Dialogue processing apparatus, a vehicle having same, and a dialogue processing method

A vehicle comprising:

• a microphone configured to receive a natural language speech utterance by being always opened and configured to convert the natural language speech utterance into an electrical input signal;
• a processor;
• a storage device configured to store the level of a background signal;
• and
• a non-transitory memory configured to store computer executable code, the non-transitory memory coupled to the processor to cause the processor to implement:
• a monitor to receive the electrical input signal and to compare a level of the electrical input signal with a reference level in real time and configured to determine that a natural language speech utterance of a user is input when the level of the electrical input signal is greater than the reference level;
• a speech recognizer to receive, from the monitor, the electrical input signal and to output an utterance in the form of text by performing speech recognition on the electrical input signal, wherein the speech recognition extracts a feature vector of the electrical input signal by applying a feature extraction technique and compares the extracted vector with a trained reference pattern;
• a natural language processor to extract a domain and a keyword based on the utterance regardless of whether the utterance includes a trigger word, when it is determined that the natural language speech utterance of the user is input;
• and
• a dialogue manager configured to determine whether a previous context is maintained based on the domain and the keyword,
• wherein the reference level is determined based on a level of a background signal that is input in a state in which the natural language speech utterance is not uttered,
• wherein the storage device determines that the electrical input signal is the background signal when the level of the electrical input signal is equal to or less than the reference level, and
• further wherein the storage device updates the stored level of the background signal using the level of the electrical input signal.

A dialogue processing method comprising:

• receiving a natural language speech utterance and converting the received natural language speech utterance into an electrical signal;
• comparing a level of the electrical input signal input via a microphone that is always opened with a reference level in real time;
• determining that the natural language speech utterance of a user is input when the level of the electrical input signal is greater than the reference level;
• performing speech recognition on the electrical input signal and outputting an utterance in the form of text, wherein the speech recognition extracts a feature vector of the electrical input signal by applying a feature extraction technique and compares the extracted vector with a trained reference pattern;
• and
• extracting a domain and a keyword based on the utterance regardless of whether the utterance includes a trigger word, when it is determined that the natural language speech utterance of the user is input,
• wherein the reference level is determined based on a level of a background signal that is input in a state in which the natural language speech utterance is not uttered, and
• further comprising:
• determining that the electrical input signal is the background signal when the level of the input signal is equal to or less than the reference level;
• and
• updating a stored level of the background signal using the level of the electrical input signal.