-
End-to-End Learning of Pulse-Shaper and Receiver Filter in the Presence of Strong Intersymbol Interference
Authors:
Søren Føns Nielsen,
Francesco Da Ros,
Mikkel N. Schmidt,
Darko Zibar
Abstract:
We numerically demonstrate that joint optimization of FIR based pulse-shaper and receiver filter results in an improved system performance, and shorter filter lengths (lower complexity), for 4-PAM 100 GBd IM/DD systems.
We numerically demonstrate that joint optimization of FIR based pulse-shaper and receiver filter results in an improved system performance, and shorter filter lengths (lower complexity), for 4-PAM 100 GBd IM/DD systems.
△ Less
Submitted 22 May, 2024;
originally announced May 2024.
-
End-to-End Optimization of Directly Modulated Laser Links using Chirp-Aware Modeling
Authors:
Sergio Hernandez F.,
Christophe Peucheret,
Francesco Da Ros,
Darko Zibar
Abstract:
The rate and reach of directly-modulated laser links is often limited by the interplay between chirp and fiber chromatic dispersion. We address this by optimizing the transmitter, receiver, bias and peak-to-peak current to the laser jointly. Our approach outperforms Volterra post-equalization at various symbol rates.
The rate and reach of directly-modulated laser links is often limited by the interplay between chirp and fiber chromatic dispersion. We address this by optimizing the transmitter, receiver, bias and peak-to-peak current to the laser jointly. Our approach outperforms Volterra post-equalization at various symbol rates.
△ Less
Submitted 16 May, 2024;
originally announced May 2024.
-
End-to-end Optimization of Optical Communication Systems based on Directly Modulated Lasers
Authors:
Sergio Hernandez F.,
Christophe Peucheret,
Francesco Da Ros,
Darko Zibar
Abstract:
The use of directly modulated lasers (DMLs) is attractive in low-power, cost-constrained short-reach optical links. However, their limited modulation bandwidth can induce waveform distortion, undermining their data throughput. Traditional distortion mitigation techniques have relied mainly on the separate training of transmitter-side pre-distortion and receiver-side equalization. This approach ove…
▽ More
The use of directly modulated lasers (DMLs) is attractive in low-power, cost-constrained short-reach optical links. However, their limited modulation bandwidth can induce waveform distortion, undermining their data throughput. Traditional distortion mitigation techniques have relied mainly on the separate training of transmitter-side pre-distortion and receiver-side equalization. This approach overlooks the potential gains obtained by simultaneous optimization of transmitter (constellation and pulse sha**) and receiver (equalization and symbol demap**). Moreover, in the context of DML operation, the choice of laser-driving configuration parameters such as the bias current and peak-to-peak modulation current has a significant impact on system performance. We propose a novel end-to-end optimization approach for DML systems, incorporating the learning of bias and peak-to-peak modulation current to the optimization of constellation points, pulse sha** and equalization. The simulation of the DML dynamics is based on the use of the laser rate equations at symbol rates between 15 and 25 Gbaud. The resulting output sequences from the rate equations are used to build a differentiable data-driven model, simplifying the calculation of gradients needed for end-to-end optimization. The proposed end-to-end approach is compared to 3 additional benchmark approaches: the uncompensated system without equalization, a receiver-side finite impulse response equalization approach and an end-to-end approach with learnable pulse shape and nonlinear Volterra equalization but fixed bias and peak-to-peak modulation current. The numerical simulations on the four approaches show that the joint optimization of bias, peak-to-peak current, constellation points, pulse sha** and equalization outperforms all other approaches throughout the tested symbol rates.
△ Less
Submitted 16 May, 2024;
originally announced May 2024.
-
BICM-compatible Rate Adaptive Geometric Constellation Sha** Using Optimized Many-to-one Labeling
Authors:
Metodi Plamenov Yankov,
Smaranika Swain,
Ognjen Jovanovic,
Darko Zibar,
Francesco Da Ros
Abstract:
In this paper, a rate adaptive geometric constellation sha** (GCS) scheme which is fully backward-compatible with existing state of the art bit-interleaved coded modulation (BICM) systems is proposed and experimentally demonstrated. The system relies on optimization of the positions of the quadrature amplitude modulation (QAM) points on the I/Q plane for maximized achievable information rate, wh…
▽ More
In this paper, a rate adaptive geometric constellation sha** (GCS) scheme which is fully backward-compatible with existing state of the art bit-interleaved coded modulation (BICM) systems is proposed and experimentally demonstrated. The system relies on optimization of the positions of the quadrature amplitude modulation (QAM) points on the I/Q plane for maximized achievable information rate, while maintaining quantization and fiber nonlinear noise robustness. Furthermore, `dummy' bits are multiplexed with coded bits before map** to symbols. Rate adaptivity is achieved by tuning the ratio of coded and `dummy' bits, while maintaining a fixed forward error-correction block and a fixed modulation format size. The points' positions and their labeling are optimized using automatic differentiation. The proposed GCS scheme is compared to a time-sharing hybrid (TH) QAM modulation and the now mainstream probabilistic amplitude sha** (PAS) scheme. The TH without sha** is outperformed for all studied data rates in a simulated linear channel by up to 0.7 dB. In a linear channel, PAS is shown to outperform the proposed GCS scheme, while similar performances are reported for PAS and the proposed GCS in a simulated nonlinear fiber channel. The GCS scheme is experimentally demonstrated in a multi-span recirculating loop coherent optical fiber transmission system with a total distance of up to 3000 km. Near-continuous zero-error flexible throughput is reported as a function of the transmission distance. Up to 1-2 spans of increased reach gains are achieved at the same net data rate w.r.t. conventional QAM. At a given distance, up to 0.79 bits/2D symbol of gain w.r.t. conventional QAM is achieved. In the experiment, similar performance to PAS is demonstrated.
△ Less
Submitted 13 March, 2024; v1 submitted 10 November, 2023;
originally announced December 2023.
-
Differentiable Machine Learning-Based Modeling for Directly-Modulated Lasers
Authors:
Sergio Hernandez,
Ognjen Jovanovic,
Christophe Peucheret,
Francesco Da Ros,
Darko Zibar
Abstract:
End-to-end learning has become a popular method for joint transmitter and receiver optimization in optical communication systems. Such approach may require a differentiable channel model, thus hindering the optimization of links based on directly modulated lasers (DMLs). This is due to the DML behavior in the large-signal regime, for which no analytical solution is available. In this paper, this p…
▽ More
End-to-end learning has become a popular method for joint transmitter and receiver optimization in optical communication systems. Such approach may require a differentiable channel model, thus hindering the optimization of links based on directly modulated lasers (DMLs). This is due to the DML behavior in the large-signal regime, for which no analytical solution is available. In this paper, this problem is addressed by develo** and comparing differentiable machine learning-based surrogate models. The models are quantitatively assessed in terms of root mean square error and training/testing time. Once the models are trained, the surrogates are then tested in a numerical equalization setup, resembling a practical end-to-end scenario. Based on the numerical investigation conducted, the convolutional attention transformer is shown to outperform the other models considered.
△ Less
Submitted 4 January, 2024; v1 submitted 27 September, 2023;
originally announced September 2023.
-
Low-complexity Samples versus Symbols-based Neural Network Receiver for Channel Equalization
Authors:
Yevhenii Osadchuk,
Ognjen Jovanovic,
Stenio M. Ranzini,
Roman Dischler,
Vahid Aref,
Darko Zibar,
Francesco Da Ros
Abstract:
Low-complexity neural networks (NNs) have successfully been applied for digital signal processing (DSP) in short-reach intensity-modulated directly detected optical links, where chromatic dispersion-induced impairments significantly limit the transmission distance. The NN-based equalizers are usually optimized independently from other DSP components, such as matched filtering. This approach may re…
▽ More
Low-complexity neural networks (NNs) have successfully been applied for digital signal processing (DSP) in short-reach intensity-modulated directly detected optical links, where chromatic dispersion-induced impairments significantly limit the transmission distance. The NN-based equalizers are usually optimized independently from other DSP components, such as matched filtering. This approach may result in lower equalization performance. Alternatively, optimizing a NN equalizer to perform functionalities of multiple DSP blocks may increase transmission reach while kee** the complexity low. In this work, we propose a low-complexity NN that performs samples-to-symbol equalization, meaning that the NN-based equalizer includes match filtering and downsampling. We compare it to a samples-to-sample equalization approach followed by match filtering and downsampling in terms of performance and computational complexity. Both approaches are evaluated using three different types of NNs combined with optical preprocessing. We numerically and experimentally show that the proposed samples-to-symbol equalization approach applied for 32 GBd on-off keying (OOK) signals outperforms the samples-domain alternative kee** the computational complexity low. Additionally, the different types of NN-based equalizers are compared in terms of performance with respect to computational complexity.
△ Less
Submitted 28 August, 2023;
originally announced August 2023.
-
Rate Adaptive Geometric Constellation Sha** Using Autoencoders and Many-To-One Map**
Authors:
Metodi P. Yankov,
Ognjen Jovanovic,
Darko Zibar,
Francesco Da Ros
Abstract:
A many-to-one map** geometric constellation sha** scheme is proposed with a fixed modulation format, fixed FEC engine and rate adaptation with an arbitrarily small step. An autoencoder is used to optimize the labelings and constellation points' positions.
A many-to-one map** geometric constellation sha** scheme is proposed with a fixed modulation format, fixed FEC engine and rate adaptation with an arbitrarily small step. An autoencoder is used to optimize the labelings and constellation points' positions.
△ Less
Submitted 19 July, 2023;
originally announced July 2023.
-
Data-Driven Modeling of Directly-Modulated Lasers
Authors:
Sergio Hernandez Fernandez,
Christophe Peucheret,
Ognjen Jovanovic,
Francesco Da Ros,
Darko Zibar
Abstract:
The end-to-end optimization of links based on directly-modulated lasers may require an analytically differentiable channel. We overcome this problem by develo** and comparing differentiable laser models based on machine learning techniques.
The end-to-end optimization of links based on directly-modulated lasers may require an analytically differentiable channel. We overcome this problem by develo** and comparing differentiable laser models based on machine learning techniques.
△ Less
Submitted 15 May, 2023;
originally announced May 2023.
-
Subspace tracking for independent phase noise source separation in frequency combs
Authors:
Aleksandr Razumov,
Holger R. Heebøll,
Mario Dummont,
Osama Terra,
Bozhang Dong,
Jasper Riebesehl,
Poul Varming,
Jens E. Pedersen,
Francesco Da Ros,
John E. Bowers,
Darko Zibar
Abstract:
Advanced digital signal processing techniques in combination with ultra-wideband balanced coherent detection have enabled a new generation of ultra-high speed fiber-optic communication systems, by moving most of the processing functionalities into digital domain. In this paper, we demonstrate how digital signal processing techniques, in combination with ultra-wideband balanced coherent detection c…
▽ More
Advanced digital signal processing techniques in combination with ultra-wideband balanced coherent detection have enabled a new generation of ultra-high speed fiber-optic communication systems, by moving most of the processing functionalities into digital domain. In this paper, we demonstrate how digital signal processing techniques, in combination with ultra-wideband balanced coherent detection can enable optical frequency comb noise characterization techniques with novel functionalities. We propose a measurement method based on subspace tracking, in combination with multi-heterodyne coherent detection, for independent phase noise sources identification, separation and measurement. Our proposed measurement technique offers several benefits. First, it enables the separation of the total phase noise associated with a particular comb-line or -lines into multiple independent phase noise terms associated with different noise sources. Second, it facilitates the determination of the scaling of each independent phase noise term with comb-line number. Our measurement technique can be used to: identify the most dominant source of phase noise; gain a better understanding of the physics behind the phase noise accumulation process; and confirm, already existing, and enable better phase noise models. In general, our measurement technique provides new insights into noise behavior of optical frequency combs.
△ Less
Submitted 18 September, 2023; v1 submitted 15 May, 2023;
originally announced May 2023.
-
Rate Adaptive Autoencoder-based Geometric Constellation Sha**
Authors:
Ognjen Jovanovic,
Metodi P. Yankov,
Francesco Da Ros,
Darko Zibar
Abstract:
An autoencoder is used to optimize bit-to-symbol map**s for geometric constellation sha**. The map**s allow for net rate adaptivity without additional hardware complexity, while achieving up to 300km of transmission distance compared to uniform QAM.
An autoencoder is used to optimize bit-to-symbol map**s for geometric constellation sha**. The map**s allow for net rate adaptivity without additional hardware complexity, while achieving up to 300km of transmission distance compared to uniform QAM.
△ Less
Submitted 28 November, 2022;
originally announced January 2023.
-
Reservoir Computing-based Multi-Symbol Equalization for PAM 4 Short-reach Transmission
Authors:
Yevhenii Osadchuk,
Ognjen Jovanovic,
Darko Zibar,
Francesco Da Ros
Abstract:
We propose spectrum-sliced reservoir computer-based (RC) multi-symbol equalization for 32-GBd PAM4 transmission. RC with 17 symbols at the output achieves an order of magnitude reduction in multiplications/symbol versus single output case while maintaining simple training.
We propose spectrum-sliced reservoir computer-based (RC) multi-symbol equalization for 32-GBd PAM4 transmission. RC with 17 symbols at the output achieves an order of magnitude reduction in multiplications/symbol versus single output case while maintaining simple training.
△ Less
Submitted 29 November, 2022;
originally announced December 2022.
-
Geometric Constellation Sha** for Fiber-Optic Channels via End-to-End Learning
Authors:
Ognjen Jovanovic,
Francesco Da Ros,
Darko Zibar,
Metodi P. Yankov
Abstract:
End-to-end learning has become a popular method to optimize a constellation shape of a communication system. When the channel model is differentiable, end-to-end learning can be applied with conventional backpropagation algorithm for optimization of the shape. A variety of optimization algorithms have also been developed for end-to-end learning over a non-differentiable channel model. In this pape…
▽ More
End-to-end learning has become a popular method to optimize a constellation shape of a communication system. When the channel model is differentiable, end-to-end learning can be applied with conventional backpropagation algorithm for optimization of the shape. A variety of optimization algorithms have also been developed for end-to-end learning over a non-differentiable channel model. In this paper, we compare gradient-free optimization method based on the cubature Kalman filter, model-free optimization and backpropagation for end-to-end learning on a fiber-optic channel modeled by the split-step Fourier method. The results indicate that the gradient-free optimization algorithms provide a decent replacement to backpropagation in terms of performance at the expense of computational complexity. Furthermore, the quantization problem of finite bit resolution of the digital-to-analog and analog-to-digital converters is addressed and its impact on geometrically shaped constellations is analysed. Here, the results show that when optimizing a constellation with respect to mutual information, a minimum number of quantization levels is required to achieve sha** gain. For generalized mutual information, the gain is maintained throughout all of the considered quantization levels. Also, the results implied that the autoencoder can adapt the constellation size to the given channel conditions.
△ Less
Submitted 17 May, 2023; v1 submitted 8 November, 2022;
originally announced November 2022.
-
Experimental Validation of Spectral-Spatial Power Evolution Design Using Raman Amplifiers
Authors:
Mehran Soltani,
Francesco Da Ros,
Andrea Carena,
Darko Zibar
Abstract:
We experimentally validate a machine learning-enabled Raman amplification framework, capable of jointly sha** the signal power evolution in two domains: frequency and fiber distance. The proposed experiment addresses the amplification in the whole C-band, by optimizing four first-order counter-propagating Raman pumps.
We experimentally validate a machine learning-enabled Raman amplification framework, capable of jointly sha** the signal power evolution in two domains: frequency and fiber distance. The proposed experiment addresses the amplification in the whole C-band, by optimizing four first-order counter-propagating Raman pumps.
△ Less
Submitted 16 May, 2022;
originally announced June 2022.
-
Flexible Raman Amplifier Optimization Based on Machine Learning-aided Physical Stimulated Raman Scattering Model
Authors:
Metodi Plamenov Yankov,
Francesco Da Ros,
Uiara Celine de Moura,
Andrea Carena,
Darko Zibar
Abstract:
The problem of Raman amplifier optimization is studied. A differentiable interpolation function is obtained for the Raman gain coefficient using machine learning (ML), which allows for the gradient descent optimization of forward-propagating Raman pumps. Both the frequency and power of an arbitrary number of pumps in a forward pum** configuration are then optimized for an arbitrary data channel…
▽ More
The problem of Raman amplifier optimization is studied. A differentiable interpolation function is obtained for the Raman gain coefficient using machine learning (ML), which allows for the gradient descent optimization of forward-propagating Raman pumps. Both the frequency and power of an arbitrary number of pumps in a forward pum** configuration are then optimized for an arbitrary data channel load and span length. The forward propagation model is combined with an experimentally-trained ML model of a backward-pum** Raman amplifier to jointly optimize the frequency and power of the forward amplifier's pumps and the powers of the backward amplifier's pumps. The joint forward and backward amplifier optimization is demonstrated for an unrepeatered transmission of 250 km. A gain flatness of $<$ 1~dB over 4 THz is achieved. The optimized amplifiers are validated using a numerical simulator.
△ Less
Submitted 13 June, 2022;
originally announced June 2022.
-
Spectral and spatial power evolution design with machine learning-enabled Raman amplification
Authors:
Mehran Soltani,
Francesco Da Ros,
Andrea Carena,
Darko Zibar
Abstract:
We present a machine learning (ML) framework for designing desired signal power profiles over the spectral and spatial domains in the fiber span. The proposed framework adjusts the Raman pump power values to obtain the desired two-dimensional (2D) profiles using a convolutional neural network (CNN) followed by the differential evolution (DE) technique. The CNN learns the map** between the 2D pro…
▽ More
We present a machine learning (ML) framework for designing desired signal power profiles over the spectral and spatial domains in the fiber span. The proposed framework adjusts the Raman pump power values to obtain the desired two-dimensional (2D) profiles using a convolutional neural network (CNN) followed by the differential evolution (DE) technique. The CNN learns the map** between the 2D profiles and their corresponding pump power values using a data-set generated by exciting the amplification setup. Nonetheless, its performance is not accurate for designing 2D profiles of practical interest, such as a 2D flat or a 2D symmetric (with respect to the middle point in distance). To adjust the pump power values more accurately, the DE fine-tunes the power values initialized by the CNN to design the proposed 2D profile with a lower cost value. In the fine-tuning process, the DE employs the direct amplification model which consists of 8 bidirectional propagating pumps, including 2 second-order and 6 first order, in an 80 km fiber span. We evaluate the framework to design broadband 2D flat and symmetric power profiles, as two goals for wavelength division multiplexing (WDM) system performing over the whole C-band. Results indicate the framework's ability to achieve maximum power excursion of 2.81 dB for a 2D flat, and maximum asymmetry of 14% for a 2D symmetric profile.
△ Less
Submitted 25 November, 2021;
originally announced December 2021.
-
End-to-end Learning of a Constellation Shape Robust to Channel Condition Uncertainties
Authors:
Ognjen Jovanovic,
Metodi P. Yankov,
Francesco Da Ros,
Darko Zibar
Abstract:
Vendor interoperability is one of the desired future characteristics of optical networks. This means that the transmission system needs to support a variety of hardware with different components, leading to system uncertainties throughout the network. For example, uncertainties in signal-to-noise ratio and laser linewidth can negatively affect the quality of transmission within an optical network…
▽ More
Vendor interoperability is one of the desired future characteristics of optical networks. This means that the transmission system needs to support a variety of hardware with different components, leading to system uncertainties throughout the network. For example, uncertainties in signal-to-noise ratio and laser linewidth can negatively affect the quality of transmission within an optical network due to e.g. mis-parametrization of the transceiver signal processing algorithms. In this paper, we propose to geometrically optimize a constellation shape that is robust to uncertainties in the channel conditions by utilizing end-to-end learning. In the optimization step, the channel model includes additive noise and residual phase noise. In the testing step, the channel model consists of laser phase noise, additive noise and blind phase search as the carrier phase recovery algorithm. Two noise models are considered for the additive noise: white Gaussian noise and nonlinear interference noise model for fiber nonlinearities. The latter models the behavior of an optical fiber channel more accurately because it considers the nonlinear effects of the optical fiber. For this model, the uncertainty in the signal-to-noise ratio can be divided between amplifier noise figures and launch power variations. For both noise models, our results indicate that the learned constellations are more robust to uncertainties in channel conditions compared to a standard constellation scheme such as quadrature amplitude modulation and standard geometric constellation sha** techniques.
△ Less
Submitted 22 April, 2022; v1 submitted 16 November, 2021;
originally announced November 2021.
-
End-to-end Learning of a Constellation Shape Robust to Variations in SNR and Laser Linewidth
Authors:
Ognjen Jovanovic,
Metodi P. Yankov,
Francesco Da Ros,
Darko Zibar
Abstract:
We propose an autoencoder-based geometric sha** that learns a constellation robust to SNR and laser linewidth estimation errors. This constellation maintains sha** gain in mutual information (up to 0.3 bits/symbol) with respect to QAM over various SNR and laser linewidth values.
We propose an autoencoder-based geometric sha** that learns a constellation robust to SNR and laser linewidth estimation errors. This constellation maintains sha** gain in mutual information (up to 0.3 bits/symbol) with respect to QAM over various SNR and laser linewidth values.
△ Less
Submitted 1 June, 2021;
originally announced June 2021.
-
Inverse design of Raman amplifier in frequency and distance domain using Convolutional Neural Networks
Authors:
Mehran Soltani,
Francesco Da Ros,
Andrea Carena,
Darko Zibar
Abstract:
We present a Convolutional Neural Network (CNN) architecture for inverse Raman amplifier design. This model aims at finding the pump powers and wavelengths required for a target signal power evolution, both in distance along the fiber and in frequency. Using the proposed framework, the prediction of the pump configuration required to achieve a target power profile is demonstrated numerically with…
▽ More
We present a Convolutional Neural Network (CNN) architecture for inverse Raman amplifier design. This model aims at finding the pump powers and wavelengths required for a target signal power evolution, both in distance along the fiber and in frequency. Using the proposed framework, the prediction of the pump configuration required to achieve a target power profile is demonstrated numerically with high accuracy in C-band considering both counter-propagating and bidirectional pum** schemes. For a distributed Raman amplifier based on a 100 km single-mode fiber, a low mean set (0.51, 0.54 and 0.64 dB) and standard deviation set (0.62, 0.43 and 0.38 dB) of the maximum test error are obtained numerically employing 2 and 3 counter, and 4 bidirectional propagating pumps, respectively.
△ Less
Submitted 21 February, 2021;
originally announced March 2021.
-
Gradient-free training of autoencoders for non-differentiable communication channels
Authors:
Ognjen Jovanovic,
Metodi Plamenov Yankov,
Francesco Da Ros,
Darko Zibar
Abstract:
Training of autoencoders using the back-propagation algorithm is challenging for non-differential channel models or in an experimental environment where gradients cannot be computed. In this paper, we study a gradient-free training method based on the cubature Kalman filter. To numerically validate the method, the autoencoder is employed to perform geometric constellation sha** on differentiable…
▽ More
Training of autoencoders using the back-propagation algorithm is challenging for non-differential channel models or in an experimental environment where gradients cannot be computed. In this paper, we study a gradient-free training method based on the cubature Kalman filter. To numerically validate the method, the autoencoder is employed to perform geometric constellation sha** on differentiable communication channels, showing the same performance as the back-propagation algorithm. Further investigation is done on a non-differentiable communication channel that includes: laser phase noise, additive white Gaussian noise and blind phase search-based phase noise compensation. Our results indicate that the autoencoder can be successfully optimized using the proposed training method to achieve better robustness to residual phase noise with respect to standard constellation schemes such as Quadrature Amplitude Modulation and Iterative Polar Modulation for the considered conditions.
△ Less
Submitted 19 August, 2021; v1 submitted 21 December, 2020;
originally announced December 2020.
-
Experimental Demonstration of Optoelectronic Equalization for Short-reach Transmission with Reservoir Computing
Authors:
Stenio M. Ranzini,
Roman Dischler,
Francesco da Ros,
Henning Buelow,
Darko Zibar
Abstract:
A receiver with shared complexity between optical and digital domains is experimentally demonstrated. Reservoir computing is used to equalize up to 4 directly-detected optically filtered spectral slices of a 32 GBd OOK signal over up to 80 km of SMF.
A receiver with shared complexity between optical and digital domains is experimentally demonstrated. Reservoir computing is used to equalize up to 4 directly-detected optically filtered spectral slices of a 32 GBd OOK signal over up to 80 km of SMF.
△ Less
Submitted 8 October, 2020;
originally announced October 2020.
-
End-to-end optimization of coherent optical communications over the split-step Fourier method guided by the nonlinear Fourier transform theory
Authors:
Simone Gaiarin,
Francesco Da Ros,
Rasmus T. Jones,
Darko Zibar
Abstract:
Optimizing modulation and detection strategies for a given channel is critical to maximize the throughput of a communication system. Such an optimization can be easily carried out analytically for channels that admit closed-form analytical models. However, this task becomes extremely challenging for nonlinear dispersive channels such as the optical fiber. End-to-end optimization through autoencode…
▽ More
Optimizing modulation and detection strategies for a given channel is critical to maximize the throughput of a communication system. Such an optimization can be easily carried out analytically for channels that admit closed-form analytical models. However, this task becomes extremely challenging for nonlinear dispersive channels such as the optical fiber. End-to-end optimization through autoencoders (AEs) can be applied to define symbol-to-waveform (modulation) and waveform-to-symbol (detection) map**s, but so far it has been mainly shown for systems relying on approximate channel models. Here, for the first time, we propose an AE scheme applied to the full optical channel described by the nonlinear Schr\{"o}dinger equation (NLSE). Transmitter and receiver are jointly optimized through the split-step Fourier method (SSFM) which accurately models an optical fiber. In this first numerical analysis, the detection is performed by a neural network (NN), whereas the symbol-to-waveform map** is aided by the nonlinear Fourier transform (NFT) theory in order to simplify and guide the optimization on the modulation side. This proof-of-concept AE scheme is thus benchmarked against a standard NFT-based system and a threefold increase in achievable distance (from 2000 to 6640 km) is demonstrated.
△ Less
Submitted 7 October, 2020; v1 submitted 11 September, 2020;
originally announced September 2020.
-
Phase Compensation for Continuous Variable Quantum Key Distribution
Authors:
Hou-Man Chin,
Darko Zibar,
Nitin Jain,
Tobias Gehring,
Ulrik L. Andersen
Abstract:
The tracking and compensation of phase noise is critical to reducing excess noise for continuous variable quantum key distribution schemes. This work demonstrates the effectiveness of unscented Kalman filter for phase noise compensation.
The tracking and compensation of phase noise is critical to reducing excess noise for continuous variable quantum key distribution schemes. This work demonstrates the effectiveness of unscented Kalman filter for phase noise compensation.
△ Less
Submitted 17 May, 2020;
originally announced May 2020.
-
End-to-end Learning for GMI Optimized Geometric Constellation Shape
Authors:
Rasmus T. Jones,
Metodi P. Yankov,
Darko Zibar
Abstract:
Autoencoder-based geometric sha** is proposed that includes optimizing bit map**s. Up to 0.2 bits/QAM symbol gain in GMI is achieved for a variety of data rates and in the presence of transceiver impairments. The gains can be harvested with standard binary FEC at no cost w.r.t. conventional BICM.
Autoencoder-based geometric sha** is proposed that includes optimizing bit map**s. Up to 0.2 bits/QAM symbol gain in GMI is achieved for a variety of data rates and in the presence of transceiver impairments. The gains can be harvested with standard binary FEC at no cost w.r.t. conventional BICM.
△ Less
Submitted 19 July, 2019;
originally announced July 2019.
-
An ultra-fast method for gain and noise prediction of Raman amplifiers
Authors:
Ann Margareth Rosa Brusin,
Vittorio Curri,
Darko Zibar,
Andrea Carena
Abstract:
A machine learning method for prediction of Raman gain and noise spectra is presented: it guarantees high-accuracy (RMSE < 0.4 dB) and low computational complexity making it suitable for real-time implementation in future optical networks controllers.
A machine learning method for prediction of Raman gain and noise spectra is presented: it guarantees high-accuracy (RMSE < 0.4 dB) and low computational complexity making it suitable for real-time implementation in future optical networks controllers.
△ Less
Submitted 2 May, 2019;
originally announced May 2019.
-
Optical Frequency Comb Noise Characterization Using Machine Learning
Authors:
Giovanni Brajato,
Lars Lundberg,
Victor Torres-Company,
Darko Zibar
Abstract:
A novel tool, based on Bayesian filtering framework and expectation maximization algorithm, is numerically and experimentally demonstrated for accurate frequency comb noise characterization. The tool is statistically optimum in a mean-square-error-sense, works at wide range of SNRs and offers more accurate noise estimation compared to conventional methods.
A novel tool, based on Bayesian filtering framework and expectation maximization algorithm, is numerically and experimentally demonstrated for accurate frequency comb noise characterization. The tool is statistically optimum in a mean-square-error-sense, works at wide range of SNRs and offers more accurate noise estimation compared to conventional methods.
△ Less
Submitted 23 April, 2019;
originally announced April 2019.
-
Dual-polarization NFDM transmission with continuous and discrete spectral modulation
Authors:
F. Da Ros,
S. Civelli,
S. Gaiarin,
E. P. da Silva,
N. De Renzis,
M. Secondini,
D. Zibar
Abstract:
Nonlinear distortion experienced by signals during their propagation through optical fibers strongly limits the throughput of optical communication systems. Recently, a strong research focus has been dedicated to nonlinearity mitigation and compensation techniques. At the same time, a more disruptive approach, the nonlinear Fourier transform (NFT), aims at designing signaling schemes more suited t…
▽ More
Nonlinear distortion experienced by signals during their propagation through optical fibers strongly limits the throughput of optical communication systems. Recently, a strong research focus has been dedicated to nonlinearity mitigation and compensation techniques. At the same time, a more disruptive approach, the nonlinear Fourier transform (NFT), aims at designing signaling schemes more suited to the nonlinear fiber channel. In a short period, impressive results have been reported by modulating either the continuous spectrum or the discrete spectrum. Additionally, very recent works further introduced the opportunity to modulate both spectra for single polarization transmission. Here, we extend the joint modulation scheme to dual-polarization transmission by introducing the framework to construct a dual-polarization optical signal with the desired continuous and discrete spectra. After a brief analysis of the numerical algorithms used to implement the proposed scheme, the first experimental demonstration of dual-polarization joint nonlinear frequency division multiplexing (NFDM) modulation is reported for up to 3200 km of low-loss transmission fiber. The proposed dual-polarization joint modulation schemes enables to exploit all the degrees of freedom for modulation (both polarizations and both spectra) provided by a single-mode fiber (SMF).
△ Less
Submitted 26 February, 2019;
originally announced March 2019.
-
Optical wavelength conversion of high bandwidth phase-encoded signals in a high FOM 50cm CMOS compatible waveguide
Authors:
Francesco Da Ros,
Edson Porto da Silva,
Darko Zibar,
Sai T. Chu,
Brent E. Little,
Roberto Morandotti,
Michael Galili,
David J. Moss,
Leif K. Oxenløwe
Abstract:
We demonstrate wavelength conversion of QAM signals including 32GBd QPSK and 10GBd 16QAM in a 50cm long high index doped glass spiral waveguide. The quality of the generated idlers over a 10nm bandwidth is sufficient to achieve a BER performance below the HD FEC threshold (less than 3.8 x 10-3), with an OSNR penalty of less than 0.3 dB compared to the original signal. Our results confirm that this…
▽ More
We demonstrate wavelength conversion of QAM signals including 32GBd QPSK and 10GBd 16QAM in a 50cm long high index doped glass spiral waveguide. The quality of the generated idlers over a 10nm bandwidth is sufficient to achieve a BER performance below the HD FEC threshold (less than 3.8 x 10-3), with an OSNR penalty of less than 0.3 dB compared to the original signal. Our results confirm that this is a promising platform for nonlinear optical signal processing, a result of both very low linear propagation loss (less than 0.07 dB/cm) and the large material bandgap that ensures negligible nonlinear loss at telecom wavelengths.
△ Less
Submitted 7 August, 2018;
originally announced August 2018.
-
Effect of filter shape on excess noise performance in continuous variable quantum key distribution with Gaussian modulation
Authors:
Hou-Man Chin,
Nitin Jain,
Darko Zibar,
Tobias Gehring,
Ulrik L. Andersen
Abstract:
An attractive implementation for quantum cryptography is the continuous variable variation, as it relies on standard telecommunication components. Modulating the quantum signal using a Gaussian format is attractive since it has been proven to be secure. This work investigates the effect of the roll-off of a root raised cosine pulse sha** and matched filter on the excess noise performance of a Ga…
▽ More
An attractive implementation for quantum cryptography is the continuous variable variation, as it relies on standard telecommunication components. Modulating the quantum signal using a Gaussian format is attractive since it has been proven to be secure. This work investigates the effect of the roll-off of a root raised cosine pulse sha** and matched filter on the excess noise performance of a Gaussian modulated quantum key distribution system in a simulated back to back configuration. Contrary to intuition, it is found that the roll-off parameter does not significantly impact the performance of the system.
△ Less
Submitted 14 August, 2018;
originally announced August 2018.
