摘要 |
PROBLEM TO BE SOLVED: To provide an on-vehicle voice recognition device capable of enhancing the accuracy of a recognition result by providing a proper ejecting value in accordance with a primary noise level in a compartment. SOLUTION: In this voice recognition device, the noise level in the compartment of a vehicle is estimated based on the driving parameter of the vehicle and a threshold is set in accordance with the estimated noise level and distance values indicating the nearness of similarities among reference voice congnition pattern data of preliminarily determined plural character strings and input voice pattern data corresponding to an inputted voice signal are respectively calculated and when the distance value of the minimum value among the calculated distance values is equal to or smaller than the thereshold, a character string corresponding to the distance value of the minimum value is recognized to be the character string indicated by the inputted voice signal.
|