System and method for customizing smart home speech interfaces using personalized speech profiles
Systems and methods are provided in which a speaker profile-data inquiry is transmitted to a mobile device associated with a first speaker. In response to the speaker-profile-data inquiry, speaker profile data associated with the first speaker is received. Audio data representing a voice input is received. The first speaker is identified as providing the voice input, the identification being based on a comparison of characteristics of the received audio data with the speaker profile data of a plurality of speakers for whom speaker profile data is stored. An instruction, which includes a speaker-relative signifier, is determined from the received audio data, and determining the instruction includes determining a referent of the speaker-relative signifier based on the first speaker profile data. An action indicated by the instruction is performed.
1. A method comprising:
transmitting a speaker-profile-data inquiry to a mobile device associated with a first user;
in response to the speaker-profile-data inquiry, receiving first speaker profile data associated with the first user, the first speaker profile data including voice-recognition data of the first user;
receiving audio data representing a voice input;
identifying the first user as providing the voice input, the identification being based on a comparison of characteristics of the received audio data with speaker profile data of a plurality of speakers for whom speaker profile data is stored, including the first speaker profile data;
determining an instruction from the received audio data, wherein the instruction references a second user other than the first user and includes a speaker-relative signifier, and wherein determining the instruction includes determining a referent of the speaker-relative signifier based on second stored speaker profile data associated with the second user; and
performing an action indicated by the instruction.
2. The method of claim 1 , wherein the speaker profile data of the plurality of speakers comprises voice-recognition data of the plurality of speakers.
3. The method of any of claim 2 , wherein the comparison comprises a comparison of the characteristics of the received audio data with the voice-recognition data.
4. The method of claim 1 , wherein the speaker profile data comprises speaker-customized vocabulary data.
5. The method of claim 1 , further comprising discovering the mobile device via a discovery protocol, wherein the speaker-profile-data inquiry is transmitted in response to discovering the mobile device.
6. The method of claim 1 , wherein the method is performed by a home automation hub.
7. The method of claim 1 , wherein identifying the first speaker comprises determining a first confidence score based on the received audio data and on the first speaker profile data.
8. The method of claim 7 , wherein identifying the first speaker from a plurality of speakers further comprises comparing the first confidence score to a threshold confidence score.
9. The method of claim 7 , wherein identifying the first speaker further comprises comparing the first confidence score to a second confidence score determined based on the received audio data and on speaker profile data associated with a user other than the first speaker.
10. The method of claim 1 , wherein the first speaker profile data includes at least one relationship mapping between the first user and the second user, and wherein a determination that the instruction references the second user is made based on the relationship mapping.
11. The method of claim 10 , wherein the first speaker profile data includes contact information, and wherein the relationship mapping between the first user and the second user is based on the contact information.
12. The method of claim 1 , wherein the first speaker profile data comprises a mapping between the referent and the speaker-relative signifier.
13. The method of claim 1 , wherein the speaker-relative signifier refers to a familial relation or a location associated with a user.
14. The method of claim 1 , wherein performing the action indicated by the instruction comprises a home automation hub controlling a second device or service.
15. The method of claim 1 , wherein performing the action indicated by the instruction comprises presenting multimedia content via a home automation hub.
16. The method of claim 1 , wherein performing the action indicated by the instruction comprises setting a user-access control rule via a home automation hub.
17. The method of claim 1 , wherein the speaker-relative signifier refers to a media object identified in the second stored speaker profile data, and wherein the instruction is an instruction to play the media object.
18. The method of claim 1 , wherein the speaker-relative signifier refers to a relationship mapping in the second stored speaker profile data, and wherein the instruction is an instruction to contact a person identified by the relationship mapping.
19. A system comprising:
a processor; and
a non-transitory storage medium storing instructions operative, when executed by the processor, to perform the functions of:
transmitting a speaker-profile-data inquiry to a mobile device associated with a first user;
in response to the speaker-profile-data inquiry, receiving first speaker profile data associated with the first user, the first speaker profile data including voice-recognition data of the first user;
receiving audio data representing a voice input;
identifying the first user as providing the voice input, the identification being based on a comparison of characteristics of the received audio data with speaker profile data of a plurality of speakers for whom speaker profile data is stored, including the first speaker profile data;
determining an instruction from the received audio data, wherein the instruction references a second user other than the first user and includes a speaker-relative signifier, and wherein determining the instruction includes determining a referent of the speaker-relative signifier based on second stored speaker profile data associated with the second user; and
performing an action indicated by the instruction.
20. The system of claim 19 , wherein the first speaker profile data includes at least one relationship mapping between the first user and the second user, and wherein a determination that the instruction references the second user is made based on the relationship mapping.