Identification of target scatterers in radar images using radial basis function neural networks

Radio engineering. Electronics. Telecommunication systems


Аuthors

Efimov E. N.*, Shevgunov T. Y.**

Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia

*e-mail: omegatype@gmail.com
**e-mail: shevgunov@gmail.com

Abstract

This paper introduces a neural network based technique for the identification of complex targets in ultra-wideband radar systems. This technique is based on the known multi-scatterer complex-valued model for representing two-dimension radar images synthesized by common ultra-wideband radar systems. The neural networks are built via structure diagrams representing the whole net implementing the desired transformation as a set of interacting adaptive elements, each of them in turn implements the partially transfer functions of bidirectional (forward and backward propagation) data signal processing. The class of radial basis feed-forward artificial neural networks learned by error back-propagation supervised methods is chosen due to the ability of such networks to highlight the local feature of the presented data. It was shown that a stable approximation result could be achieved if the number of elements in hidden layer of the networks corresponds to the number of scatterers necessary for the accurate fitting of artificial targets. The simulation showed that the importances of scatterer parameters are different, thus the estimation of center positions of scatterers can be carried out more accurately than their widths. The accuracy of the scatterer parameters estimation depends on the distance between a chosen scatterer and others and signal to noise ratio. The simulation results reveal the high sensitivity of the accuracy to the level of noise while scatterers are placed closely in the radar image. The proposed neural processing method is a relatively new but promising approach for the problem to be solved. The main benefits and drawbacks of this methods are examined in the conclusion as well as comparison with another system identification technique traditionally used for the considered problem.

Keywords:

, ultra-wideband radar, multi-scatterer model, artificial neural network, radial basis functions, system identification, ultra-wideband systems, mathematical model simulation

References

  1. Tatuzov A.L., Nejronnyeseti v zadachah radiolokacii (Neural networksin the problemsof radar), Moscow, Radiotehnika, 2009, 432 p.
  2. Rutkovskaja D., Pilin'skij M., Rutkovskij L. Nejronnye seti, geneticheskie algoritmy i nechjotkie sistemy (Neural networks, genetic algorithms andfuzzy systems), Moscow, Gorjachaja linija — Telekom, 2006, 452 p.
  3. Immoreev I.Ja., Radiotehnika i elektronika, 2009, vol. 54, no.1, pp. 5–31.
  4. Efimov E.N., Shevgunov T.Ja. Elektronnyi zhurnal “Zhurnal radiojelektroniki, ISSN 1684-1719, 2012, no.8, available at: http://jre.cplire.ru/win/aug12/index.html (accessed 08.2012).
  5. Konovaljuk M.A., Kuznecov Ju.V., Baev A.B. Vestnik Moskovskogo aviatsionnogo instituta, 2010, vol. 17, no. 3, pp. 193-198.
  6. August W. Rihaczek, Stephen J. Hershkowitz. Theory and practice of radar target identification, Artech House, 2000, 738 p.
  7. Hagan M.T., Demuth H.B., Beale M.H., Neural Network Design, Martin Hagan, 2002, 736 p.
  8. Tadeusevich R. Elementarnoe vvedenie v tekhnologiyu neironnykh setei s primerami programm (An elementary introduction to the technology of neural networks with examples of programs), Moscow, Gorjachajalinija - Telekom, 2011, 408 p.
  9. Hajkin S., Neironnye seti (Neural Networks), Moscow, 2006, 1104 p.
  10. T. Shevgunov, E. Efimov, Patent RU 2012610557, 10.01.2012.
  11. Efimov E.N., Shevgunov T.Ja. Еlektronnyi zhurnal “Trudy MAI”, 2012, no. 51, available at: http://www.mai.ru/science/trudy/published.php?ID=28948(26.03.2012).
  12. Samarasinghe S. Neural Networks for Applied Sciences and Engineering: From Fundamentals to Complex Pattern Recognition, Auerbach Publications; 1st ed., 2006, 570 p.
  13. Hassoun M. H. Fundamentals of Artificial Neural Networks, MIT Press, 1995, 511 p.
  14. Konovalyuk M.A., Gorbunova A.A., KuznetsovYu.V., Baev A.B. Materialy 4 vserossiiskoi konferentsii «Radiolokatsiya i radiosvyaz'», Moscow, 2010, pp. 123–128.

Download

mai.ru — informational site MAI

Copyright © 2000-2021 by MAI

Вход