On the Issue of Speaker's Identification in Communication Networks and Terminal Equipment of Onboard System

Abstract

The work is dedicated to the issue of speech messages quality providing in communication networks, in particular, to study the influence of the state of functional systems of the speaker on the accuracy of his speech identification. The study allowed drawing a conclusion on quite clearly formalized correspondence between changes in speech signal parameters and disease groups that define them. Such formalization can be used to increase the identification reliability level of speech sources, and for timely and rapid diagnosis of a certain set of diseases, including their early development stage. Adequate identification of speech messages sources belongs to a common task group of ensuring effective and safe operation of automated control systems as well as access control systems. The second group of tasks, in some cases, may be associated with the first one and is caused by the need of personnel health state monitoring implementing management functions of important and challenging complex systems and processes. Also, the detection of several diseases signs at their early stages is very important task to ensure the implementation of research and production programs made by members of various expeditions teams, in particular, the crews working on board space stations. As a rule, the correspondent decision - making is usually based on the correlation analysis of speech signals, as well as on the evaluation of phonogram statistical parameters. The paper proposes and justifies the algorithm for speech source identification, in condition of a set of endogenous and exogenous factor presence, affecting its informative parameters.