Neuro-Fuzzy Network for Equalization of Different Channel Models

  • Amira A. Elbibas, Issmail M. Ellabib, Yousef Hwegy
  • Published 2013


the rapidly increasing need for information communication requires higher speed and efficient data transmission over communication channels. The rate of data transmissions over these channels is limited due to the effects of inter-symbol interference and additive noise. The conventional techniques used to reduce the effect of channel distortion are based on linear equalizer. However, these techniques are proved to perform poorly owing to nonstationary characteristics of the communication channel. Due to non-stationary characteristics of the channels, this paper addresses equalization problem by adapting the structure of Neuro-Fuzzy network to solve the equalization problem and explores its performance on different channel models. The proposed equalizer is compared with conventional equalizer and neural network equalizer in terms of performance and computation time. Both linear and nonlinear with time-varying and time-invariant channel models are considered with different coefficient. The performance is evaluated in terms of bit-error rate (BER) for different noise powers in the channels. The results obtained demonstrated a significant improvement of the proposed equalizer in terms of both the performance and the computation time compared with LMS and RBF. Keywords—ANFIS, Channel equalizer, digital communication, inter-symbols interference.

11 Figures and Tables

Cite this paper

@inproceedings{Elbibas2013NeuroFuzzyNF, title={Neuro-Fuzzy Network for Equalization of Different Channel Models}, author={Amira A. Elbibas and Issmail M. Ellabib and Yousef Hwegy}, year={2013} }