S.N. Kirillov – Dr. Sc. (Eng.), Professor, Head of Department «Radio control and communication», V.F. Utkin Ryazan State Radio Engineering University; Honored Worker of the Russian Higher School, Academician of the International Academy Of communications, corresponding Member of the Academy of Military Sciences

E-mail: kirillov.lab@mail.ru

A.N. Kochurkin – Doctor of higher qualification category, Head of LOR-Department, Regional clinical hospital n. a. N.Ah. Semashko (Ryazan)

E-mail: akuchurkin@list.ru

D.Yu. Mamushev – Candidate of Department «Radio control and communication», V.F. Utkin Ryazan State Radio Engineering University

E-mail: mamushev_du@bk.ru

The method of support of diagnostic decisions in diseases of the throat and nose of the upper respiratory tract (URT) based on the analysis of the parameters of the speech signal (SS). In addition to clinical methods of diagnosis of these diseases, an important task is the use of objective non-invasive methods, in particular in the parameters of MS. This will allow us to conduct medical examinations of children and adults pre-to determine the presence of upper respiratory tract disease and make decisions about the need for further treatment. To assess the diagnostic significance of SS parameters, it is proposed to analyze the changes in the values of these parameters in various diseases of URT on the basis of the method of group consideration of the argument (mgua).

The aim of the work is to investigate the influence of URT diseases on the parameters of MS, in particular to establish significant parameters of SS, allowing to diagnose such diseases as VDP, nasal septum curvature, polypous sinusitis, rhinitis, tonsillitis, as well as to justify the algorithm of classification of these diseases based on mgua. 

The influence of trade on such group characteristics of the PC, as the settings format, Mel-frequency cepstral coefficients, parameters of fundamental frequency (), energy of the signal, linear prediction coefficients, studied the parameters of nonlinear models RS, developed by the authors. At the same time, the frequency of OT, formants, dynamic range of the phonogram, jitter, shimmer were used previously in known works to determine violations of voice function in children and adults in the case of laryngeal pathologies and lung disease. The rest of the selected parameters of MS were not found in the works devoted to the definition of LOR diseases in adults, as well as in the problem of classification of diseases of URT and for the problem solved in the article are analyzed for the first time. Identified significant parameters for diagnosing diseases such as a deviated septum, polypous sinusitis, rhinitis, tonsillitis.

Justified algorithm for classification of these diseases in the parameters of the RS on the basis of the GMDH is used to identify the system characteristics and management of a small number of the original data. Application of the mgua algorithm for the problem solved in the article includes two points: 

1) Training to assess the significant parameters of the RS, the result is a vector of coefficients , and the space of diagnosis values for each disease sample and healthy people sample. 

2)Classification of the vector of parameters of the intended patient as belonging to one of the spaces, signs of patients or healthy people. The result is the probability of correct classification, depending on the number of significant parameters obtained in the previous stages of the study.

The dependence of SS parameters on the disease of URT in adults was established. Justified non-invasive method of supporting diagnostic solutions for diseases of the SDT in the parameters of the SS on the basis of GMDH. Application of the method allowed to establish a minimum sufficient number of parameters of MS, for which the probability of correct classification of the disease is 0.86-0.93. The method can be used both for remote preliminary diagnosis and in case of insufficient clinical symptoms in a comprehensive examination of patients.

1. Pankova V.B. Osnovnye napravleniya profilaktiki patologii lor-organov // XVI Rossijskij kongress otorinolaringologov «Nauka i praktika v otorinolaringologii». M. 2017. S. 19–21.
2. Kochneva A.O., Egorov V.I., Logutova L.S., Zajdieva YA.Z. Izuchenie vliyaniya patologii lor-organov na reproduktivnuyu sistemu zhenshchin detorodnogo vozrasta // XIV Rossijskij kongress otorinolaringologov «Nauka i praktika v otorinolaringologii». M. 2015. S. 18. 
3. Dyagtereva D.V. Ob"ektivnye neinvazivnye metody issledovaniya golosovoj funkcii u detej: Dis. … kand. med. nauk. M. 2014. 124 s.
4. Patent № 2598051 (RF), MPK A61B8/08. Sposob opredeleniya izmenenij golosovoj funkcii / E.A. Galifanov, V.A. Nevzorova, S.A. Artyushkin,  E.V. Krukovich, T.V. Tilik, L.B. Ardeeva, A.A. Kabieva.  
5. Aithal V.U. et al. Acoustic analysis of voice in normal and high pitch phonation: a comparative study // Folia Phoniatr Logop. 2012. V. 64. № 1. P. 48–53.
6. Ivahnenko A.G. Dolgosrochnoe prognozirovanie i upravlenie slozhnymi sistemami. Kiev: Tekhnika. 1975. 312 s. 
7. Bol'shev L.N., Smirnov N.V. Tablicy matematicheskoj statistiki. M.: Glavnaya redakciya fiziko-matematicheskoj literatury. 1983. 416 s.
8. Rabiner L., Gold B. Teoriya i primenenie cifrovoj obrabotki signalov. M.: Mir. 1978. 848 s. 
9. Kirillov S.N., Mamushev D.YU., Locmanov A.A. Nelinejnaya parametricheskaya model' rechevyh signalov v zadache identifikacii golosa diktora //  Sb. trydov XIII Mezhdunar. nauchnoj konf. «Informatizaciya i informacionnaya bezopasnost' pravoohranitel'nyh organov». M.: Akademiya upravleniya MVD Rossii. 2004. S. 394–397.