S.V. Chelebaev - Ph.D. (Eng.), Associate Professor, Department «Biomedical and Semiconductor Electronics», Ryazan State Radio Engineering University. E-mail: sergey_chel_r@rambler.ru
Y.A. Chelebaeva - Student, Department «Biomedical and Semiconductor Electronics», Ryazan State Radio Engineering University. E-mail: sergey_chel_r@rambler.ru

Now the device of artificial neural networks is an effective remedy on the basis of which process of the information form conversion devices formalized synthesis, difinienda as neuronetwork converters is carried out. These devices operate with the variables provided in the form of frequency, duration of time slot, the period of signal change, its amplitude in the tension form, number - impulse or positional codes, and represent an analog-to-digital neuronet. The linear and non-linear analog-to-digital converters neuronetwork synthesis questions implementing the functional dependences of one variable, are illuminated in known publications. However the neuronetwork converters synthesis questions, implementing functional dependences of two variables aren't described. Reproduction need of two and bigger variables number functions of usually face in case of the ballistic and navigation tasks solution, tasks of control and monitoring by moving objects, different technological processes, and also the tasks connected to research and simulation of autoregulation difficult systems. Converters synthesis pocedures of time-and-frequency signals parameters in a digital code of two variables on the example of the perceptron networks application are developed. Converters structures of time-and-frequency signals parameters in a digital unitary and position code of two variables on a basis one - and three-layer perceptrons are respectively constructed. The network training example of the frequency converter in a position code of two variables on the three-layer perceptron basis is given. The neuronetwork converter structure of the time interval in a positional code of two variables on the basis of three-layer perceptron is implemented on hardware description language VHDL.

 

1. Galushkin A.I. Nejronnye seti: osnovy teorii / M.: Gorjachaja linija-Telekom. 2010.
2. Loktjuhin V.N., Chelebaev S.V. Nejrosetevye preobrazovateli impul'sno-analogovoj informacii: organizacija, sintez, realizacija / Pod obsh. red. A.I. Galushkina. M.: Gorjachaja linija - Telekom. 2008.
3. David W. Tank, John J. Hopfield. Simple "neural" optimization networks: an A/D converter, signal decision circuit and a linear programming circuit // IEEE Circuit and Systems. V. CAS-33. May 1986. S. 533-541.
4. Bang W. Lee, Bing J. Shen. Design of a neural-based A/D converter using modified Hopfield Network // IEEE Solid-State Circuits. V. SC-24. Aug 1989. P. 1120-1135.
5. Avitabile G., Forti M., Manetti S., Marini M. On a class of nonsymmetrical neural networks with application to ADC // IEEE Circuit and Systems. V. CAS-38. Feb 1991. P. 202-209.
6. Loktjuhin V.N., Chelebaev S.V., Antonenko A.V. Nejrosetevye analogo-cifrovye preobrazovateli / pod obsh. red. A.I. Galushkina. M.: Gorjachaja linija-Telekom. 2010.
7. Loktjuhin V.N., Chelebaev S.V. Principy primenenija tehnologii iskusstvennyh nejronnyh setej dlja proektirovanija preobrazovatelej chastotno-vremennyh parametrov signalov v kod // Datchiki i sistemy. 2006. №2. S. 43-56.
8. Jasnickij L.N. Vvedenie v iskusstvennyj intellekt / Ucheb. posobie dlja studentov VUZov. M.: Izdatel'skij centr "Akademija". 2005.
9. Loktjuhin V.N., Chelebaev S.V. Primenenie rekurentnyh setej dlja postroenija impul'sno-cifrovyh preobrazovatelej // Vestnik Rjazanskogo gosudarstvennogo radiotehnicheskogo universiteta. 2006. № 19. S. 95-103.
10. Loktjuhin V.N., Chelebaev S.V., Antonenko A.V. Algoritmy nastrojki perseptronnyh setej na preobrazovanie chastotno-vremennyh parametrov signalov v kod // Vestnik Rjazanskogo gosudarstvennogo radiotehnicheskogo universiteta. 2009. № 30. S. 27-35.