Machine learning techniques in optical communication

150

Linear signal processing algorithms are effective in dealing with linear transmission channel and linear signal detection, while the nonlinear signal processing algorithms, from the machine learning community, are effective in dealing with nonlinear transmission channel and nonlinear signal detection. In this paper, a brief overview of the various machine learning methods and their application in optical communication is presented and discussed. Moreover, supervised machine learning methods, such as neural networks and support vector machine, are experimentally demonstrated for in-band optical signal to noise ratio (OSNR) estimation and modulation format classification, respectively. The proposed methods accurately evaluate optical signals employing up to 64 quadrature amplitude modulation (QAM), at 32 Gbaud, using only directly-detected data.

Section: Clustering K-meansmentioning

confidence: 99%

Machine Learning Techniques for Optical Performance Monitoring From Directly Detected PDM-QAM Signals

Thrane

Wass

Piels

et al. 2017

150

“…Some popular implementations are based on Artificial Neural Networks (ANNs) ( [7]) and Support Vector Machines (SVMs) ( [8], [9]). In [10] techniques based on nonlinear state-space based Bayesian filtering and Gaussian Mixture Models (GMMs) are presented. A more recent study of a photonic machine learning implementation for signal recovery in optical communications can be found in [11].…”

Section: Introductionmentioning

confidence: 99%

A Neuromorphic Silicon Photonics Nonlinear Equalizer For Optical Communications With Intensity Modulation and Direct Detection

Katumba

Yin

Dambre

et al. 2019

We present the design and numerical study of a nonlinear equalizer for optical communications based on silicon photonics and reservoir computing. The proposed equalizer leverages the optical information processing capabilities of integrated photonic reservoirs to combat distortions both in metro links of a few hundred kilometers and in high-speed short-reach intensitymodulation-direct-detection links. We show nonlinear compensation in unrepeated metro links of up to 200 km that outperform electrical feedforward equalizers based equalizers, and ultimately any linear compensation device. For a high-speed short-reach 40-Gb/s link based on a distributed feedback laser and an electroabsorptive modulator, and considering a hard decision forward error correction limit of 0.2 × 10 −2 , we can increase the reach by almost 10 km. Our equalizer is compact (only 16 nodes) and operates in the optical domain without the need for complex electronic DSP, meaning its performance is not bandwidth constrained. The approach is, therefore, a viable candidate even for equalization techniques far beyond 100G optical communication links.

“…Moreover, all proposed nonlinearity compensators present high complexity [3][4][5][6][7] being impractical for real-time communications. The aforementioned random noises of the network can be partially tackled by low-complex digital machine learning algorithms that perform nonlinear equalization (NLE), such as unsupervised and supervised algorithms: machine learning clustering (MLC) with K-means and Gaussian mixture [8][9][10], and classification machines [11], e.g. artificial neural networks (ANN) [12][13][14] and convolutional neural network-based deep learning [15,16].…”

mentioning

confidence: 99%

Blind Nonlinearity Equalization by Machine-Learning-Based Clustering for Single- and Multichannel Coherent Optical OFDM

Giacoumidis

Matin

Wei

et al. 2018

Abstract-Fiber-induced intra-and inter-channel nonlinearities are experimentally tackled using blind nonlinear equalization (NLE) by unsupervised machine learning based clustering (MLC) in ~46 Gb/s single-channel and ~20 Gb/s (middle-channel) multi-channel coherent multi-carrier signals (OFDM-based). To that end we introduce, for the first time, Hierarchical and FuzzyLogic C-means (FLC) based clustering in optical communications. It is shown that among the two proposed MLC algorithms, FLC reveals the highest performance at optimum launched optical powers (LOPs), while at very high LOPs Hierarchical can compensate more effectively nonlinearities only for low-level modulation formats. When employing BPSK and QPSK, FLC outperforms K-means, Fast-Newton support vector machines, supervised artificial neural networks and NLE with deterministic Volterra analysis. In particular, for the middle channel of a QPSK WDM coherent optical OFDM system at optimum -5 dBm of LOP and 3200 km of transmission, FLC outperforms Volterra-NLE by 2.5 dB in Q-factor. However, for a 16-QAM single-channel system at 2000 km, the performance benefit of FLC over IVSTF reduces to ~0.4 dB at a LOP of 2 dBm (optimum). Even when using novel sophisticated clustering designs in 16 clusters, no more than additional ~0.3 dB Q-factor enhancement is observed. Finally, in contrast to the deterministic Volterra-NLE, MLC algorithms can partially tackle the stochastic parametric noise amplification.