Abstract
This chapter applies the natural gradient (NG) descent for adaptive identification of nonlinear channels with memory. The nonlinear channel is comprised of a discrete-time linear filter H followed by a zero-memory nonlinearity g(.). The adaptive system is a neural network which is composed of a linear adaptive filter Q followed by a two-layer memoryless nonlinear neural network (NN). It is shown that the NG learning method significantly outperforms the ordinary gradient descent method in terms of convergence speed and mean squared error (MSE) performance. The chapter studies the sensitivity of the different NN parameters to the natural gradient descent and gives applications to channel tracking and to the modeling of high power amplifiers.
1 In this chapter, the use of the word ‘backpropagation’ (BP) without specifying the gradient descent method, refers to the classical BP algorithm trained with the ordinary gradient descent. (section 2.1).
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Amari S I (1998) Natural Gradient Works Efficiently in Learning. Neural Computation 10, 251–276
Amari S I, Park H, and Fukumizu K (2000) Adaptive method for realizing natu- ral gradient learning for multi-layer perceptrons. Neural Computation 12, 1399–1409
Benedetto S, Biglieri E (1999) Digital Transmission With Wireless Applications, Kluwer Academic Publishers
Bershad N, Celka P, and Vesin J M (1999) Stochastic analysis of gradient adaptive identification of nonlinear systems with memory for gaussian data and noisy input and output measurements. IEEE Trans. Signal Processing 47, 675–689
Bershad N, Celka P, and Vesin J M (2000) Analysis of stochastic gradient tracking of time-varying polynomial Wiener systems. IEEE Trans. Signal Processing 48, 1676–1686.
Ghogho M, Meddeb S, and Bakkoury J (1997) Identification ot time-varying nonlinear channels using polynomial filters. Proc. IEEE Workshop on Non Linear Signal and Image Processing, Michigan, USA
Haykin S (1997) Neural Networks: A Comprehensive Foundation. IEEE Press.
Haykin S (1996) Adaptive Filter Theory. Prentice Hall
Hetrakul P, Taylor D (1976) T hle ettects ot transponder nonimearity on bmary CPSK digital transmission. IEEE Trans. Communications 29, 546–553
Ibnkahla M (2000) Applications of neural networks to digital communications: A survey. Signal Processing 80, 1185–1215
Ibnkahla M, Bershad N J, Sombrin J and Castanié F (1998) Neural network modeling and identification of non linear channels with memory: Algorithms, applications and analytic models. IEEE Trans. Signal Processing 46, 1208–1220
Ibnkahla M (2002) Statistical analysis of neural network modeling and identification of nonlinear channels with memory. IEEE Trans. Signal Processing 50, 1508–1517
Narendra KS, Parthasarathy F (1990) Identification and control of dynamical systems using neural networks. IEEE Trans. Neural Networks 1, 4–27
Prakriya S, Hatzinakos D (1995) Blind identification of LTI-ZMNL-LTI nonlinear channel models. IEEE Trans. Signal Processing 43, 3007–3013
Ralston J, Zoubir A and Bouashash B (1997) Identification of a class of nonlinear systems under stationary non Gaussian Excitation. IEEE Trans. Signal Processing 45, 719–735
Rumelhart D, Hinton G and Williams R (1986) Learning internal representations by error propagation. In: Rumelhart D, McClelland J Eds. Parallel Distributed Processing. MIT Press, pp. 318–362
Saleh A (1981) Frequency-independent and frequency—dependent nonlinear models of TWT amplifiers. IEEE Trans. Communications 29
Sjoberg J et al. (1995) Nonlinear black box modeling in system identification: A unified overview. Automatica 31, 1691–1724
Thomas M, Weidner M, and Durrani S (1974) Digital amplitude-phase keying with Mary alphabets. IEEE Trans. Communications 22, 168–180
Berman A, Mahle C (1970) Non linear phase shift in traveling-wave tubes as applied to multiple access communications satellites. IEEE Trans. Communications 18, 37–48
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Ibnkahla, M. (2004). Nonlinear Channel Identification Using Natural Gradient Descent: Application to Modeling and Tracking. In: Soft Computing in Communications. Studies in Fuzziness and Soft Computing, vol 136. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-45090-0_3
Download citation
DOI: https://doi.org/10.1007/978-3-540-45090-0_3
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-53623-6
Online ISBN: 978-3-540-45090-0
eBook Packages: Springer Book Archive