CMA-based nonlinear blind equaliser modelled by a two-layer feedforward neural network

Combining the ‘mixture of experts’ (ME) model with the expectation-maximisation (EM) algorithm, the author proposes a constant modulus algorithm-based (CMA-based) blind equalisation scheme for equalising the nonlinear channel. The equaliser considered in the paper is assumed as a two-layer feedforward neural network (FNN). The hidden units of the FNN equaliser are approximated with multiple linear systems around multiple points of its input space, and the FNN equaliser is then remodelled with an ME architecture. Based on the ME model, FNN equaliser parameters are estimated with the statistical EM algorithm in a faster convergence speed than the back-propagating algorithm, in which M-operation of the EM algorithm is in fact the CMA-based linear finite impulse response equalisation.

[1]  Jerry M. Mendel,et al.  Tutorial on higher-order statistics (spectra) in signal processing and system theory: theoretical results and some applications , 1991, Proc. IEEE.

[2]  Shun-ichi Amari,et al.  Information geometry of the EM and em algorithms for neural networks , 1995, Neural Networks.

[3]  Jitendra K. Tugnait,et al.  Blind equalization and channel estimation with partial response input signals , 1997, IEEE Trans. Commun..

[4]  Bernard Widrow,et al.  30 years of adaptive neural networks: perceptron, Madaline, and backpropagation , 1990, Proc. IEEE.

[5]  Michael I. Jordan,et al.  Hierarchical Mixtures of Experts and the EM Algorithm , 1994, Neural Computation.

[6]  Michael I. Jordan,et al.  Convergence results for the EM approach to mixtures of experts architectures , 1995, Neural Networks.

[7]  Chrysostomos L. Nikias,et al.  Blind equalization using a tricepstrum-based algorithm , 1991, IEEE Trans. Commun..

[8]  John R. Treichler,et al.  Practical blind demodulators for high-order QAM signals , 1998 .

[9]  Jitendra K. Tugnait,et al.  Identification of linear stochastic systems via second- and fourth-order cumulant matching , 1987, IEEE Trans. Inf. Theory.

[10]  J. Treichler,et al.  A new approach to multipath correction of constant modulus signals , 1983 .

[11]  Jacques Sombrin,et al.  Neural network modeling and identification of nonlinear channels with memory: algorithms, applications, and analytic models , 1998, IEEE Trans. Signal Process..

[12]  J. Cadzow.  Blind deconvolution via cumulant extrema , 1996, IEEE Signal Process. Mag..

[13]  Vojin Zivojnovic,et al.  Minimum Fisher information of moment-constrained distributions with application to robust blind identification , 1998, Signal Process..

[14]  Chuanyi Ji,et al.  An Efficient EM-based Training Algorithm for Feedforward Neural Networks , 1997, Neural Networks.

[15]  Geoffrey E. Hinton,et al.  Adaptive Mixtures of Local Experts , 1991, Neural Computation.