The speech interface for intelligent robots

An appropriate interface is not only to providing a better controlled environment, but also increasing the willing of the using by operator. This paper presents a robot speech user interface's design. The system is based on a continuous Gaussian Hidden Markov model as a basis for using Viterbi algorithm in the speech of modeling and identification procedure. In Visual Basic 6.0 environment, the system designed 7 robot's commands used to manipulate the robot to forward, backward, turn around ...etc seven different movements. Under usually circumstances, most of them would make it have a correct operation; the identification rate up to about 93%, its effect is quite good.