Power-Efficient Voice Activation

• US 20130339028A1
• Filed: 06/15/2012
• Published: 12/19/2013
• Est. Priority Date: 06/15/2012
• Status: Active Grant

• Current Assignee: Monterey Research, LLC (Vector Capital Corporation)
• Original Assignee: Spansion LLC (Infineon Technologies AG)
• Inventors: Liu, Chen, Olson, Jens, ROSNER, Stephan
• Granted Patent: US 9,142,215 B2
• Time in Patent Office: Days
• US Class Current: 704/275
• 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 15/222 Barge in, i.e. overridable guidance for interrupting prompts
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: H04R 29/004 for microphones H04R29/007 takes precedence

Abstract

A voice activation system is provided. The voice activation system includes a first stage configured to output a first activation signal if at least one energy characteristic of a received audio signal satisfies at least one threshold and a second stage configured to transition from a first state to a second state in response to the first activation signal and, when in the second state, to output a second activation signal if at least a portion of a profile of the audio signal substantially matches at least one predetermined profile.

Claims

1. A voice activation system, comprising:

a first stage configured to output a first activation signal if at least one energy characteristic of a received audio signal satisfies at least one threshold; and

a second stage configured to transition from a first state to a second state in response to the first activation signal and, when in the second state, to output a second activation signal if at least a portion of a profile of the audio signal substantially matches at least one predetermined profile.

2. The voice activation system of claim 1, further comprising:

a speech recognition engine coupled to the second stage and configured to transition from a first state to a second state based on the second activation signal.

3. The voice activation system of claim 2, wherein the second state of the speech recognition engine is a fully operational state.

4. The voice activation system of claim 2, wherein the second state of the speech recognition is a wake-up word detection state and wherein the speech recognition is configured to transition to a fully operational state if the speech recognition engine recognizes at least one wake-up word in the audio signal.

5. The voice activation system of claim 4, wherein the speech recognition engine is configured to transition to the first state if the speech engine does not recognize at least one wake-up word in the audio signal.

6. The voice activation system of claim 4, further comprising:

a control module configured to enable the speech recognition engine to transition to the wake-up word detection state.

7. The voice activation system of claim 6, wherein the control module is configured to enable the speech recognition engine to transition to the wake-up word detection state based on at least input from a user.

8. The voice activation system of claim 1, wherein the first stage is configured to compare an energy level of the audio signal to a threshold of the at least one thresholds.

9. The voice activation system of claim 1, wherein the first stage is configured to compare a ratio of high frequency energy to low frequency energy in the audio signal to a threshold of the one or more thresholds.

10. The voice activation system of claim 1, wherein the second stage is configured to compare at least one of a time domain profile or a frequency domain profile to the at least one predetermined profile.

11. The voice activation system of claim 1, wherein the second stage is configured to extract a feature of the audio signal and to compare the feature to the at least one predetermined profile.

12. A voice activation method, comprising:

comparing at least one energy characteristic of an audio signal to at least one threshold using a first stage of a voice activation system;

transitioning a second stage of the voice activation system from a first state to a second stage if the audio signal satisfies the threshold;

comparing at least a portion of a profile of the audio signal to at least one predetermined profile using the second stage of the voice activation system while the second stage of the voice activation system is in the second state; and

transitioning a speech recognition engine of the voice activation system from a first state to a second state if the least a portion of a profile of the audio signal substantially matches the at least one predetermined profile.

13. The method of claim 12, wherein the second state of the speech recognition engine is a fully operational state.

14. The method of claim 12, wherein the second state of the speech recognition is a wake-up word detection state, the method further comprising:

determining whether at least one wake-up word is present in the audio signal using the speech recognition engine while the speech recognition engine is in the wake-up word detection state; and

transitioning the speech recognition engine from the wake-up word detection state to a fully operational state if at least one wake-up word is present in the audio signal.

15. The method of claim 14, further comprising:

enabling the speech recognition state to transition to the wake-up word detection state.

16. The method of claim 12, wherein comparing at least one energy characteristic of an audio signal comprises:

comparing an energy level of the audio signal to a threshold of the at least one thresholds.

17. The method of claim 12, wherein comparing at least one energy characteristic of an audio signal comprises:

comparing a ratio of high frequency energy to low frequency energy in the audio signal to a threshold of the one or more thresholds.

18. The method of claim 12, wherein comparing at least a portion of a profile of the audio signal comprises:

comparing at least one of a time domain profile or a frequency domain profile to the at least one predetermined profile.

19. The method of claim 12, wherein comparing at least a portion of a profile of the audio signal comprises:

extracting a feature of the audio signal; and

comparing the feature to the at least one predetermined profile.

20. A voice activation system, comprising:

a microphone configured to output an analog electrical signal corresponding to received sound waves;

an analog-to-digital converter configured to covert the analog electrical signal to a digital signal;

a first stage configured to output a first activation signal if at least one energy characteristic of the digital signal satisfies at least one threshold;

a second stage configured to transition from a stand-by state to a fully-operational state in response to the first activation signal and, when in the fully-operational state, to output a second activation signal if at least a portion of a profile of the audio signal substantially matches at least one predetermined profile; and

a speech recognition engine configured to transition from a first state to a second state based on the second activation signal.