Methods and systems for determining and using a confidence level in speech systems

• US 10,276,159 B2
• Filed: 05/10/2016
• Issued: 04/30/2019
• Est. Priority Date: 05/10/2016
• Status: Expired due to Fees

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: Wang, Guoqing, Jiang, Chuan, Tian, Kai, Yang, Wei
• Primary Examiner(s): Godbold, Douglas
• Application Number:
  US15/150,975
• Publication Number:
  US 20170330560A1
• Time in Patent Office: 1,085 Days
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context

Abstract

Methods and systems are provided for processing speech inputs for a controlling one or more vehicle systems of a vehicle. In one embodiment, a method includes: receiving speech input from an audio channel; performing, by a processor, speech recognition on the speech input to obtain recognized results; determining, by a processor, an accuracy level of the audio channel based on a comparison of the recognized results and predictive phraseology; determining, by a processor, an integrity level of the audio channel based on situational awareness information; communicating the recognized results, accuracy level, and the integrity level to a vehicle system; and selectively using the recognized results by the vehicle system based on the accuracy level and the integrity level.

Claims

1. A method of validating speech inputs for controlling one or more vehicle systems of a vehicle, comprising:

validating the speech inputs periodically and upon events by using predictive phraseology stored in a data storage device of the vehicle, and by using situational information received by the one or more vehicle systems to determine eligible voice features of the vehicle by;

receiving speech input from an audio channel;

performing, by a processor, speech recognition on the speech input to obtain recognized results wherein the speech input has not yet been validated;

determining by the processor whether one or more conditions exist related to the speech input in order to validate the speech input wherein if the one or more conditions exist then validating an accuracy and an integrity of an audio channel of the speech input by;

determining for validating the accuracy of the audio channel, by a processor, an accuracy level of the audio channel based on whether the one or more conditions exist for updating a confidence level of a speech output and by computing a comparison of the recognized results and the predictive phraseology in order to assign a descriptive level to the audio channel based on the comparison, and

determining for validating the integrity of the audio channel, by a processor, an integrity level of the audio channel based on the context of the speech input as indicated by current configuration settings of the vehicle and on situational awareness information provided by the vehicle or remote systems by;

computing a comparison of the content of the speech input to an expected content, and assigning the integrity level based on the computed comparison for associating a tag indicative of the integrity level of the audio channel with messages of the audio channel;

communicating descriptive labels with the recognized results, accuracy level, and the integrity level to the vehicle system wherein the descriptive labels are associated with the tag indicative of the accuracy level and integrity level of the audio channel; and

selectively using the recognized results by the vehicle system based on the accuracy level and the integrity level to determine the eligible voice features of the vehicle.

2. The method of claim 1, wherein the determining the integrity level is performed based on the performance of the determining the accuracy level.

3. The method of claim 2, further comprising maintaining the accuracy level while updating the integrity level.

4. The method of claim 3, further comprising determining the descriptive label based on a cause for the update in the integrity level.

5. The method of claim 3, further comprising determining the descriptive label based on the situational awareness information.

6. A system for validating speech inputs for controlling one or more vehicle systems of a vehicle, comprising:

a non-transitory computer readable medium comprising;

a first module that performs, by a processor, speech recognition on a speech input received from an audio channel to obtain recognized results;

a second module that validates the speech input by a determination, by a processor, of an accuracy level of the audio channel based on whether one or more conditions exist for updating a confidence level of a speech output and by a computation of a comparison of the recognized results and predictive phraseology for an assignment of a numeric value based on the comparison; and

a third module that further validates the speech input by a determination, by a processor, of an integrity level of the audio channel based on the context of the speech input as indicated by current configuration settings of the vehicle and on situational awareness information provided by the vehicle or remote systems by;

a computation of a comparison of the content of the speech input to an expected content, and an assignment of the integrity level based on the computed comparison in order to communicate the recognized results, the accuracy level, and the integrity level to a vehicle system to selectively determine eligible voice features of the vehicle.

7. The system of claim 6, further comprising a vehicle system that receives the recognized results, accuracy level, and the integrity level and that selectively uses the recognized results based on the accuracy level and the integrity level.

8. The system of claim 6, wherein the third module performs the determining of the integrity level based on the performance of the determining the accuracy level.

9. The system of claim 8, wherein the second module further maintains the accuracy level while the third module further updates the integrity level.

10. The system of claim 6, wherein the third module associates a descriptive label with the integrity level, and communicates the descriptive label to the vehicle system.

11. The system of claim 10, wherein the third module determines the descriptive label based on a cause for the update in the integrity level.

12. The system of claim 10, wherein the third module determines the descriptive label based on the situational awareness information.

13. An aircraft, comprising:

a speech input device that receives speech from a user;

a plurality of speech dependent aircraft systems that validate the speech from the user periodically and upon events by using predictive phraseology, and by using situational information received by the speech dependent aircraft systems to determine eligible voice enabled features of the aircraft; and

a non-transitory computer readable medium, comprising;

a first module that performs, by a processor, speech recognition on a speech input received from an audio channel to obtain recognized results;

a second module that validates the speech input by a determination, by a computation by a processor, of an accuracy level of the audio channel based on whether any of the one or more conditions exist for updating a confidence level of a speech output and by a comparison of the recognized results and the predictive phraseology; and

a third module that further validates the speech input by a determination, by a processor, of an integrity level of the audio channel based on situational awareness information provided by the vehicle or remote systems by;

a computation by the processor of a comparison of content of the speech input to an expected content, and assigning the integrity level based on the computed comparison in order to communicate the recognized results, the accuracy level, and the integrity level to a vehicle system to selectively determine eligible speech dependent aircraft systems.

14. The aircraft of claim 13, wherein the at least one of the plurality of speech dependent aircraft systems selectively uses the recognized results based on the accuracy level and the integrity level.

15. The aircraft of claim 13, wherein the third module performs the determining of the integrity level based on the performance of the determining the accuracy level.

16. The aircraft of claim 15, wherein the second module further maintains the accuracy level while the third module further updates the integrity level.

17. The aircraft of claim 13, wherein the third module associates a descriptive label with the integrity level, and communicates the descriptive label to the at least one of the plurality of speech dependent aircraft systems.

18. The aircraft of claim 17, wherein the third module determines the descriptive label based on a cause for the update in the integrity level.

19. The aircraft of claim 17, wherein the third module determines the descriptive label based on the aircraft situational awareness information.