Autoencoder-Optimized Geometric Constellation Shaping for Unamplified Coherent Optical Links

Oliveira, Beatriz M.; Neves, Manuel; Guiomar, Fernando P.; Medeiros, Maria C. R.; Monteiro, Paulo

doi:10.1364/cleo_si.2022.sw4e.7

Cited by 4 publications

(4 citation statements)

References 4 publications

(8 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…41,42 The joint optimization can focus on several aspects of the signaling, e.g. constellation shaping and decision regions, [43][44][45][46][47][48][49][50] pulse shaping and receiver filtering, 41,42 etc. Similarly, the optimization objective can be varied from absolute system throughput [43][44][45][46][47] to robustness to uncertainty in channel 41,48 or transciever 50,51 parameters.…”

Section: End-to-end Learning Of Phase-noise Robust Communicationmentioning

confidence: 99%

“…constellation shaping and decision regions, [43][44][45][46][47][48][49][50] pulse shaping and receiver filtering, 41,42 etc. Similarly, the optimization objective can be varied from absolute system throughput [43][44][45][46][47] to robustness to uncertainty in channel 41,48 or transciever 50,51 parameters. The challenge of applying end-toend learning mainly consists of requiring a differentiable transmitter-to-receiver model in order to optimize the encoder-decoder pair through gradient-based methods.…”

Section: End-to-end Learning Of Phase-noise Robust Communicationmentioning

confidence: 99%

See 1 more Smart Citation

Machine learning meets photonics

Da Ros,

Yankov,

Zibar

2023

Emerging Topics in Artificial Intelligence (ETAI) 2023

View full text Add to dashboard Cite

Machine learning (ML) is becoming a ubiquitous and powerful tool helping to address challenges in countless fields. Applications of ML addressing optics challenges have been extensively studied in recent years opening up new research directions. In particular, here, we review some of our current efforts and provide examples of successful applications of ML to the characterization of photonic devices, design, and modeling of optical subsystems, and complete end-to-end optical system optimization. ML and statistical tools can yield additional insight from measurement data, e.g. by targeted filtering of noise sources. They have also been shown to assist complex or inaccurate physics-based models through black and grey-box modeling of photonics components or subsystems. Such ML-aided models have enabled easier optimization and design (including inverse design) of optical systems.

show abstract

Section: End-to-end Learning Of Phase-noise Robust Communicationmentioning

confidence: 99%

Section: End-to-end Learning Of Phase-noise Robust Communicationmentioning

confidence: 99%

Machine learning meets photonics

Da Ros,

Yankov,

Zibar

2023

Emerging Topics in Artificial Intelligence (ETAI) 2023

View full text Add to dashboard Cite

show abstract

“…G EOMETRIC constellation shaping has become attractive in communication systems due to its flexible design for different impairments [1], e.g., tolerance to laser phase noise [2], [3], fiber Kerr nonlinearities [4]- [6], transceiver components impairments [7], and low signal-to-noise ratio (SNR) channels [8]. In contrast to probabilistic constellation shaping (PCS) [9], [10], geometric constellation shaping (GCS) modifies the Euclidean geometry of the constellation points while keeping their occurrence probability uniform [11], [12,Ch.…”

Section: Introductionmentioning

confidence: 99%

Physical Realizations of Multidimensional Voronoi Constellations in Optical Communication Systems

Mirani

Vijayan

et al. 2023

J. Lightwave Technol.

View full text Add to dashboard Cite

Multidimensional geometric shaping has been shown to outperform uniform quadrature amplitude modulation (QAM) in optical communication systems but the complexity of symbol decision and bit mapping can often be significant as dimensionality increases. In this paper, a low-complexity geometric shaping method based on multidimensional lattices is investigated both in experiments and simulations. The modulation formats designed based on this method are called Voronoi constellations (VCs) and we study them in 8, 16, and 32 dimensions. We obtain transmission reach improvements of up to 22 and 70% for VCs compared to 4QAM and 16QAM, respectively, in nonlinear long-haul fiber transmission. Moreover, we compare different physical realizations of multidimensional VCs over wavelengths, polarizations, and time slots in both the Gaussian and nonlinear fiber channels. We demonstrate that different physical realizations perform similarly in the fiber-optic back-to-back channel. However, in long-haul transmission systems, spreading the dimensions over time slots can increase the transmission reach up to 4% compared to wavelengths and polarizations. Furthermore, the mutual information and generalized mutual information are estimated and compared to QAM formats at the same spectral efficiencies.

show abstract

“…End-to-end learning, which was introduced in [8], utilizes an autoencoder (AE) [9] to optimize the communication system and it has gathered traction as a method to approach GCS due to its versatility to the employed channel model. Endto-end learning has proven to be effective for GCS in optical communication systems [6], [10]- [15], mainly focusing on the mitigation of the nonlinear effects of the optical fiber. Geometric constellation shaping considering the BPS algorithm was explored in [16], [17], where a constellation was optimized to be robust to channel uncertainties with BPS at the receiver.…”

Section: Introductionmentioning

confidence: 99%

Geometric Constellation Shaping for Fiber-Optic Channels via End-to-End Learning

Jovanovic

Ros

Zibar

et al. 2023

J. Lightwave Technol.

View full text Add to dashboard Cite

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. Users may download and print one copy of any publication from the public portal for the purpose of private study or research.  You may not further distribute the material or use it for any profit-making activity or commercial gain  You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

show abstract

Autoencoder-Optimized Geometric Constellation Shaping for Unamplified Coherent Optical Links

Abstract: Using end-to-end deep learning, we experimentally demonstrate the optimized design of geometric constellation shaping for coherent unamplified links. A power budget gain of more than 2 dB is demonstrated for 8-ary constellations.

Cited by 4 publications

References 4 publications

Machine learning meets photonics

Machine learning meets photonics

Physical Realizations of Multidimensional Voronoi Constellations in Optical Communication Systems

Geometric Constellation Shaping for Fiber-Optic Channels via End-to-End Learning

Contact Info

Product

Resources

About