Granted Patent B2

US 12,614,549 · App. 18/431,516 · Granted Apr 28, 2026

Inventors: Frank Maker (Livermore, CA); Andrey Eltsov (Pleasanton, CA); Robert Curtis (Los Gatos, CA); Gregory Medding (San Jose, CA)

Assignee: Roku, Inc.

G10L15/22G10L15/08G10L21/0232G10L2015/088G10L2021/02082

Abstract

Claims (63)

1 . A computer implemented method for selecting a software agent, comprising:

receiving, by at least one processor, a modulated audio stream;

calculating, using automated speech recognition, a first confidence score that the modulated audio stream contains a first trigger word, wherein the first trigger word is mapped to a first software agent among a plurality of software agents stored in a voice platform;

calculating, using automated speech recognition, a second confidence score that the modulated audio stream contains a second trigger word, wherein the second trigger word is mapped to a second software agent, and the second trigger word is different from the first trigger word, and wherein the second software agent is configured to process the modulated audio stream differently than the first software agent;

determining the first confidence score is higher than the second confidence score; and

in response to the determining, transmitting a request to the voice platform to select the first software agent from among the plurality of software agents stored in the voice platform.

2 . The computer implemented method of claim 1 , further comprising:

normalizing the first confidence score based on a first confidence score criteria.

3 . The computer implemented method of claim 1 , further comprising:

determining the first confidence score exceeds a threshold value.

4 . The computer implemented method of claim 1 , further comprising:

determining that the first confidence score was calculated within a timeout period.

5 . The computer implemented method of claim 1 , wherein the determining further comprises:

determining the first confidence score and the second confidence score are within an error range; and

determining the first confidence score is higher than the second confidence score based on a priority criteria.

6 . The computer implemented method of claim 1 , wherein the receiving the modulated audio stream comprises:

receiving the modulated audio stream from an audio responsive electronic device, wherein the audio responsive electronic device is configured to:

output a tone at a speaker;

determine an echo cancellation delay based on a time of receipt of the tone at a microphone; and

perform echo cancellation on the modulated audio stream based on the echo cancellation delay.

7 . The computer implemented method of claim 1 , further comprising:

in response to the transmitting, providing an indication that the first software agent was selected to an audio responsive electronic device.

8 . A media device, comprising:

a memory; and

at least one processor coupled to the memory and configured to:

receive a modulated audio stream;

calculate, using automated speech recognition, a first confidence score that the modulated audio stream contains a first trigger word, wherein the first trigger word is mapped to a first software agent among a plurality of software agents stored in a voice platform;

calculate, using automated speech recognition, a second confidence score that the modulated audio stream contains a second trigger word, wherein the second trigger word is mapped to a second software agent, and the second trigger word is different from the first trigger word, and wherein the second software agent is configured to process the modulated audio stream differently than the first software agent;

determine the first confidence score is higher than the second confidence score; and

in response to the determination, transmit a request to the voice platform to select the first software agent from among the plurality of software agents stored in the voice platform.

9 . The media device of claim 8 , wherein the at least one processor is further configured to:

normalize the first confidence score based on a first confidence score criteria.

10 . The media device of claim 8 , wherein the at least one processor is further configured to:

determine the first confidence score exceeds a threshold value.

11 . The media device of claim 8 , wherein the at least one processor is further configured to:

determine that the first confidence score was calculated within a timeout period.

12 . The media device of claim 8 , wherein to determine the first confidence score is higher than the second confidence score, the at least one processor is further configured to:

determine the first confidence score and the second confidence score are within an error range; and

determine the first confidence score is higher than the second confidence score based on a priority criteria.

13 . The media device of claim 8 , wherein the at least one processor is further configured to:

receive the modulated audio stream from an audio responsive electronic device, wherein the audio responsive electronic device is configured to:

output a tone at a speaker;

determine an echo cancellation delay based on a time of receipt of the tone at a microphone; and

perform echo cancellation on the modulated audio stream based on the echo cancellation delay.

14 . The media device of claim 8 , wherein the at least one processor is further configured to:

in response to the transmission, provide an indication that the first software agent was selected to an audio responsive electronic device.

15 . A non-transitory computer-readable medium having instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to perform operations comprising:

receiving a modulated audio stream;

calculating, using automated speech recognition, a first confidence score that the modulated audio stream contains a first trigger word, wherein the first trigger word is mapped to a first software agent among a plurality of software agents stored in a voice platform;

calculating, using automated speech recognition, a second confidence score that the modulated audio stream contains a second trigger word, wherein the second trigger word is mapped to a second software agent, and the second trigger word is different from the first trigger word, and wherein the second software agent is configured to process the modulated audio stream differently than the first software agent;

determining the first confidence score is higher than the second confidence score; and

in response to the determining, transmitting a request to the voice platform to select the first software agent from among the plurality of software agents stored in the voice platform.

16 . The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

normalizing the first confidence score based on a first confidence score criteria.

17 . The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

determining the first confidence score exceeds a threshold value.

18 . The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

determining that the first confidence score was calculated within a timeout period.

19 . The non-transitory computer-readable medium of claim 15 , wherein the determining further comprises:

determining the first confidence score and the second confidence score are within an error range; and

determining the first confidence score is higher than the second confidence score based on a priority criteria.

20 . The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

in response to the transmitting, providing an indication that the first software agent was selected to an audio responsive electronic device.

Assignments (2)

Continuity (4)

References Cited (227)