A.I.Loskutov - Ph.D., Head of Department of telemetry systems and integrated information processing, Military Space Academy named after A.F. Mozhaysky (St.Peterburg). E-mail: rujenz@mail.ru A.J. Rashhupkin - Applicant, Department of telemetry systems and integrated information processing, Military Space Academy named after A.F. Mozhaysky (St.Peterburg). E-mail: raschupkin@yandex.ru A.S. Dunikov - Post-Graduate Student, Department of telemetry systems and integrated information processing, Military Space Academy named after A.F. Mozhaysky (St.Peterburg). E-mail: artem.sever1@yandex.ru

With the development of rocket and space technology to the radio telemetry system of increasingly high requirements in terms of reliability and accuracy of receiving telemetry data. Analysis of the processes of transmission and reception via radio channel telemetry data showed that an important element influencing the quality of receiving the information system is the symbol tim-ing. One of the promising of approaches which would ensure stability of the system symbol synchronization at a small signal / noise ratio, an approach based on the construction of symbol timing as intellectual recognition system. At the same time the most effi-cient apparatus for implementation of this approach relate symbol synchronization artificial neural networks. The proposed neural network algorithm symbol synchronization based on the classification of spectral images obtained on the basis of spectral and time-frequency analysis signal fragments. The drawn theoretical analysis of spectral and time-frequency signal analysis, and simulation results showed that the most suitable method for spectral images of binary symbols group tele-metry signal and their boundaries is a method based on a calculation of the power spectral density of the signal analyzed frag-ment. The simulation results have allowed to identify the signs that describe the spectral images of the binary symbols group telemetry signal and their boundaries, which include: - The number of of spectral peaks in the image; - value of the frequency of a spectral component with maximum amplitude; - Width of the spectrum in the image (fact determines plateout of the spectrum). Proposed to in Article symbol synchronization algorithm based on neural networks using spectral analysis group telemetry signal allows to combine the advantages high informative methods of digital signal processing and of neural networks the ability to effi-ciently process the data in terms of distortion accepted information. When using this algorithm and the process of determining the boundaries of a demodulation symbol is performed within a single of processing contour information registered surface equipment. Consequently, the use of symbol synchronization algorithm based on neural networks will result in faster occurrence in synchro-nization ground receiving and recording equipment.

 

1. Skljar B. Cifrovaja svjaz. Teoreticheskie osnovy i prakticheskoe primenenie. Izd. 2-e, ispr. M.: Izdatelskijj dom «Viljams». 2003. 1104 s.
2. Belickijj V.I., Zverev V.I. Telemetrija. SPb.: MO SSSR. 1984. 465 s.
3. Sergienko A.B.Cifrovaja obrabotka signalov. izd. 3-e. SPb.: BKHV-Peterburg. 2011. 768 s.
4. Marpl-ml. S.L.Cifrovojj spektralnyjj analiz i ego prilozhenija: Per. s angl. M.: Mir. 1990. 547 s.
5. Malla S. Vejjvlety v obrabotke signalov: Per. s angl. M.: Mir. 2005. 671 s.
6. KHarkevich A.A. Spektry i analiz. Izd. 4-e. M.: Gosudarstvennoe izdatelstvo fiziko-matematicheskojj literatury. 1962. 236 s.
7. Fukunaga K. Vvedenie v statisticheskuju teoriju raspoznavanija obrazov: Per. s angl. / Pod red. A.A. Dorofejuk. M.: Glavnaja redakcija fiziko-matematicheskojj literatury. 1979. 368 s.
8. Mesteckijj L.M.Matematicheskie metody raspoznavanija obrazov. M.: MGU. 2004. 85 s.
9. Loskutov A.I., Vecherkin V.B., SHestopalova O.L. Avtomatizacija kontrolja sostojanija slozhnykh tekhnicheskikh sistem na osnove ispolzovanija konechno-avtomatnojj modeli i nejjrosetevykh struktur // Informacionno-upravljajushhie sistemy. 2012. № 2. S. 74-81.