Research and Implementation of a Speaker-Dependent Speech Recognition Algorithm

Abstract

The goal of this research to realize the control humanoid type prosthetic movement used for speech recognition principles. Also, the basic principles of speaker-dependent speech recognition algorithms are introduced, a speaker-dependent speech recognition algorithm in the implementation process of an upper machine is described, and the feature template library of voice command is established. Ultimately, this information was used to test the performance of a speech recognition system. Experimental results show that this method could detect blinks at high accuracy signals and obtain an eigenvector, which achieved the goal of accurate prosthetic limb movement control.