摘要 |
The present invention concerns a speech recognition system (8) and a method of recognizing spoken utterances of users. The speech recognition system (8) comprises the plurality of speech handling units (81-84), a scheduler (85) and a plurality of speech recognizers (86-88). A prompt is presented to the user. The speech handling unit (81) receives a spoken utterance from the user in response to that prompt. It generates a speech signal from the received spoken utterance, detects the start and the end of said utterance and creates a speech data set containing the content of the utterance. It sends a speech recognition request message to the scheduler (85). The scheduler (85) dynamically allocates one of the speech recognizers (86-88) to the speech data set, when the speech data set is available. The speech handling unit (81) transfers the speech data set to the allocated speech recognizer (86).
|