Автоматизированная система дикторонезависимого голосового русскогоязычного управления операционной системой Windows

This study examines the methods and ways of organizing the recognition of Russian speech commands. The aim of the study is to establish a system for recognizing speaker independent Russian voice commands. Among the methods considered algorithms such as MFCC (Mel-Frequency Cepstral Coefficients), LPCC (Linear Predictive Cepstral Coefficients), PLP (Perceptual Linear Predictive). It is also not rejected from the consideration of ready-made software solutions and modules from leading companies with expertise in the field of speech technology. It was supposed to use one of the algorithms, which itself most prove obtain evidence on the quality of voice in the synthesis of a neural network learning system. During the analysis and testing is not one of the algorithms is not proven itself to create speaker independent system. In the end, was selected recognition module Speech API from Microsoft. The only drawback is the lack of module support for the Russian language, but the problem is solved by creating a class of software that converts typed words in Russian English-language transcription of the module supports recognition. To demonstrate the potential of using this method has been developed automated system for Russian-speaking speaker independent voice control of the operating system Microsoft Windows.