METHODS AND APPARATUS FOR DETECTING A VOICE COMMAND

• US 20140278435A1
• Filed: 03/12/2013
• Published: 09/18/2014
• Est. Priority Date: 03/12/2013
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Ganong, William F. III, Sejnoha, Vladimir, Van Mulbregt, Paul Adrian, Wilson, Glenn
• Granted Patent: US 9,361,885 B2
• Time in Patent Office: Days
• US Class Current: 704/275
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• 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 2015/223 Execution procedure of a spoken command

Abstract

Some aspects include a method of monitoring an acoustic environment of a mobile device operating in a low power mode, the mobile device having a first and second processor, the method comprises receiving acoustic input while the mobile device is operating in the low power mode, performing at least one first processing stage on the acoustic input using the first processor, prior to engaging the second processor, to evaluate whether the acoustic input includes a voice command, performing at least one second processing stage on the acoustic input using the second processor to evaluate whether the acoustic input includes a voice command if further processing is needed to determine whether the acoustic input includes a voice command, and initiating responding to the voice command when either the at least one first processing stage or the at least one second processing stage determines that the acoustic input includes a voice command.

Claims

1. A method of monitoring an acoustic environment of a mobile device for voice commands when the mobile device is operating in a low power mode, the mobile device having a first processor and a second processor, the method comprising:

receiving acoustic input while the mobile device is operating in the low power mode;

performing at least one first processing stage on the acoustic input using the first processor, prior to engaging the second processor to process the acoustic input, to evaluate whether the acoustic input includes a voice command;

performing at least one second processing stage on the acoustic input using the second processor to evaluate whether the acoustic input includes a voice command if further processing is needed to determine whether the acoustic input includes a voice command; and

initiating responding to the voice command when either the at least one first processing stage or the at least one second processing stage determines that the acoustic input includes a voice command.

2. The method of claim 1, wherein the first processor is a secondary processor and the second processor is a primary processor.

3. The method of claim 2, wherein the secondary processor is a digital signal processor capable of operating at a lower power than the primary processor.

4. The method of claim 2, wherein performing the at least one first processing stage includes performing at least one voice activity detection processing stage including performing at least one of spectral analysis on the acoustic input to evaluate whether the spectrum of the acoustic input is indicative of voice activity, periodicity analysis to evaluate whether the signal periodicity is indicative of voice activity, and phone loop analysis to evaluate whether the acoustic input includes speech.

5. The method of claim 2, wherein performing the at least one first processing stage comprises performing limited vocabulary automatic speech recognition on the acoustic input to detect whether the acoustic input includes an explicit voice trigger.

6. The method of claim 2, wherein performing the at least one second processing stage comprises performing at least one speech processing stage including at least one of automatic speech recognition on the acoustic input, at least one classification processing stage, and at least one natural language processing stage to understand the meaning of the voice command.

7. The method of claim 4, wherein performing the at least one second processing stage comprises performing limited vocabulary automatic speech recognition on the acoustic input to detect whether the acoustic input includes an explicit voice trigger.

8. The method of claim 2, further comprising providing at least one low power indication that the acoustic input was received and is being processed prior to activating the primary processor or turning on a display of the mobile device.

9. The method of claim 8, wherein the at least one low power indication includes illuminating one or more light emitting diodes of the mobile device.

10. The method of claim 8, wherein the at least one low power indication includes at least one audio signal indicating that the acoustic input was received and is being processed.

11. The method of claim 1, wherein the mobile device is capable of connecting to at least one server via a network, the method further comprising transmitting at least a portion of the acoustic information to the at least one server via the network for processing by the at least one server.

12. The method of claim 1, further comprising transitioning the mobile device to an active mode to respond to the voice command when either the at least one first processing stage or the at least one second processing stage determines that the acoustic input includes a voice command.

13. The method of claim 1, wherein at least one task specified by the voice command is performed with the mobile device remaining in the low power mode.

14. The method of claim 1, wherein responding to the voice command is performed with the mobile device remaining in the low power mode.

15. The method of claim 1, further comprising performing one or more default actions when a specific task could not be ascertained from the acoustic input.

16. The method of claim 1, wherein performing the at least one first processing stage comprises evaluating at least one contextual cue selected from a group consisting of:

time of day, motion of the mobile device, location of the mobile device, an occurrence of an event on the mobile device, recent activity on the mobile device and state of a battery powering the mobile device.

17. The method of claim 1, wherein performing the at least one second processing stage comprises evaluating at least one contextual cue selected from a group consisting of:

time of day, motion of the mobile device, location of the mobile device, an occurrence of an event on the mobile device, recent activity on the mobile device and state of a battery powering the mobile device.

18. At least one computer readable medium storing with instructions that, when executed on at least one processor, perform a method of monitoring an acoustic environment of a mobile device to receive acoustic input from the environment of the mobile device while the mobile device is operating in a low power mode, the mobile device having a first processor and second processor, the method comprising:

performing at least one first processing stage on the acoustic input using the first processor, prior to engaging the second processor to process the acoustic input, to evaluate whether the acoustic input includes a voice command;

performing at least one second processing stage on the acoustic input using the second processor to evaluate whether the acoustic input includes a voice command if further processing is needed to determine whether the acoustic input includes a voice command; and

initiating responding to the voice command when either the at least one first processing stage or the at least one second processing stage determines that the acoustic input includes a voice command.

19. A mobile device comprising:

at least one input to receive acoustic input from an environment of the mobile device while the mobile device is operating in the low power mode;

a first processor to perform at least one first processing stage on the acoustic input to evaluate whether the acoustic input includes a voice command;

a second processor to perform at least one second processing stage on the acoustic input to evaluate whether the acoustic input includes a voice command, wherein the at least one first processing stage is performed prior to engaging the second processor, wherein the at least one second processing stage is performed if further processing is needed to determine whether the acoustic input includes a voice command, and wherein response to the voice command is initiated when either the at least one first processing stage or the at least one second processing stage determines that the acoustic input includes a voice command.