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

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.

[1]  Timothy J. O'Shea,et al.  Physical layer deep learning of encodings for the MIMO fading channel , 2017, 2017 55th Annual Allerton Conference on Communication, Control, and Computing (Allerton).

[2]  Henk Wymeersch,et al.  End-to-end Autoencoder for Superchannel Transceivers with Hardware Impairment , 2021, 2021 Optical Fiber Communications Conference and Exhibition (OFC).

[3]  Darko Zibar,et al.  Machine Learning Techniques in Optical Communication , 2015, Journal of Lightwave Technology.

[4]  Timothy J. O'Shea,et al.  Deep architectures for modulation recognition , 2017, 2017 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN).

[5]  Woongsup Lee,et al.  A Novel PAPR Reduction Scheme for OFDM System Based on Deep Learning , 2018, IEEE Communications Letters.

[6]  Henk Wymeersch,et al.  Achievable Information Rates for Nonlinear Fiber Communication via End-to-end Autoencoder Learning , 2018, 2018 European Conference on Optical Communication (ECOC).

[7]  Henry D. Pfister,et al.  Nonlinear Interference Mitigation via Deep Neural Networks , 2017, 2018 Optical Fiber Communications Conference and Exposition (OFC).

[8]  Maxim Kuschnerov,et al.  Flex-grid optical networks: spectrum allocation and nonlinear dynamics of super-channels. , 2013, Optics express.

[9]  Talha Rahman,et al.  Nonlinear digital pre-distortion of transmitter components , 2015, 2015 European Conference on Optical Communication (ECOC).

[10]  Jakob Hoydis,et al.  Waveform Learning for Next-Generation Wireless Communication Systems , 2021, ArXiv.

[11]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[12]  Mingliang Li,et al.  End-to-end Learning for Optical Fiber Communication with Data-driven Channel Model , 2020, 2020 Opto-Electronics and Communications Conference (OECC).

[13]  Ginni Khanna,et al.  A Robust Adaptive Pre-Distortion Method for Optical Communication Transmitters , 2016, IEEE Photonics Technology Letters.

[14]  Benjamin J. Puttnam,et al.  Geometric Constellation Shaping for Fiber Optic Communication Systems via End-to-end Learning , 2018, ArXiv.

[15]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[16]  Y. Nishida,et al.  Gain-flattened tellurite-based EDFA with a flat amplification bandwidth of 76 nm , 1998, IEEE Photonics Technology Letters.

[17]  Biing-Hwang Juang,et al.  Channel Agnostic End-to-End Learning Based Communication Systems with Conditional GAN , 2018, 2018 IEEE Globecom Workshops (GC Wkshps).

[18]  Stephan ten Brink,et al.  Deep-learning Autoencoder for Coherent and Nonlinear Optical Communication , 2020, ArXiv.

[19]  G. Liga,et al.  End-to-End Learning in Optical Fiber Communications: Experimental Demonstration and Future Trends , 2020, 2020 European Conference on Optical Communications (ECOC).

[20]  Mikael Mazur,et al.  Joint Superchannel Digital Signal Processing for Ultimate Bandwidth Utilization. , 2019, 1911.02326.

[21]  Xue Chen,et al.  Data-driven Optical Fiber Channel Modeling: A Deep Learning Approach , 2020, Journal of Lightwave Technology.

[22]  Jiri Matas,et al.  All you need is a good init , 2015, ICLR.

[23]  Jakob Hoydis,et al.  Trainable Communication Systems: Concepts and Prototype , 2020, IEEE Transactions on Communications.

[24]  Jakob Hoydis,et al.  End-to-End Learning of Communications Systems Without a Channel Model , 2018, 2018 52nd Asilomar Conference on Signals, Systems, and Computers.

[25]  Laurent Schmalen,et al.  Concept and Experimental Demonstration of Optical IM/DD End-to-End System Optimization using a Generative Model , 2019, 2020 Optical Fiber Communications Conference and Exhibition (OFC).

[26]  Stefano Ermon,et al.  Neural Joint Source-Channel Coding , 2018, ICML.

[27]  Maxim Kuschnerov,et al.  Digital Compensation of Bandwidth Limitations for High-Speed DACs and ADCs , 2016, Journal of Lightwave Technology.

[28]  A. Carena,et al.  Optimization of DSP-based Nyquist-WDM PM-16QAM transmitter , 2012, 2012 38th European Conference and Exhibition on Optical Communications.

[29]  Polina Bayvel,et al.  End-to-End Deep Learning of Optical Fiber Communications , 2018, Journal of Lightwave Technology.

[30]  Francesco Da Ros,et al.  End-to-end Learning of a Constellation Shape Robust to Variations in SNR and Laser Linewidth , 2021, 2021 European Conference on Optical Communication (ECOC).

[31]  Stephan ten Brink,et al.  Deep Learning Based Communication Over the Air , 2017, IEEE Journal of Selected Topics in Signal Processing.

[32]  Matthew D. Zeiler ADADELTA: An Adaptive Learning Rate Method , 2012, ArXiv.

[33]  Maurice O'Sullivan,et al.  Advances in High-Speed DACs, ADCs, and DSP for Optical Coherent Transceivers , 2014, Journal of Lightwave Technology.

[34]  Keang-Po Ho,et al.  Coherent Optical OFDM Transmitter Design Employing Predistortion , 2008, IEEE Photonics Technology Letters.

[35]  Ming Liu,et al.  Neural Network Assisted Active Constellation Extension for PAPR Reduction of OFDM System , 2019, 2019 11th International Conference on Wireless Communications and Signal Processing (WCSP).

[36]  Metodi P. Yankov,et al.  End-to-end Learning for GMI Optimized Geometric Constellation Shape , 2019, 45th European Conference on Optical Communication (ECOC 2019).

[37]  Bin Chen,et al.  End-to-End Learning of Geometrical Shaping Maximizing Generalized Mutual Information , 2019, 2020 Optical Fiber Communications Conference and Exhibition (OFC).

[38]  Peter J Winzer,et al.  Fiber-optic transmission and networking: the previous 20 and the next 20 years [Invited]. , 2018, Optics express.

[39]  Stephan ten Brink,et al.  OFDM-Autoencoder for End-to-End Learning of Communications Systems , 2018, 2018 IEEE 19th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC).

[40]  Chao Lu,et al.  Machine Learning Methods for Optical Communication Systems , 2017 .

[41]  Polina Bayvel,et al.  End-to-End Optimized Transmission over Dispersive Intensity-Modulated Channels Using Bidirectional Recurrent Neural Networks , 2019, Optics express.

[42]  Jakob Hoydis,et al.  An Introduction to Deep Learning for the Physical Layer , 2017, IEEE Transactions on Cognitive Communications and Networking.

[43]  Ginni Khanna,et al.  Digital pre-compensation techniques enabling high-capacity bandwidth variable transponders , 2018 .

[44]  Henk Wymeersch,et al.  Over-the-fiber Digital Predistortion Using Reinforcement Learning , 2021, 2021 European Conference on Optical Communication (ECOC).

[45]  Michail Matthaiou,et al.  Deep Learning-Aided Multicarrier Systems , 2020, IEEE Transactions on Wireless Communications.

[46]  Jakob Hoydis,et al.  Joint Learning of Geometric and Probabilistic Constellation Shaping , 2019, 2019 IEEE Globecom Workshops (GC Wkshps).

[47]  Nathan Srebro,et al.  The Marginal Value of Adaptive Gradient Methods in Machine Learning , 2017, NIPS.

[48]  Jakob Hoydis,et al.  Model-Free Training of End-to-End Communication Systems , 2018, IEEE Journal on Selected Areas in Communications.

[49]  Sergio Escalera,et al.  Beyond One-hot Encoding: lower dimensional target embedding , 2018, Image Vis. Comput..

[50]  Zhu Han,et al.  Machine Learning Paradigms for Next-Generation Wireless Networks , 2017, IEEE Wireless Communications.