Journal of information and communication convergence engineering

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지
DOI QR코드

DOI QR Code

Parameter Estimation of Recurrent Neural Equalizers Using the Derivative-Free Kalman Filter

  • Kwon, Oh-Shin (Department of Control, Robotics and Systems Engineering, Kunsan National University)
  • Received : 2010.05.11
  • Accepted : 2010.05.24
  • Published : 2010.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

For the last decade, recurrent neural networks (RNNs) have been commonly applied to communications channel equalization. The major problems of gradient-based learning techniques, employed to train recurrent neural networks are slow convergence rates and long training sequences. In high-speed communications system, short training symbols and fast convergence speed are essentially required. In this paper, the derivative-free Kalman filter, so called the unscented Kalman filter (UKF), for training a fully connected RNN is presented in a state-space formulation of the system. The main features of the proposed recurrent neural equalizer are fast convergence speed and good performance using relatively short training symbols without the derivative computation. Through experiments of nonlinear channel equalization, the performance of the RNN with a derivative-free Kalman filter is evaluated.

Keywords

Acknowledgement

Supported by : Kunsan National University

References

  1. A. F. Atiya and A. G.Parlos, "New results on recurrent network training: Unifying the algorithms and accelerating convergence," IEEE Transactions on Neural Networks, vol.11, pp.697-709,May 2000. https://doi.org/10.1109/72.846741
  2. G.. Kechriotis, E. Zervas, and E. S. Manolakos, "Using recurrent neural networks for adaptive communication channel equalizations," IEEE Transactions on Neural Networks, vol.5, pp.267-278,March 1994. https://doi.org/10.1109/72.279190
  3. J.S. Choi, M. Bouchard and T. H. Yeap, "Decision Feedback Recurrent Neural Equalization with Fast Convergence Rate," IEEE transactions on Neural Networks, vol. 16, No. 3, May 2005.
  4. R. J. Williams and D. Zipser, "A learning algorithm for continually running fully recurrent neural networks, "Neural Computation, vol. 1, pp. 270-280, 1989. https://doi.org/10.1162/neco.1989.1.2.270
  5. S. Julier, J. Uhlmann, and H. F. Durrant-Whyte, "A new method for the nonlinear transformation of means and covariances in filters and estimators," IEEE Transactions on Automatic Control, vol. 45, pp. 477-482, March 2000. https://doi.org/10.1109/9.847726
  6. S. J. Julier and J. K. Uhlmann, "A new extension of the Kalman filter to nonlinear systems," in Proceedings of Aerospace: The 11th International Symposium on Aerospace/Defence Sensing, Simulation and Controls, 1997.
  7. S. Haykin, Neural Networks: a comprehensive Foundation, 2nd Ed. Upper Saddle River, NJ: Prentice Hall, 1999.
  8. E. A. Wan and R. van der Merwe, "The unscented Kalman filter," in Kalman Filtering and Neural Networks, Edited by S.Haykin. John Wiley and Sons, Inc., 2001.
  9. S. Haykin, Adaptive Filter Theory, 4th Ed. Upper Saddle River, NJ: Prentice Hall, 2002.
  10. C. Cowan and S. Semnani, "Time-variant equalization using a novel non-linear adaptive structure," International Journal of Adaptive Control and Signal Processing, vol. 12, no. 2, pp. 195- 206, 1998. https://doi.org/10.1002/(SICI)1099-1115(199803)12:2<195::AID-ACS487>3.0.CO;2-K
  11. O.S. Kwon, J. S. Choi, M. Bouchard and T.H. Yeap, "A Derivative-Free Kalman Filter for Parameter Estimation of Recurrent Neural Networks and Its Applications to Nonlinear Channel Equalization,"4th international ICSC Symposium on EIS 2004.