Method for automatically and dynamically switching between speech technologies

• US 6,704,707 B2
• Filed: 03/14/2001
• Issued: 03/09/2004
• Est. Priority Date: 03/14/2001
• Status: Expired due to Term

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Anderson, Andrew V., Bennett, Steven M.
• Primary Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US09/808,699
• Publication Number:
  US 20020133337A1
• Time in Patent Office: 1,091 Days
• Field of Search: 704/231, 704/236, 704/243, 704/244, 704/246, 704/250, 704/251, 704/255, 704/270, 379/88.01, 379/88.04
• US Class Current: 704/231
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A method for switching between speech recognition technologies. The method includes reception of an initial recognition request accompanied by control information. Recognition characteristics are determined using the control information and then a switch is configured based upon the particular characteristic. Alternatively, the switch may be configured based upon system load levels and resource constraints.

Claims

1. A method for switching between speech recognition technologies, the method comprising:

a) receiving a recognition request, wherein the recognition request is accompanied by control information;

b) determining recognition characteristics of the recognition request using the control information;

c) configuring a switch based upon the recognition characteristics; and

d) applying a speech recognition technology determined by the configuration of the switch.

2. The method of claim 1, the method further comprising monitoring at least one of resource constraints and load levels and configuring a switch is also based upon at least one of the resource constraints and load levels.

3. The method of claim 2, the method further comprising reconfiguring the switch as necessary based upon at least one of the resource constraints and load levels after the initial application of a speech technology.

4. The method of claim 3, wherein the method further comprises buffering signals from the recognition request and voice data prior to configuring the switch.

5. The method of claim 1, wherein the method is applied to several recognition requests received by the system at substantially the same time.

6. The method of claim 1, wherein the control information includes a priority rating usable to determine the speech technology to apply.

7. The method of claim 1, wherein the request is from a new audio stream.

8. The method of claim 1, wherein the request is from new data in an acknowledged audio stream.

9. The method of claim 1, wherein configuring a switch is also based upon recognizer characteristics.

10. An article containing software instructions that, when executed, result in:

a) reception of a recognition request, wherein the recognition request is accompanied by control information;

b) determination of recognition characteristics of the recognition request using the control information;

c) configuration of a switch based upon the recognition characteristics; and

d) application of a speech technology determined by the configuration of the switch.

11. The article of claim 10, the article further including instructions that, when executed, result in monitoring at least one of resource constraints and load levels and configuring a switched based upon at least one of the resource constraints and load levels.

12. The article of claim 10, wherein the instructions are executed for several recognition requests received by the system at substantially the same time.

13. The article of claim 10, wherein the control information includes a priority rating usable to determine the speech technology to apply.

14. The article of claim 10, the article further including instructions that, when executed, result in reconfiguration of the switch as necessary based upon at least one of resource constraints and load levels after a speech technology has been applied.

15. The article of claim 10, wherein the article further includes instructions that, when executed, results in buffering of signals from the recognition request and voice data prior to reconfiguring the switch.

16. A speech recognition manager, comprising:

a) a control module operable to receive input audio streams and indicate recognition characteristics of the input audio streams;

b) a director operable to receive control information accompanying the input audio stream;

c) a switch operable to direct the input audio streams in accordance with configuration instructions from the director; and

d) at least two recognition engines operable to receive an input audio stream from the switch when the switch is configured to direct the stream to that recognition engine, and to perform speech conversion.

17. The manager of claim 16 wherein the manager further comprises an output switch operable to select between at least two output streams from the recognition engines.

18. The manager of claim 16, wherein the at least two recognition engines further comprises an array of recognition engines.

19. A method for switching between speech recognition technologies, the method comprising:

a) receiving a recognition request;

b) monitoring at least one of resource constraints and system load levels;

c) configuring a switch based upon at least one of the resource constraints and system load levels; and

d) applying a speech recognition technology determined by the configuration of the switch.

20. The method of claim 19, wherein the request is from a new audio stream.

21. The method of claim 20, wherein a priority code accompanying the request is used in configuring the switch.

22. The method of claim 19, wherein monitoring at least one of the resource constraints and system load levels is an ongoing process.

23. The method of claim 19, wherein the method is applied to several recognition requests received by the system at substantially the same time.

24. The method of claim 19, wherein the method further comprises buffering signals from the recognition request and voice data prior to configuring the switch.

25. The method of claim 19, wherein the request is from new data in an acknowledged audio stream.

26. The method of claim 21, wherein a priority code accompanying the request is used in configuring the switch.

27. The method of claim 19, wherein configuring the switch is also based upon recognizer characteristics.

28. An article containing software instructions that, when executed result in:

a) reception of a recognition request;

b) monitoring at least one of resource constraints and load levels;

c) configuring a switched based upon at least one of the resource constraints and system load levels; and

d) applying a speech recognition technology determined by the configuration of the switch.

29. The article of claim 28, wherein the request is from a new audio stream.

30. The article of claim 28, wherein the request is from new data in an acknowledged audio stream.