Abstract:This paper extends previous studies work on the application of Machine Learning (ML) to distortion compensation in optical communication systems with optical fibers nonlinear coefficients (γ) that exceed current industry standards e.g. . To quantify the improvement afforded by ML methods under different transmission conditions the procedures are applied to a model of a typical single-frequency optical communication system with, a 3200 km fiber length, double polarization, and a 16-QAM modulation format. The performance of both transmitters and receivers that incorporate Neural Networks (NNs) are then examined over a wide range of values. In all cases considered, the system Q-factor is improved although the degree of enhancement is dependent on the signal to noise ratio. The ML structures that were investigated include Siamese neural networks (SNN) implemented at the receiver end as well as two-stage architectures that employ NNs at the transmitter together with a classifier at the receiver side. Classifiers ranging from simple decision tree structures to boosting, forests, extra trees, and Multi-layer perceptron (MLP) were further examined and found to provide significant enhancement for . The optimal performance for highly nonlinear systems was achieved for two-stage systems with random forest or extra tree classifiers. Finally, empirical equations for each ML technique were derived that relate the Q-factor enhancement and the value of and the number of triplet terms that are input into the neural network. These results could potentially be employed to relax manufacturing constraints and accordingly reduce system costs.

Artificial Neural Networks for Fiber Nonlinear Noise Estimation

OSNR and nonlinear noise power estimation for optical fiber communication systems using LSTM based deep learning technique

Neural Network Training for OSNR Estimation -- From Prototype to Product

Long Short-Term Memory Neural Network (LSTM-NN) Enabled Accurate Optical Signal-to-Noise Ratio (OSNR) Monitoring

Demonstration of photonic neural network for fiber nonlinearity compensation in long-haul transmission systems

Optical Signal-to-noise Ratio Monitoring Using Uncorrelated Beat Noise

Joint Fiber Nonlinear Noise Estimation, OSNR Estimation and Modulation Format Identification Based on Asynchronous Complex Histograms and Deep Learning for Digital Coherent Receivers

Joint OSNR monitoring and modulation format identification in digital coherent receivers using deep neural networks

Optical Performance Monitoring in Transparent Fiber-Optic Networks Using Neural Networks and Asynchronous Amplitude Histograms

OSNR monitoring for QPSK and 16-QAM systems in presence of fiber nonlinearities for digital coherent receivers

Artificial Neural Network Compensation for an Optical Fiber Temperature Sensing System Based on Optical Time Offset

Machine Learning Model Training Framework for Nonlinear Signal-to-Noise Ratio Estimation in Heterogeneous Optical Networks

Artificial neural networks for nonlinear pulse shaping in optical fibers

Machine learning compensation of fiber nonlinear noise

Neural Network-based Fiber Nonlinearity Mitigation in High-speed Coherent Optical Transmission Systems

Fiber Link Anomaly Detection and Estimation Based on Signal Nonlinearity

Going Deeper into OSNR Estimation with CNN

Intelligent optical performance monitor using multi-task learning based artificial neural network

Artificial intelligence based optical performance monitoring

Nonlinear Recovering Technique Based on Neural Network for Frequency Modulated Continuous Wave Optical Fiber Sensors

A Novel Method to Monitor OSNR Using a Mach-Zehnder Interferometer