# US 12,603,089 B2 · Electronic apparatus performing speech recognition and method for controlling thereof > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,603,089 B2 - Document ID: US12603089B2 - Status: Granted - Kind: B2 - Application number: 18/384,607 - Filed: 2023-10-27 - Published: 2024-03-21 - Granted: 2026-04-14 - Earliest claimed: 2022-09-15 - Assignees: Samsung Electronics Co., Ltd. - Applicants: Samsung Electronics Co., Ltd. - Inventors: Saeyoung Kim; Yongkook Kim; Junghoe Kim; Hyeontaek Lim; Boseok Moon - Web page: https://patentgrep.com/patent/US12603089B2 ## CPC - G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue - G06F 3/167: {Audio in a user interface, e.g. using voice commands for navigating, audio feedback} - G06F 40/247: Thesauruses; Synonyms - G06F 40/30: Semantic analysis - G10L 15/1815: {Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning} - G10L 15/187: Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams - G10L 2015/223: {Execution procedure of a spoken command} - G10L 2015/228: {of application context} - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Disclosed is an electronic apparatus which performs speech recognition. The apparatus includes: at least one sensor; a memory storing context data each corresponding to a plurality of commands; a microphone; and at least one processor configured to acquire a text corresponding to a voice command in case that the voice command is received through the microphone, identify context data corresponding to an identified command in case that a command having similarity of a first threshold value or more with the acquired text is identified among the plurality of commands stored in the memory, acquire a control signal corresponding to the identified command in case that current context data including context data acquired through the at least one sensor is identified as having similarity of a second threshold value or more with the context data corresponding to the identified command, and perform an operation based on the control signal. ## Claims ### Claim 1 An apparatus comprising: at least one sensor; a memory storing a plurality of context data each corresponding to one of a plurality of commands; a microphone; and at least one processor configured to: acquire text data corresponding to a voice command received through the microphone, identify at least one candidate command, among the plurality of commands, based on a comparison between the acquired text data and the plurality of commands, identify a first command, among the at least one candidate command based on intent data corresponding to each of the at least one candidate command, each of the intent data indicating an intent of the user corresponding to the respective candidate command, identify first context data, among the plurality of context data, corresponding to the identified first command, acquire a control signal corresponding to the identified first command based on a comparison between the first context data corresponding to the identified first command and second context data acquired through the at least one sensor corresponding to the voice command, and perform an operation based on the control signal. ### Claim 2 (depends on 1) The apparatus as claimed in claim 1, wherein the at least one processor is further configured to identify the acquired text data as first data in phoneme units, identify a plurality of similarity values, each corresponding to a level of similarity between the acquired text data and a respective one of the plurality of commands by comparing the identified first data with second data in the phoneme units corresponding to each the plurality of commands, and identify the first command based on a first similarity value, corresponding to the level of similarity between the acquired text data and the first command, being greater than or equal to a first threshold value. ### Claim 3 (depends on 2) The apparatus as claimed in claim 2, wherein the at least one processor is configured to: identify the first context data based on the first context data having a second similarity value, corresponding to a level of similarity between the first context data and the second context data, being greater than or equal to a second threshold value. ### Claim 4 (depends on 3) The apparatus as claimed in claim 3, wherein the memory stores text information for each of the plurality of commands and intent data corresponding to the plurality of commands, and wherein the at least one processor is further configured to: acquire the intent data corresponding to the voice command when the second context data is identified as having the similarity satisfying the second threshold value or more with first context data corresponding to the first command, and identify the first command by comparing the acquired intent data with the intent data for each text corresponding to each of the plurality of commands stored in the memory. ### Claim 5 (depends on 2) The apparatus as claimed in claim 2, wherein the at least one processor is further configured to: identify the first command corresponding to the acquired text data among the plurality of commands stored in the memory based on a first speech recognition module, and acquire a second command corresponding to the voice command based on a second speech recognition module to store the acquired second command and context data corresponding to the acquired second command in the memory. ### Claim 6 (depends on 5) The apparatus as claimed in claim 5, wherein the at least one processor is further configured to perform speech recognition on the voice command based on the second speech recognition module in case that the command having the similarity satisfying the first threshold value or more with the acquired text is not identified based on the first speech recognition module. ### Claim 7 (depends on 1) The apparatus as claimed in claim 1, wherein the first context data and the second context data comprise a plurality of types of data comprising at least one of position data of the electronic apparatus, state data of the electronic apparatus, environment data of the electronic apparatus, or context data of a user, and wherein the at least one processor is further configured to identify similarity between the first context data and the second context data by comparing each of the plurality of types of data included in the first context data with each of the plurality of types of data included in the second context data. ### Claim 8 (depends on 1) The apparatus as claimed in claim 1, wherein the at least one processor is further configured to: identify the first command further based on a priority associated with each of the intent data corresponding to each of the at least one commands. ### Claim 9 (depends on 1) The apparatus as claimed in claim 1, wherein the at least one processor is further configured to delete at least one of the plurality of commands from the memory based on at least one of use history or priority of the plurality of commands stored in the memory. ### Claim 10 (depends on 1) The apparatus as claimed in claim 1, wherein the at least one processor is further configured to: identify a weight corresponding to each type of a plurality of data of different types included in the second context data, and identify similarity between the first context data and the second context data based on the identified weight. ### Claim 11 (depends on 1) The apparatus as claimed in claim 1, wherein the at least one processor is further configured to: identify the first command by comparing the acquired text data in phoneme units with second data in the phoneme units corresponding to each the plurality of commands. ### Claim 12 A method for controlling an electronic apparatus which performs speech recognition, the method comprising: acquiring text data corresponding to a voice command; identifying at least one candidate command, among a plurality of commands, based on a comparison between the acquired text data and the plurality of commands; identifying a first command, among the at least one candidate command based on intent data corresponding to each of the at least one candidate command, each of the intent data indicating an intent of the user corresponding to the respective candidate command; identifying first context data, among a plurality of context data, corresponding to the identified first command, each of the plurality of context data corresponding to one of a plurality of commands; acquiring a control signal corresponding to the identified first command based on a comparison between the first context data corresponding to the identified first command and second context data acquired through at least one sensor corresponding to the voice command; and performing an operation based on the control signal. ### Claim 13 (depends on 12) The method as claimed in claim 12, wherein the identifying of the first command comprises: identifying the acquired text data as first data in phoneme units; identifying a plurality of similarity values, each corresponding to a level of similarity between the acquired text data and a respective one of the plurality of commands by comparing the identified first data with second data in the phoneme units corresponding to each the plurality of commands, and identifying the first command based on a first similarity value corresponding to the level of similarity between the acquired text data and the first command being greater than or equal to a first threshold value. ### Claim 14 (depends on 13) The method as claimed in claim 13, wherein the acquiring first context data comprises: identifying the first context data based on the first context data having a second similarity value corresponding to a level of similarity between the first context data and the second context data being greater than or equal to a second threshold value. ### Claim 15 (depends on 12) The method as claimed in claim 12, wherein the first context data and the second context data comprise a plurality of types of data comprising at least one of position data of the electronic apparatus, state data of the electronic apparatus, environment data of the electronic apparatus, or context data of a user, and wherein the method further comprises: identifying similarity between the first context data and the second context data by comparing each of the plurality of data included in the first context data with each of the plurality of data included in the second context data. ### Claim 16 (depends on 12) The method as claimed in claim 12, wherein the identifying of the first command comprises: identifying the first command further based on a priority associated with each of the intent data corresponding to each of the plurality of commands. ### Claim 17 (depends on 12) The method as claimed in claim 12, wherein the first command is identifying by comparing the acquired text data in phoneme units with second data in the phoneme units corresponding to each the plurality of commands.