D. S. Makhov – Ph.D. (Eng.), Doctoral Candidate, Krasnodar highest military college n.a. General of the Army S.M. Shtemenko

E-mail: sinedvoham@mail.ru

One of the problems of application of systems with unmanned aerial vehicles is ensuring reliability of the transmitted data in radio channels. This problem is caused by increase in volume of the target information transferred by the unmanned aerial vehicle to a control ground station, restrictions for frequency band of channels, high speed of change of operating conditions of complexes, environment of distribution of signal, multibeam propagation of signal, power limiting on the transmitter, mass-dimensional characteristics and conditions of interfering situation. The classic methods of increase in noise stability on the basis of use of effective algorithms of digital information processing solve this problem partially. Development of new approaches and methods of search of optimization of the parameters influencing reliability of the transmitted data in radio channels of complexes with unmanned aerial vehicles is required. One of the approaches with flexibility at achievement of the optimal solution due to balance between encroaching speed to the decision and approximation accuracy is the mathematical apparatus of neural networks.

In the article the approach to representation of systems with unmanned aerial vehicles in the form of system with many inputs and many outputs which optimum parameters are defined on the basis of neural network approach has been offered. Creation of mathematical models of systems of information transfer in radio channels in the form of neural networks allows using advantages of the principles of functioning of neural networks for solution of problems of optimization of radio channels parameters. The mathematical model describing functioning of complex with unmanned aerial vehicles in the form of twolayer perceptron has been given.

The description of radio channels of systems with unmanned aerial vehicles by the modified mathematical model of neural network allows considering the nature of scaling and complication of system with many inputs and many outputs. The method of optimization of coefficients of transfer of radio channel in the form of weight functions of the received mathematical model of neural network with the return distribution of mistake has been offered. The method allows to optimize coefficients of neural network due to redistribution of capacities in system radio channels with many inputs and many outputs. Optimization is performed on the basis of formulation of requirements of the teacher in the form of required distribution of level of signals at the system exits.

The results of numerical researches of the proposed method have been given. They show optimization of weighting coefficients during the iterative procedure and approach to the required decision. The conducted researches give the optimal solution on the basis of the offered approach. Results of iterative optimization of weighting coefficients and signals in the system outputs have been given. The restrictions of achievement of the optimal solution which removal is the purpose of further researches have been given.

1. Makarenko S.I. Robototekhnicheskie kompleksy voennogo naznacheniya – sovremennoe sostoyanie i perspektivy razvitiya. Sistemy upravleniya, svyazi i bezopasnosti. 2016. № 2. S. 73–132. [in Russian]
2. Polynkin A.V., Le Kh.T. Issledovanie kharakteristik radiokanala svyazi s bespilotnymi letatel'nymi apparatami. Izvestiya TulGU. Tekhnicheskie nauki. 2013. Vyp. 7. Ch. 2. S. 98–106. [in Russian]
3. Marinho M.A.M., Ferreira Junior R.S. Antenna array based positioning scheme for unmanned aerial vehicles. Conference of Wireless System Application. Stuttgart, Germany. 2013. P. 1–6.
4. Boev N.M. Analiz komandno-telemetricheskoj radiolinii svyazi s bespilotnymi letatel'nymi apparatami. Vestnik Sibirskogo gosudarstvennogo aerokosmicheskogo universiteta im. akademika M.F. Reshetneva. SibGU. 2012. S. 86–91. [in Russian]
5. Boev N.M. Sposoby povysheniya energeticheskoj i spektral'noj effektivnosti tsifrovykh sistem svyazi bespilotnykh letatel'nykh apparatov. Radiofizika, radiotekhnika, svyaz'. Trudy MFTI. 2014. T. 6. № 2. S. 162–166. [in Russian]
6. Tuzov G.I., Sivov V.A., Prytkov V.I. i dr. Pomekhozashchishchennost' radiosistem so slozhnymi signalami. M.: Radio i svyaz'. 1985.
7. Bakulin M.G., Varukina L.A., Krejndelin V.B. Tekhnologiya MIMO: printsipy i algoritmy. M.: Goryachaya liniya – Telekom. 2014.
8. Lazarev V.M., Sviridov A.P. Nejroseti i nejrokomp'yutery. Monografiya. M.: Akademiya. 2011.
9. Sklyar B. Tsifrovaya svyaz'. Teoreticheskie osnovy i prakticheskoe primenenie. Izd. 2-e, ispr.: Per. s angl. M.: ID «Vil'yams». 2003.
10. Ermolaev V.T., Mavrychev E.A., Flaksman A.G. Umen'shenie veroyatnosti bitovoj oshibki pri parallel'noj peredache informatsii v MIMO sisteme. Izvestiya VUZov. Radiofizika. 2003. T. 46. № 3. S. 251–260. [in Russian]
11. Galushkin A.I. Teoriya nejronnykh setej. Kn. 1. Ucheb. posobie dlya vuzov / Pod obshch. red. A.I. Galushkina. M.: IPRZhR. 2000.
12. Manuilov B.D., Manuilov M.B. Ustrojstva SVCh i antenny. Rostov-na-Donu: MO RF. 2006.
13. Korn G., Korn T. Spravochnik po matematike (dlya nauchnykh rabotnikov i inzhenerov). M.: Nauka. 1974.
14. Zyuko A.G., Klovskij D.D., Nazarov M.V., Fink L.M. Teoriya peredachi signalov. Uchebnik dlya vuzov. M.: Svyaz'. 1980.