Joint Superchannel Digital Signal Processing for Ultimate Bandwidth Utilization

Mazur, Mikael; Schröder, Jochen; Karlsson, Magnus; Andrekson, Peter A.

doi:10.48550/arxiv.1911.02326

Cited by 2 publications

(5 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to (5), the dependence of L Bt (θ 1 , θ 2 , θ 3 , θ 4 ) on θ 1 , θ 2 , θ 3 is implicit through the distribution of the downsampled signal y k , which is a function of the channel input g(s k ), where g(•) denotes the joint transfer function of the DAC, PA, Update NN2 parameters according to (17) 16: until Stop criterion is satisfied and IQM, and sk is dependent on NNs 1-3 as can be seen in Fig. 3.…”

Section: B Learning With a Channel Modelmentioning

confidence: 96%

“…In this paper, for the sake of simplicity, the ENOB is assumed to be constant over the considered bandwidth and is set to 6. 5 We model the ENOB noise introduced by the DAC as AWGN with variance determined by the ENOB of the device [47]…”

Section: Wdm System and Main Hardware Limitationsmentioning

confidence: 99%

“…To minimize the guard bands between channels, it is common to employ pulse shaping to create a near-rectangular spectrum in the frequency domain with a bandwidth close to the symbol rate. However, in practice, generating a rectangular spectrum is difficult due to the finite pulse-shaping (PS) filter and transceiver hardware impairments, requiring computation expensive digital signal processing to eliminate performance degradation caused by inter-channel interference (ICI) [5]. Guard bands therefore remain a major contributor to SE loss in WDM systems.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Model-Based End-to-End Learning for WDM Systems With Transceiver Hardware Impairments

Song

Häger

Schröder

et al. 2022

IEEE J. Select. Topics Quantum Electron.

View full text Add to dashboard Cite

We propose an autoencoder (AE)-based transceiver for a wavelength division multiplexing (WDM) system impaired by hardware imperfections. We design our AE following the architecture of conventional communication systems. This enables to initialize the AE-based transceiver to have similar performance to its conventional counterpart prior to training and improves the training convergence rate. We first train the AE in a singlechannel system, and show that it achieves performance improvements by putting energy outside the desired bandwidth, and therefore cannot be used for a WDM system. We then train the AE in a WDM setup. Simulation results show that the proposed AE significantly outperforms the conventional approach. More specifically, it increases the spectral efficiency of the considered system by reducing the guard band by 37% and 50% for a rootraised-cosine filter-based matched filter with 10% and 1% roll-off, respectively. An ablation study indicates that the performance gain can be ascribed to the optimization of the symbol mapper, the pulse-shaping filter, and the symbol demapper. Finally, we use reinforcement learning to learn the pulse-shaping filter assuming that the channel model is unknown. Simulation results show that the reinforcement-learning-based algorithm achieves similar performance as the standard supervised end-to-end learning approach assuming perfect channel knowledge.

show abstract

Section: B Learning With a Channel Modelmentioning

confidence: 96%

Section: Wdm System and Main Hardware Limitationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Model-Based End-to-End Learning for WDM Systems With Transceiver Hardware Impairments

Song

Häger

Schröder

et al. 2022

IEEE J. Select. Topics Quantum Electron.

View full text Add to dashboard Cite

show abstract

“…Similar to (5), the dependence of L Bt (θ 1 , θ 2 , θ 3 , θ 4 ) on θ 1 , θ 2 , θ 3 is implicit through the distribution of the downsampled signal y k , which is a function of the channel input g(s k ), where g(•) denotes the joint transfer function of the DAC, PA, and IQM, and sk is dependent on NNs 1-3 as can be seen in Fig. 3.…”

Section: B Learning With a Channel Modelmentioning

confidence: 96%

“…As a result, the proposed scheme has decreased training complexity and increased interpretability as compared to a conventional AE. 5 1) Transmitter: At the transmitter, the symbol mapper, the PS filter, and the DPD of the conventional communication system are replaced by three NNs. We denote these three NNs by f θ1 (•), f θ2 (•), and f θ3 (•), where θ 1 , θ 2 , and θ 3 are the sets of trainable parameters.…”

Section: A Autoencoder Designmentioning

confidence: 99%

Model-Based End-to-End Learning for WDM Systems With Transceiver Hardware Impairments

Song¹,

Häger²,

Schröder³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

We propose an autoencoder (AE)-based transceiver for a wavelength division multiplexing (WDM) system impaired by hardware imperfections. We design our AE following the architecture of conventional communication systems. This enables to initialize the AE-based transceiver to have similar performance to its conventional counterpart prior to training and improves the training convergence rate. We first train the AE in a singlechannel system, and show that it achieves performance improvements by putting energy outside the desired bandwidth, and therefore cannot be used for a WDM system. We then train the AE in a WDM setup. Simulation results show that the proposed AE significantly outperforms the conventional approach. More specifically, it increases the spectral efficiency of the considered system by reducing the guard band by 37% and 50% for a rootraised-cosine filter-based matched filter with 10% and 1% roll-off, respectively. An ablation study indicates that the performance gain can be ascribed to the optimization of the symbol mapper, the pulse-shaping filter, and the symbol demapper. Finally, we use reinforcement learning to learn the pulse-shaping filter under the assumption that the channel model is unknown. Simulation results show that the reinforcement-learning-based algorithm achieves similar performance to the standard supervised endto-end learning approach assuming perfect channel knowledge.

show abstract

Joint Superchannel Digital Signal Processing for Ultimate Bandwidth Utilization

Cited by 2 publications

References 48 publications

Model-Based End-to-End Learning for WDM Systems With Transceiver Hardware Impairments

Model-Based End-to-End Learning for WDM Systems With Transceiver Hardware Impairments

Model-Based End-to-End Learning for WDM Systems With Transceiver Hardware Impairments

Contact Info

Product

Resources

About