Method and system for using sound related vehicle information to enhance speech recognition

• US 9,263,040 B2
• Filed: 01/17/2012
• Issued: 02/16/2016
• Est. Priority Date: 01/17/2012
• Status: Active Grant

• Current Assignee: GM Global Technology Operations LLC (General Motors Company)
• Original Assignee: GM Global Technology Operations LLC (General Motors Company)
• Inventors: Tzirkel-Hancock, Eli, Tsimhoni, Omer
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US13/351,314
• Publication Number:
  US 20130185065A1
• Time in Patent Office: 1,491 Days
• Field of Search: 704/233, 704/275
• US Class Current: 1/1
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems
• 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 2015/226 using non-speech characteristics
• CPC Class Codes: G10L 21/0208 Noise filtering

Abstract

An audio signal may be received, in a processor associated with a vehicle. Sound related vehicle information representing one or more sounds may be received by the processor. The sound related vehicle information may or may not include an audio signal. A speech recognition process or system may be modified based on the sound related vehicle information.

Claims

1. A method comprising:

receiving, in a processor associated with a vehicle, an audio signal;

receiving, in the processor, sound related vehicle information representing one or more sounds, the sound related vehicle information not comprising an audio signal;

selecting a previously generated acoustic model from among a plurality of acoustic models, to decode speech; and

modifying a speech recognition process of the previously generated acoustic model based on the sound related vehicle information, wherein the modifying includes;

determining a pre-trained filter according to filter parameters based on the sound related vehicle information; and

applying the pre-trained filter to the audio signal.

2. The method of claim 1, comprising determining interference profile records based on the sound related vehicle information.

3. The method of claim 2, wherein modifying the speech recognition process based on the sound related vehicle information comprises:

selecting an acoustic model based on the interference profile records; and

decoding speech using the acoustic model.

4. The method of claim 1, wherein the filter is in an automatic speech recognition system front-end and

wherein the applying the filter to the audio signal is in the automatic speech recognition system front-end.

5. The method of claim 2, wherein determining the interference profile records based on the sound related vehicle information comprises determining the interference profile records using logic operations based on multiple types of sound related vehicle information.

6. The method of claim 1, comprising:

measuring a success of speech recognition modification based on the sound related vehicle information in enhancing speech recognition functionality; and

adapting the speech recognition modification based on the measured success.

7. The method of claim 1, comprising:

generating a response to speech commands in the audio signal; and

outputting the response to a vehicle occupant.

8. A system comprising:

a memory;

a processor associated with a vehicle to;

receive an audio signal;

receive sound related vehicle information representing one or more sounds, the sound related vehicle information not including an audio signal;

select a previously generated acoustic model from among a plurality of acoustic models, to decode speech; and

modify a speech recognition process of the previously generated acoustic model based on the sound related vehicle information, wherein to modify a speech recognition process includes to;

determine a pre-trained filter according to filter parameters based on the sound related vehicle information; and

apply the pre-trained filter to the audio signal.

9. The system of claim 8, wherein the processor is to determine interference profile records based on the sound related vehicle information.

10. The system of claim 9, wherein to modify the speech recognition process based on the sound related vehicle information the processor is to:

select an acoustic model based on the interference profile records; and

decode speech using the acoustic model.

11. The system of claim 8, wherein the filter is in a automatic speech recognition system front-end and

wherein to apply the filter to the audio signal is in the automatic speech recognition system front-end.

12. The system of claim 9, wherein to determine the interference profile records based on the sound related vehicle information the processor is to determine the interference profile records by quantizing sound related vehicle information.

13. The system of claim 9, wherein to determine the interference profile records based on the sound related vehicle information the processor is to determine the interference profile records using a table.

14. The system of claim 8, wherein the processor is to:

measure a success of speech recognition modification based on the sound related vehicle information in enhancing speech recognition functionality; and

adapt the speech recognition modification based on the measured success.

15. A method comprising:

measuring, using one or more microphones associated with a vehicle, sounds;

selecting a previously generated acoustic model from among a plurality of acoustic models, to decode speech;

transmitting a signal representing the sounds to an automatic speech recognition system according to the previously generated acoustic model;

receiving, at a controller, information related to the operation of vehicle systems causing sound;

calculating interference profile records based on the information, the interference profile records representing noise types and noise levels; and

altering speech recognition based on the interference profile records, wherein the altering speech recognition includes;

choosing a pre-trained filter according to filter settings based on the information related to the operation of vehicle systems causing sound; and

applying the pre-trained filter to the signal.

16. The method of claim 15, wherein altering speech recognition based on the interference profile records comprises:

determining an acoustic model based on the interference profile records; and

decoding speech in the signal using the acoustic model.

17. The method of claim 15, wherein the filter is in an automatic speech recognition system front-end and wherein applying the filter to the signal is in the automatic speech recognition system front-end.