Preventing false wake word detections with a voice-controlled device

• US 9,368,105 B1
• Filed: 06/26/2014
• Issued: 06/14/2016
• Est. Priority Date: 06/26/2014
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Freed, Ian W., Barton, William Folwell, Prasad, Rohit
• Primary Examiner(s): Augustin, Marcellus
• Application Number:
  US14/316,612
• Time in Patent Office: 719 Days
• Field of Search: 704/4, 704208-254
• US Class Current: 1/1
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

Natural language controlled devices may be configured to activate command recognition in response to one or more wake words. Techniques are provided to allow for multiple operating modes in which different recognition parameters are employed in recognizing wake words that activate the natural language control functionality of a computing device.

Claims

1. A computing system comprising:

one or more hardware processors;

one or more computer-readable storage media storing computer-executable instructions maintained in the one or more computer-readable storage media and executed by the one or more hardware processors to cause the one or more hardware processors to perform acts comprising;

in a first operating mode;

receiving a first input audio signal corresponding to a first user utterance, wherein the first user utterance comprises a wake word and a first command;

determining that the first user utterance comprises the wake word;

transmitting at least a portion of the first input audio signal to a remote device, wherein the remote device causes a first action to be performed based at least in part on the first command;

receiving an output audio signal from the remote device, wherein the output audio signal corresponds to the first action;

playing the output audio signal;

receiving a second input audio signal corresponding to a second user utterance, wherein the second user utterance comprises a second command to enter a second operating mode;

determining that the second user utterance comprises the second command to enter the second operating mode;

entering the second operating mode, wherein entering the second operating mode includes at least one of maintaining a connection to the remote device or continuing to play the output audio signal;

in the second operating mode;

receiving a third input audio signal corresponding to a third user utterance, wherein the third user utterance comprises the wake word and a third command;

determining that the third command does not comprise a command to change to the first operating mode;

at least partly in response to determining that the third command does not comprise a command to change to the first operating mode, refraining from performing an action corresponding to the third command;

receiving a fourth input audio signal corresponding to a fourth user utterance, wherein the fourth user utterance comprises a fourth command to enter the first operating mode;

determining that the fourth user utterance comprises the fourth command to enter the first operating mode; and

entering the first operating mode, with the first operating mode and the second operating mode.

2. The system as recited in claim 1, wherein:

the determining that the first user utterance comprises the wake word is performed using first recognition settings corresponding to the first operating mode; and

the determining that the fourth user utterance comprises the fourth command to enter the first operating mode is performed using second recognition settings corresponding to the second operating mode.

3. The system as recited in claim 2, wherein the second recognition settings include a recognition certainty that is a higher level of certainty than a recognition certainty of the first recognition settings.

4. The system as recited in claim 2, wherein the computer-executable instructions maintained in the one or more computer-readable storage media are further executed by the one or more hardware processors to cause the one or more hardware processors to perform acts comprising:

in the first operating mode;

receiving a fifth input audio signal corresponding to a fifth user utterance, wherein the fifth user utterance comprises a wake word and a fifth command;

determining that the first user utterance comprises the wake word; and

transmitting at least a portion of the first input audio signal to the remote device, wherein the remote device causes a functionality of at least one of the system or the remote device to be one of locked, restricted, disabled, unlocked, unrestricted, or enabled.

5. A method comprising:

processing, by a voice controlled device comprising a hardware processor and a microphone, a first audio input signal generated by the microphone and corresponding to a first user utterance;

recognizing a wake word based on one or more recognition settings of a first operating mode, wherein the one or more recognition settings of the first operating mode differ from one or more recognition settings of a second operating mode of the voice controlled device;

in response to the recognizing of the wake word, causing further processing of the first audio input signal;

storing an indication that the voice-controlled device is operating in the second operating mode;

receiving a second input audio signal corresponding to a second user utterance;

determining that the second utterance does not comprise a command to change to the first operating mode; and

at least partly in response to determining that the second user utterance does not comprise a command to change to the first operating mode, refraining from performing an action corresponding to the second user utterance.

6. The method as recited in claim 5, wherein a recognition certainty level of the recognition settings of the first operating mode for recognition of audio input in the first operating mode is higher than a recognition certainty level of the recognition settings of the second operating mode.

7. The method as recited in claim 5, further comprising:

in the first operating mode;

transmitting at least a portion of the first audio input signal to a remote device, wherein the remote device causes an action to be performed based at least in part on a first command included in the portion of the first audio input signal;

receiving an output audio signal from the remote device, wherein the output audio signal corresponds to the action; and

playing the output audio signal.

8. The method as recited in claim 5, wherein the voice controlled device provides additional functionality in the first operating mode not provided in the second operating mode.

9. The method as recited in claim 8, wherein the additional functionality in the first operating mode includes initializing a connection to one or more remote devices and providing an audio input signal to the remote devices for recognition of commands in the provided audio input signal by the one or more remote devices.

10. The method as recited in claim 5, further comprising:

in the second operating mode;

receiving a third input audio signal corresponding to a third user utterance, wherein the third user utterance comprises a command to enter the first operating mode;

determining that the third user utterance comprises a command to change to the first operating mode based on the one or more recognition settings of the second operating mode; and

storing an indication that the voice-controlled device is operating in the first operating mode.

11. The method as recited in claim 5, wherein operating in the second operating mode includes maintaining a connection to a remote device.

12. A computing system comprising:

one or more hardware processors;

one or more computer-readable storage media storing computer-executable instructions maintained in the one or more computer-readable storage media and executed by the one or more hardware processors to cause the one or more hardware processors to perform the acts comprising;

receiving a first audio input signal corresponding to a first user utterance and comprising a wake word;

determining that the first user utterance comprises the wake word;

transmitting at least a portion of the first audio input signal to a remote device, wherein the remote device causes the computing system to perform a function of the computing system based at least in part on a first command in the portion of the first audio input signal;

receiving a second input audio signal corresponding to a second user utterance, wherein the second user utterance comprises a second command to enter a second operating mode;

determining that the second user utterance comprises the second command to enter the second operating mode;

storing an indication that the computing system is operating in the second operating mode, wherein operating in the second operating mode includes at least one of maintaining a connection to the remote device or not interrupting the function being performed by the computing system based at least in part on the first command;

receiving a third input audio signal corresponding to a third user utterance, wherein the third user utterance comprises the wake word and a third command;

determining that the third command does not comprise a command to change to the first operating mode; and

at least partly in response to determining that the third command does not comprise a command to change to the first operating mode, refraining from performing an action corresponding to the third command.

13. The computing system as recited in claim 12, wherein the function comprises:

receiving an output audio signal from the remote device, wherein the output audio signal corresponds to the function; and

playing the output audio signal, wherein the function is not interrupted such that the playing of the output audio signal continues after entering the second operating mode.

14. The computing system as recited in claim 13, wherein the determining that the first user utterance comprises the wake word is based on one or more recognition settings of a first operating mode.

15. The computing system as recited in claim 12, the computer-executable instructions maintained in the one or more computer-readable storage media further executed by the one or more hardware processors to cause the one or more hardware processors to perform acts comprising:

in the second operating mode;

receiving a fourth input audio signal corresponding to a fourth user utterance, wherein the fourth user utterance comprises a fourth command to enter the first operating mode;

determining that the fourth command comprises a command to change to the first operating mode; and

storing an indication that the computing system is operating in the first operating mode.

16. The computing system as recited in claim 15, wherein the determining that the first user utterance comprises the wake word is performed using first recognition settings corresponding to the first operating mode;

and

the determining that the fourth command comprises a command to enter the first operating mode is performed using second recognition settings corresponding to the second operating mode.