Adaptive confidence thresholds in telematics system speech recognition

• US 20060074651A1
• Filed: 09/22/2004
• Published: 04/06/2006
• Est. Priority Date: 09/22/2004
• Status: Active Grant

• Current Assignee: General Motors LLC (General Motors Company)
• Original Assignee: General Motors Corporation (General Motors Company)
• Inventors: Arun, Uma
• Granted Patent: US 8,005,668 B2
• Time in Patent Office: Days
• US Class Current: 704/233
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering

Abstract

A method of configuring a speech recognition unit in a vehicle. The method includes receiving a noise error from the speech recognition unit responsive to a user voice command and reducing a confidence threshold for an appropriate grammar set responsive to the received noise error.

Claims

1. A method of configuring a speech recognition unit in a vehicle, the method comprising:

receiving a noise error from the speech recognition unit responsive to a user voice command; and

reducing a confidence threshold for an appropriate grammar set responsive to the received noise error.

2. The method of claim 1, further comprising:

announcing an advisory message to a user responsive to the reduction of the confidence threshold.

3. The method of claim 1, wherein reducing a confidence threshold comprises:

retrieving a current confidence threshold responsive to the received noise error;

determining whether the current confidence equals a minimum confidence threshold; and

reducing the current confidence threshold by a preset amount based on the determination.

4. The method of claim 3, wherein the confidence threshold is reduced by the preset amount two or more times responsive to at least two or more respective received noise errors.

5. The method of claim 4, wherein the confidence threshold is reduced by the preset amount until the current confidence threshold equals the minimum confidence threshold.

6. The method of claim 1, further comprising:

resetting the confidence thresholds to default when ignition is cycled.

7. The method of claim 1, further comprising:

announcing a noise reduction advisory message to a user responsive to the received noise error; and

receiving a second noise error from the speech recognition unit responsive to a repeated user voice command,

wherein the repeated user voice command is announced responsive to the noise reduction advisory message, and

wherein the reducing a confidence threshold for an appropriate grammar set is responsive to the received second noise error.

8. The method of claim 7, wherein reducing a confidence threshold is responsive to announcing more than one noise reduction advisory messages during a speech recognition session.

9. The method of claim 7, further comprising:

storing the repeated user voice command in the telematics unit; and

attempting to recognize the stored repeated user voice command at the speech recognition unit after the confidence threshold is reduced.

10. The method of claim 9, further comprising:

reducing a confidence threshold for the appropriate grammar set responsive to a third received noise error, wherein the third received noise error is responsive to the attempting to recognize the stored user voice command; and

attempting to recognize the stored repeated user voice command at the speech recognition unit after the confidence threshold is reduced twice.

14. The system of claim 7, further comprising:

means for storing a repeated user voice command in the telematics unit; and

means for attempting to recognize the stored repeated user voice command at the speech recognition unit after the confidence threshold is reduced.

11. A system for configuring a speech recognition unit in a vehicle, the system comprising:

means for receiving a noise error from the speech recognition unit responsive to a user voice command; and

means for reducing a confidence threshold for an appropriate grammar set responsive to the received noise error.

12. The system of claim 11, further comprising:

means for announcing an advisory message to a user responsive to the reduction of the confidence threshold.

13. The system of claim 11, wherein the means for reducing a confidence threshold comprise:

means for retrieving a current confidence threshold responsive to the received noise error;

means for determining whether the current confidence equals a minimum confidence threshold; and

means for reducing the current confidence threshold by a preset amount based on the determination.

15. The system of claim 11, further comprising:

means for resetting the confidence thresholds to default when ignition is cycled.

16. A computer readable medium storing a computer program comprising:

computer readable code for receiving a noise error from the speech recognition unit responsive to a user voice command; and

computer readable code for reducing a confidence threshold for an appropriate grammar set responsive to the received noise error.

17. The medium of claim 16, further comprising:

computer readable code for announcing an advisory message to a user responsive to the reduction of the confidence threshold.

18. The medium of claim 16, wherein computer readable code for reducing a confidence threshold comprises:

computer readable code for retrieving a current confidence threshold responsive to the received noise error;

computer readable code for determining whether the current confidence equals a minimum confidence threshold; and

computer readable code for reducing the current confidence threshold by a preset amount based on the determination.

19. The medium of claim 16, further comprising:

computer readable code for resetting the confidence thresholds to default when ignition is cycled.

20. The medium of claim 16, further comprising:

computer readable code for storing a repeated user voice command in the telematics unit; and

computer readable code for attempting to recognize the stored repeated user voice command at the speech recognition unit after the confidence threshold is reduced.