| 摘要 |
PROBLEM TO BE SOLVED: To precisely recognize a driver's voice by accurately detecting a voice section while taking various situations of the driver into consideration. SOLUTION: A voice recognizing device 1 detects the start of the voice input by comparing with an utterance start condition when a voice is inputted from a microphone 3, and detects the end of the voice input when a state wherein the voice input level is lower than a previously set level lasts for a voice end detection time, thereby subjecting the inputted speech to recognition processing. The utterance end detection time is set longer for vehicle information inputted from an in-vehicle CAN communication network 2 and operation information on the driver (voice input of an address, in a curve travel, turn signal switch ON, shift position R, acceleration/deceleration time). COPYRIGHT: (C)2006,JPO&NCIPI
|
