Probabilistic shaping communication system aided by neural network distribution matcher in data center optical network

Abstract

A neural network (NN)‐assisted probabilistic shaping (PS) distribution matcher is proposed, in which the model is simplified by a structured optimization method. The NN algorithm can encode the information sequence, making the signal obey the Gaussian distribution, and can directly restore the received signal. In addition, the algorithm uses the novel training method at both ends of the transmitter and receiver so that the system performance is significantly improved. PS system verification experiments have been carried out under 16QAM‐DMT modulation format. Under the hard decision forward error correction (FEC) threshold of 3.8*10−3 BER, the proposed system achieves 1.1 dB improvement compared to the traditional 16QAM‐DMT system.

End-to-End Deep Learning of Joint Geometric Probabilistic Shaping Using a Channel-Sensitive Autoencoder

“…In recent years, deep learning neural networks (DLNNs) have been widely applied to enhance the transmission performance achieved in various aspects of optical fiber communication systems [4,5], which mainly consist of three components: a transmitter, channels and a receiver. At the transmitter, neural networks (NNs) can be used for constellation shaping (CS), which includes geometric shaping (GS) [6] and probabilistic shaping (PS) [7], to promote transmission flexibility. In terms of transmission channels, NNs exhibit strong channel fitting abilities, providing rapid channel simulation [8][9][10][11].…”