Learnable digital signal processing: a new benchmark of linearity compensation for optical fiber communications

Zekun Niu; Hang Yang; Lyu Li; Minghui Shi; Guozhi Xu; Weisheng Hu; Lilin Yi

doi:10.1038/s41377-024-01556-5

Go to Nature Website

Your Location：

Home >

Browse articles >

Learnable digital signal processing: a new benchmark of linearity compensation for optical fiber communications

Original Articles

- Learnable digital signal processing: a new benchmark of linearity compensation for optical fiber communications
- Light: Science & Applications Vol. 13, Issue 9, Pages: 1931-1943(2024)
- Affiliations：
  
  State Key Lab of Advanced Optical Communication Systems and Networks, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, PR China
- Author bio：
  
  Lilin Yi (lilinyi@sjtu.edu.cn)
- Funds：
- DOI：10.1038/s41377-024-01556-5
  CLC：
- Published：30 September 2024，
  
  Published Online：13 August 2024，
  
  Received：02 February 2024，
  
  Revised：13 June 2024，
  
  Accepted：25 July 2024
- Accepted：
Scan QR Code
Niu, Z. K. et al. Learnable digital signal processing: a new benchmark of linearity compensation for optical fiber communications. Light: Science & Applications, 13, 1931-1943 (2024).
DOI：

Niu, Z. K. et al. Learnable digital signal processing: a new benchmark of linearity compensation for optical fiber communications. Light: Science & Applications, 13, 1931-1943 (2024). DOI： 10.1038/s41377-024-01556-5.

摘要

Abstract

The surge in interest regarding the next generation of optical fiber transmission has stimulated the development of digital signal processing (DSP) schemes that are highly cost-effective with both high performance and low complexity. As benchmarks for nonlinear compensation methods

however

traditional DSP designed with block-by-block modules for linear compensations

could exhibit residual linear effects after compensation

limiting the nonlinear compensation performance. Here we propose a high-efficient design thought for DSP based on the learnable perspectivity

called learnable DSP (LDSP). LDSP reuses the traditional DSP modules

regarding the whole DSP as a deep learning framework and optimizing the DSP parameters adaptively based on backpropagation algorithm from a global scale. This method not only establishes new standards in linear DSP performance but also serves as a critical benchmark for nonlinear DSP designs. In comparison to traditional DSP with hyperparameter optimization

a notable enhancement of approximately 1.21 dB in the Q factor for 400 Gb/s signal after 1600 km fiber transmission is experimentally demonstrated by combining LDSP and perturbation-based nonlinear compensation algorithm. Benefiting from the learnable model

LDSP can learn the best configuration adaptively with low complexity

reducing dependence on initial parameters. The proposed approach implements a symbol-rate DSP with a small bit error rate (BER) cost in exchange for a 48% complexity reduction compared to the conventional 2 samples/symbol processing. We believe that LDSP represents a new and highly efficient paradigm for DSP design

which is poised to attract considerable attention across various domains of optical communications.

关键词

Keywords

references

Kim, C. et al. Parity-time symmetry enabled ultra-efficient nonlinear optical signal processing.eLight4, 6 (2024)..

Yushi, C. et al. Additive manufacturing fiber preforms for structured silica fibers with bismuth and erbium dopants.Light Adv. Manuf.3, 13 (2022)..

Savory, S. J. Digital coherent optical receivers: algorithms and subsystems.IEEE J. Sel. Top. Quantum Electron.16, 1164–1179 (2010)..

Zhou, X. Efficient clock and carrier recovery algorithms for single-carrier coherent optical systems: A systematic review on challenges and recent progress.IEEE Signal Process. Mag.31, 35–45 (2014)..

Zhou, X.&Lynn, N. Advanced DSP for 400 Gb/s and beyond optical networks.J. Lightwave Technol.32, 2716–2725 (2014)..

Selmi, M., Yves, J.,&Philippe, C. Accurate digital frequency offset estimator for coherent PolMux QAM transmission systems. 2009 InProc. European Conference on Optical Communications (ECOC)1–3 (IEEE, Vienna, Austria, 2009).

Pfau, T., Sebastian, H.&Reinhold, N. Hardware-efficient coherent digital receiver concept with feedforward carrier recovery for M-QAM constellations.J. Lightwave Technol.27, 989–999 (2009)..

Kikuchi, K. Performance analyses of polarization demultiplexing based on constant-modulus algorithm in digital coherent optical receivers.Opt. Express19, 9868–9880 (2011)..

Rios-Müller, R., Jeremie, R.&Gabriel, C. Blind receiver skew compensation and estimation for long-haul non-dispersion managed systems using adaptive equalizer.J. Lightwave Technol.33, 1315–1318 (2015)..

da, S., Edson, P.&Darko, Z. Widely linear equalization for IQ imbalance and skew compensation in optical coherent receivers.J. Lightwave Technol.34, 3577–3586 (2016)..

Faruk, M. S.&Savory, S. J. Digital signal processing for coherent transceivers employing multilevel formats.J. Lightwave Technol.35, 1125–1141 (2017)..

Song, J. et al. Low-complexity FPGA implementation of 106.24 Gbps DP-QPSK coherent optical receiver with fractional oversampling rate based on one FIR filter for resampling, retiming and equalizing.J. Lightwave Technol.41, 5244–5251 (2023)..

Arikawa, M.&Kazunori, H. Frequency-domain adaptive MIMO filter with fractional oversampling using stochastic gradient descent for long-haul transmission over coupled 4-core fibers.Opt. Express31, 13104–13124 (2023)..

Wang, H. et al. Non-integer-oversampling digital signal processing for coherent passive optical networks.J. Opt. Commun. Netw.16, 4–11 (2024)..

Du, L. B. et al. Digital fiber nonlinearity compensation: toward 1-Tb/s transport.IEEE Signal Process. Mag.31, 46–56 (2014)..

Vassilieva, O., Inwoong, K.&Tadashi, I. Enabling technologies for fiber nonlinearity mitigation in high capacity transmission systems.J. Lightwave Technol.37, 50–60 (2018)..

Zhang, S. et al. Field and lab experimental demonstration of nonlinear impairment compensation using neural networks.Nat. Commun.10, 3033 (2019)..

Redyuk, A. et al. Compensation of nonlinear impairments using inverse perturbation theory with reduced complexity.J. Lightwave Technol.38, 1250–1257 (2020)..

Fu, M. et al. Low-complexity triplet-correlative perturbative fiber nonlinearity compensation for long-haul optical transmission.J. Lightwave Technol.40, 5416–5425 (2022)..

Fan, Q. et al. Advancing theoretical understanding and practical performance of signal processing for nonlinear optical communications through machine learning.Nat. Commun.11, 3694 (2020)..

Häger, C.&Henry, D. P. Physics-based deep learning for fiber-optic communication systems.IEEE J. Sel. Areas Commun.39, 280–294 (2020)..

Fan, Q., Lu, C.&Alan, P. T. L. Combined neural network and adaptive DSP training for long-haul optical communications.J. Lightwave Technol.39, 7083–7091 (2021)..

Xiang, Q. et al. Machine learning assisted modulation-format transparent and nonlinearity tolerant carrier recovery scheme for intelligent receiver.J. Lightwave Technol.38, 6007–6014 (2020)..

Nevin, J. W. et al. Machine learning for optical fiber communication systems: an introduction and overview.APL Photonics6, 121101 (2021)..

Freire, P. et al. Artificial neural networks for photonic applications—from algorithms to implementation: tutorial.Adv. Opt. Photonics15, 739–834 (2023)..

Koike-Akino, T. et al. Neural turbo equalization: Deep learning for fiber-optic nonlinearity compensation.J. Lightwave Technol.38, 3059–3066 (2020)..

Deligiannidis, S. et al. Compensation of fiber nonlinearities in digital coherent systems leveraging long short-term memory neural networks.J. Lightwave Technol.38, 5991–5999 (2020)..

Xiatao, H. et al. Design of fully interpretable neural networks for digital coherent demodulation.Opt. Express30, 35526–35538 (2022)..

Robert M. G. et al, SGD: general analysis and improved rates. InProc International Conference on Machine Learning (ICML)5200–5209 (PMLR, Long Beach, CA, USA, 2019).

Qianqian, T., Guannan, L.&Jinbo, B. Calibrating the adaptive learning rate to improve convergence of ADAM.Neurocomputing481, 333–356 (2022)..

Faruk, M. S.&Kazuro, K. Adaptive frequency-domain equalization in digital coherent optical receivers.Opt. Express19, 12789–12798 (2011)..

Ioffe, S.&Christian S. Batch normalization: accelerating deep network training by reducing internal covariate shift InProc International Conference on Machine Learning (ICML)448–456 (PMLR, Lille, France, 2015).

Paszke, A. et al. Pytorch: an imperative style, high-performance deep learning library. InProc Advances in Neural Information Processing Systems (NeurIPS)8024–8035 (Neural Information Processing Systems Foundation, Vancouver, Canada, 2019).

Liu, L. et al. Initial tap setup of constant modulus algorithm for polarization de-multiplexing in optical coherent receivers.Optical Fiber Communication Conference (OFC)1–3 (IEEE, San Diego, CA, USA, 2009).

Zhang, J., Bo, H.&Xinying, L. Improved quadrature duobinary system performance using multi-modulus equalization.IEEE Photonics Technol. Lett.25, 1630–1633 (2013)..

Haykin, S.Adaptive Filter Theory, 5th edn (Pearson, London, 2014).

Luersen, M. A.&Rodolphe, L. R. Globalized Nelder–Mead method for engineering optimization.Comput. Struct.82, 2251–2260 (2004)..

Falkner, S., Aaron K.,&Frank H. BOHB: Robust and efficient hyperparameter optimization at scale. InProc International Conference on Machine Learning (ICML)1437–1446 (PMLR, Stockholm, Sweden, 2018).

Wu, J. et al. Hyperparameter optimization for machine learning models based on Bayesian optimization.J. Electron. Sci. Technol.17, 26–40 (2019)..

Tao Z. et al. Multiplier-free intrachannel nonlinearity compensating algorithm operating at symbol rate.J. Lightwave Technol.29, 2570–2576 (2011)..

Views

Downloads

CSCD

Alert me when the article has been cited

Submit

Tools

Publicity Resources

No data

Related Author

No data

Related Institution

No data

⁰