Method for voice recognition and electronic device for performing same

A method for recognizing voice inputs in an electronic device, the method comprising:

• storing in memory a plurality of applications operable with voice inputs;
• converting, by a microphone, an utterance of a user into an audio signal;
• generating, by at least one processor, text data corresponding to the audio signal;
• transmitting, by a communication circuit, the text data to a first NLU server associated with a first application from among the stored plurality of applications and a second NLU server associated with a second application from among the stored plurality of applications, wherein the first and second NLU servers are external to the electronic device;
• respectively receiving, by the communication circuit, a first control message and a second control message from the first NLU server and the second NLU server after the first and second NLU servers analyze the text data;
• selecting one of the first control message or the second control message, based on one of the first and second control messages having a confidence above a prespecified threshold;
• when both the first and second control messages have confidences below the prespecified threshold, selecting one of the first and second control messages based on at least one of:
• which of the first and second control message is first received,
• which of the first and second control message corresponds to an application with a higher execution frequency, and
• which of the first and second control message corresponds to a most recently executed application among the first and second applications;
• and
• providing the selected control message to an application associated with one of the first and second NLU servers that transmitted the selected control message.

An electronic device comprising:

• an audio input module;
• a memory configured to store a speech recognition application, a first application, and a second application;
• a communication circuit configured to communicate with a first NLU server associated with the first application and a second NLU server associated with the second application, wherein the first and second NLU servers are external to the electronic device;
• and
• a processor electrically connected to the audio input module, the memory, and the communication circuit and configured to execute the speech recognition application, wherein the processor is configured to:
• convert an utterance of a user received through the audio input module, into an audio signal;
• transmit text data corresponding to the audio signal to the first NLU server and the second NLU server;
• receive a first control message as a result of analyzing the text data, from the first NLU server;
• receive a second control message as a result of analyzing the text data, from the second NLU server;
• select one of the first control message or the second control message, based on one of the first control message or the second control message having a confidence above a prespecified threshold;
• when the first and second control messages have confidences below the prespecified threshold, selecting one of the first control message or the second control message based on at least one of:
• which of the first and second control messages is first received,
• which of the first and second control messages has a higher execution frequency, and
• which of the first and second control messages corresponds to a most recently executed application among the first and second applications, and
• provide the first control message to the first application, when the first control message is selected;
• and
• provide the second control message to the second application, when the second control message is selected.