Joint Neural Network Equalizer and Decoder

TL;DR

This paper introduces a joint neural network-based equalizer and decoder for nonlinear channels, achieving improved performance and reduced complexity without requiring channel state information.

Contribution

The paper proposes a novel two-network approach for blind equalization and decoding, outperforming existing machine learning methods and reducing parameter count.

Findings

CNN equalizer outperforms other ML-based solutions

Model reduces parameters by about two-thirds

Efficiently handles long sequences with linear complexity

Abstract

Recently, deep learning methods have shown significant improvements in communication systems. In this paper, we study the equalization problem over the nonlinear channel using neural networks. The joint equalizer and decoder based on neural networks are proposed to realize blind equalization and decoding process without the knowledge of channel state information (CSI). Different from previous methods, we use two neural networks instead of one. First, convolutional neural network (CNN) is used to adaptively recover the transmitted signal from channel impairment and nonlinear distortions. Then the deep neural network decoder (NND) decodes the detected signal from CNN equalizer. Under various channel conditions, the experiment results demonstrate that the proposed CNN equalizer achieves better performance than other solutions based on machine learning methods. The proposed model reduces about $2/3$ of the parameters compared to state-of-the-art counterparts. Besides, our model can be easily applied to long sequence with $\mathcal{O}(n)$ complexity.