Adaptive filtering with the self-organizing map: a performance comparison
Neural Networks - 2006 Special issue: Advances in self-organizing maps--WSOM'05
Information Sciences: an International Journal
Performance analysis of a RLS-based MLP-DFE in time-invariant and time-varying channels
Digital Signal Processing
Channel equalization using neural networks: a review
IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
Channel equalization using simplified least mean-fourth algorithm
Digital Signal Processing
Information Sciences: an International Journal
An adaptive decision feedback equalizer based on the combination of the FIR and FLNN
Digital Signal Processing
| Hi-index | 0.00 |
The severely distorting channels limit the use of linear equalizers and the use of the nonlinear equalizers then becomes justifiable. Neural-network-based equalizers, especially the multilayer perceptron (MLP)-based equalizers, are computationally efficient alternative to currently used nonlinear filter realizations, e.g., the Volterra type. The drawback of the MLP-based equalizers is, however, their slow rate of convergence, which limit their use in practical systems. In this work, the effect of whitening the input data in a multilayer perceptron-based decision feedback equalizer (DFE) is evaluated. It is shown from computer simulations that whitening the received data employing adaptive lattice channel equalization algorithms improves the convergence rate and bit error rate performances of multilayer perceptron-based DFE. The adaptive lattice algorithm is a modification to the one developed by Ling and Proakis (1985). The consistency in performance is observed in both time-invariant and time-varying channels. Finally, it is found in this work that, for time-invariant channels, the MLP DFE outperforms the least mean squares (LMS)-based DFE. However, for time-varying channels comparable performance is obtained for the two configurations