International Science and Technology Journal

In recent years, a great deal of attention has been paid by several international scientific institutions and organizations to search for the development of human communication to machines (machine learning). Researchers in these institutions have gone through many attempts which most of them failed and others partially succeeded in some way that made other researchers to reach the stage of understanding the machine to human voice even if few command words to implement simple orders. Therefore, our propose in this paper is figuring out findings on machine recognition for a particular human voice by comparing the frequency spectrum of that voice with the frequency spectrum components of the voice stored in the memory of that machine using the magnitude-squared coherence function . Accordingly, a reaction is made by the machine to that command word