Deep Learning-Aided Perturbation Model-Based Fiber Nonlinearity Compensation

Luo, Shenghang; Kumar, O. S. Sunish; Lampe, Lutz; Mitra, J.

doi:10.1109/jlt.2023.3279449

Cited by 3 publications

(1 citation statement)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data-driven ML models have arisen as potent tools for addressing numerous intricate challenges in optical communications and are being envisaged as an enabler for future efficient, intelligent and reliable network infrastructures 1 , 2 . In the last few years, industry as well as academia has witnessed a significant increase in research endeavors to harness and capitalize on ML across different facets of fiber-optic communications ranging from designing of network components 3 – 7 to compensating critical transmission impairments 8 – 10 to predicting data traffic flow patterns in networks 11 , 12 . However, despite unprecedented interest in this field over the past decade, the developed ML methods have unfortunately not yet attained anticipated deployment, credibility and impact in real-world fiber-optic networks, barring the successful implementation of some non-network related business use-cases (e.g., customers’ churn prediction, customers’ segmentation, etc.)…”

Section: Introductionmentioning

confidence: 99%

Non-technological barriers: the last frontier towards AI-powered intelligent optical networks

Khan

2024

Nat Commun

View full text Add to dashboard Cite

Machine learning (ML) has been remarkably successful in transforming numerous scientific and technological fields in recent years including computer vision, natural language processing, speech recognition, bioinformatics, etc. Naturally, it has long been considered as a promising mechanism to fundamentally revolutionize the existing archaic optical networks into next-generation smart and autonomous entities. However, despite its promise and extensive research conducted over the last decade, the ML paradigm has so far not been triumphant in achieving widespread adoption in commercial optical networks. In our perspective, this is primarily due to non-addressal of a number of critical non-technological issues surrounding ML-based solutions’ development and use in real-world optical networks. The vision of intelligent and autonomous fiber-optic networks, powered by ML, will always remain a distant dream until these so far neglected factors are openly confronted by all relevant stakeholders and categorically resolved.

show abstract

Section: Introductionmentioning

confidence: 99%

Non-technological barriers: the last frontier towards AI-powered intelligent optical networks

Khan

2024

Nat Commun

View full text Add to dashboard Cite

show abstract

Transfer learning-assisted modulation classification and OSNR estimation using constellation diagrams

Jha,

Mishra

2024

J. Opt.

View full text Add to dashboard Cite

A cost-effective deep transfer learning (TL) learned with characteristics derived from the constellation diagram is proposed for executing modulation classification and optical signal-to-noise ratio (OSNR) estimation for the future generation of elastic optical networks. The information acquired with the ImageNet dataset is transferred using pre-trained TL versions including VGG19, ResNet152V2, InceptionV3, and ResNet50 to identify various modulation formats and their corresponding OSNR. Adam optimizer is employed to ascertain the appropriate settings for the hyperparameters. Numerical results demonstrate that the proposed VGG19 method achieves the highest level of accuracy (100%) in the recognition of various modulation forms. To fulfill the requirements of actual use, OSNR monitoring is also explored, with a total precision equal to 95.2%. Furthermore, a thorough investigation of the impact of training data dimensions, on TL performance is conducted. The outcomes of the proposed method demonstrate that the suggested TL-based algorithms are more accurate and need much less training and testing time than non-TL approaches.

show abstract

Performance analysis of nonlinear crosstalk of WDM systems using modulation schemes criteria

Mohammed,

Mansoor,

Abd

et al. 2024

Open Engineering

View full text Add to dashboard Cite

Nonlinearities in optical fibers are regarded as the most significant barriers that endanger the effectiveness of the optical transmission system and pose a threat to communication quality. Four-wave mixing (FWM) is one of the most important nonlinear effects that greatly reduces the wavelength-division multiplexing (WDM) system performance at high data rates over extended transmission distances. This research examines, and assesses, numerically, the behavior of a 4-channel, 40 Gbps WDM system under the effect of the FWM under various tuning parameters, including dispersion, input power, and wavelength spacing. The system model was built using OptiSystem software, and then three different modulation formats, namely, Non-return-to-zero-frequency shift keying, Return-to-zero frequency shift keying, and differential phase shift keying (DPSK) are used to assess the FWM power penalty. The results demonstrate that the FWM power penalty obtained with 1 nm wavelength separation in the DPSK method is dramatically reduced to −35 dBm. This study also demonstrates that when power variation is taken into consideration, the DPSK modulation scheme delivers a lower bit error rate in comparison to other modulation schemes.

show abstract

Deep Learning-Aided Perturbation Model-Based Fiber Nonlinearity Compensation

Cited by 3 publications

References 36 publications

Non-technological barriers: the last frontier towards AI-powered intelligent optical networks

Non-technological barriers: the last frontier towards AI-powered intelligent optical networks

Transfer learning-assisted modulation classification and OSNR estimation using constellation diagrams

Performance analysis of nonlinear crosstalk of WDM systems using modulation schemes criteria

Contact Info

Product

Resources

About