation pair is not above the second threshold value,

the delay time coefficient of the microphones of the first correlation pair is not between 121 μ

s and 400 μ

s,

the energy value of the first microphone is not lower than the energy value of the second microphone, and

the energy value of the second microphone is not higher than the predetermined energy threshold value or the energy value of the third microphone is not lower than the energy value of the second microphone, and if

the correlation coefficient of the first correlation pair is above the first threshold value,

the delay time difference of the microphones of the first correlation pair is between −

185 μ

s and 121 μ

s,

the energy value of the first microphone is higher than the predetermined energy threshold value, and

the energy value of the second microphone is lower than the energy value of the first microphone.

12. The method according to claim 8, further comprising the step of:

retaining the microphone selected for operation of the voice-control system if

the correlation coefficient of the second correlation pair is not above the third threshold value,

the delay time difference of the microphones of the second correlation pair is not between 122 μ

s and 400 μ

s, and

the energy value of the second microphone is not lower than the energy value of the third microphone or the energy value of the third microphone is not higher than the predetermined energy threshold value, if

the correlation coefficient of the first correlation pair is not above the second threshold value,

the delay time difference of the microphones of the first correlation pair is not between 121 μ

s and 400 μ

s,

the energy value of the first microphone is not lower than the energy value of the second microphone, and

the energy value of the second microphone is not higher than the predetermined energy threshold value or the energy value of the third microphone is not lower than the energy value of the second microphone, and if

the correlation coefficient of the first correlation pair is not above the first threshold value,

the delay time difference of the microphones of the first correlation pair is not between −

185 μ

s and 121 μ

s, and

the energy value of the first microphone is not higher than the predetermined energy threshold value or the energy value of the second microphone is not lower than the energy value of the first microphone.