Multilayer perceptron equalizer for optical communication systems

Abstract

Optical fibers are now widely used in communica tion systems, mainly because they offer faster data transmission speeds, compared to any other type of digital communication sys tem. Despite this great advantage, problems remain that prevent the full exploitation of optical connection: by increasing transmis sion rates over longer distances, the data is affected by nonlinear inter-symbol interference caused by dispersion phenomena in the fiber. Adaptive equalizers can be used to compensate for the effects caused by nonlinear channel responses, restoring the signal originally transmitted. The present work proposes a multilayer perceptron equalizer based on artificial neural networks. The technique was validated using a simulated optical channel, and was compared with other adaptive equalization techniques.

Novel suboptimal approaches for hyperparameter tuning of deep neural network [under the shelf of optical communication]

“…It is used in power restoration systems. Multilayer Perceptron [35] has three or more fully connected layers. It uses a nonlinear activation function (mainly hyperbolic tangent or logistic).…”

Butterfly Neural Filter Applied to Beamforming

“…There is also a new training strategy created for the NB-Butterfly, which is based on the joint error strategy presented in [19], being The results are presented using BER and mean square error (MSE) curves for the backpropagation training. In order to validate the NB-Butterfly and NB-Butterfly-JE, they were analyzed and compared with another neural beamforming strategy called the Bi-dimensional Neural Beamformer with Joint Error (BNB-JE), which is an implementation based on the Multi Layer Perceptron Bi-dimensional Neural Equalizer with Backpropagation (BNE-MLP-BP) presented in [18]- [22] and a least mean square (LMS) adaptive beamformer presented in [23]. The digital communication system where the beamformers were tested uses 4-QAM (quadrature amplitude modulation), which was simulated for four different channels that have varying levels of multi pathing and AOA between the source signal and its reflections, leading to linear and nonlinear configurations.…”

Nonlinear adaptive channel equalization using genetic algorithms

“…Besides the communication channel nature itself, unwanted distortions coming from amplifiers and converters, crosstalk interference and modulation/demodulation process are common causes for nonlinearities. Equalizing communication channels with nonlinear distortions is a well known problem, it has been the subject of many research works in the last decades [1]- [3]. The problem becomes more difficult to treat when the intersymbol interference (ISI) gets severe.…”