arXiv subjects
Publications and source records attributed to Vahid Aref.
Equalization-enhanced phase noise (EEPN) emerges as a key performance limitation in high symbol-rate coherent transmission systems. In this paper, we highlight recent advances in modeling EEPN and show that the temporal Gaussian noise model reproduces the characteristic burst-like SNR degradation, enabling efficient system simulation.
Equalization-enhanced Phase Noise causes burst-like distortions in high symbol-rate transmission systems. We propose a temporal Gaussian noise model that captures these distortions by introducing a time-varying distortion power. Validated through simulations and experiments, it enables accurate and simple performance prediction for high symbol-rate transmission systems.
Probabilistic constellation shaping enables easy rate adaption and has been proven to reduce the gap to Shannon capacity. Constellation point probabilities are optimized to maximize either the mutual information or the bit-wise mutual information. The optimization problem is however challenging even for simple channel models. While autoencoder-based machine learning has been applied successfully to solve this problem [1], it requires manual computation of additional terms for the gradient which is an error-prone task. In this work, we present novel loss functions for autoencoder-based learning of probabilistic constellation shaping for coded modulation systems using automatic differentiation and importance sampling. We show analytically that our proposed approach also uses exact gradients of the constellation point probabilities for the optimization. In simulations, our results closely match the results from [1] for the additive white Gaussian noise channel and a simplified model of the intensity-modulation direct-detection channel.
In coherent optical communication systems the laser phase noise is commonly modeled as a Wiener process. We propose a sliding-window based linearization of the phase noise, enabling a novel description. We show that, by stochastically modeling the residual error introduced by this approximation, equalization-enhanced phase noise (EEPN) can be described and decomposed into four different components. Furthermore, we analyze the four components separately and provide a stochastical model for each of them. This novel model allows to predict the impact of well-known algorithms in coherent digital signal processing (DSP) pipelines such as timing recovery (TR) and carrier phase recovery (CPR) on each of the terms. Thus, it enables to approximate the resulting signal affected by EEPN after each of these DSP steps and helps to derive appropriate ways of mitigating such effects.
We show that equalization-enhanced phase noise manifests as a time-varying, frequency-dependent phase error, which can be modeled and reversed by a time-varying all-pass finite impulse response filter.
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 keeping 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 keeping the computational complexity low. Additionally, the different types of NN-based equalizers are compared in terms of performance with respect to computational complexity.
We examine here what type of predictive modelling, classification, or regression, using neural networks (NN), fits better the task of soft-demapping based post-processing in coherent optical communications, where the transmission channel is nonlinear and dispersive. For the first time, we present possible drawbacks in using each type of predictive task in a machine learning context, considering the nonlinear coherent optical channel equalization/soft-demapping problem. We study two types of equalizers based on the feed-forward and recurrent NNs, for several transmission scenarios, in linear and nonlinear regimes of the optical channel. We point out that even though from the information theory perspective the cross-entropy loss (classification) is the most suitable option for our problem, the NN models based on the cross-entropy loss function can severely suffer from learning problems. The latter translates into the fact that regression-based learning is typically superior in terms of delivering higher Q-factor and achievable information rates. In short, we show by empirical evidence that loss functions based on cross-entropy may not be necessarily the most suitable option for training communication systems in practical scenarios when overfitting- and vanishing gradients-related problems come into play.
Autoencoder-based deep learning is applied to jointly optimize geometric and probabilistic constellation shaping for optical coherent communication. The optimized constellation shaping outperforms the 256 QAM Maxwell-Boltzmann probabilistic distribution with extra 0.05 bits/4D-symbol mutual information for 64 GBd transmission over 170 km SMF link.
We present a simple, efficient "direct learning" approach to train Volterra series-based digital pre-distortion filters using neural networks. We show its superior performance over conventional training methods using a 64-QAM 64-GBaud simulated transmitter with varying transmitter nonlinearity and noisy conditions.
We present a novel autoencoder-based learning of joint geometric and probabilistic constellation shaping for coded-modulation systems. It can maximize either the mutual information (for symbol-metric decoding) or the generalized mutual information (for bit-metric decoding).
We report on theoretical and experimental investigations of the nonlinear tolerance of single carrier and digital multicarrier approaches with probabilistically shaped constellations. Experimental transmission of PCS16QAM is assessed at 120 GBd over an ultra-long-haul distance.
Commercial coherent receivers utilize balanced photodetectors (PDs) with high single-port rejection ratio (SPRR) to mitigate the signal-signal beat interference (SSBI) due to the square-law detection process. As the symbol rates of coherent transponders are increased to 100 Gbaud and beyond, maintaining a high SPRR in a cost-effective manner becomes more and more challenging. One potential approach for solving this problem is to leverage the concept of single-ended coherent receiver (SER) where single-ended PDs are used instead of the balanced PDs. In this case, the resulting SSBI should be mitigated in the digital domain. In this paper, we show that SSBI can be effectively mitigated using various low-complexity techniques, such as the direct filed reconstruction (DFR), clipped iterative SSBI cancellation (CIC) and gradient decent (GD). In addition, we present a self-calibration technique for SERs which can be extended for characterizing the optical-to-electrical (O/E) response of a conventional balanced coherent receiver (BR). Using the developed techniques, we then experimentally demonstrate a 90 Gbaud probabilistically constellation shaped 64-QAM (PCS-64QAM) transmission using a SER, achieving a net data rate of 882 Gb/s over 100 km of standard single mode fiber (SSMF). The sensitivity penalty compared to the BR is below 0.5 dB. We expect that when the symbol rate is increased further, a SER can potentially outperform a BR, especially when applied to cost-sensitive commercial pluggable coherent transceivers
We present a novel end-to-end autoencoder-based learning for coherent optical communications using a "parallelizable" perturbative channel model. We jointly optimized constellation shaping and nonlinear pre-emphasis achieving mutual information gain of 0.18 bits/sym./pol. simulating 64 GBd dual-polarization single-channel transmission over 30x80 km G.652 SMF link with EDFAs.
We investigate end-to-end optimized optical transmission systems based on feedforward or bidirectional recurrent neural networks (BRNN) and deep learning. In particular, we report the first experimental demonstration of a BRNN auto-encoder, highlighting the performance improvement achieved with recurrent processing for communication over dispersive nonlinear channels.
We investigate methods for experimental performance enhancement of auto-encoders based on a recurrent neural network (RNN) for communication over dispersive nonlinear channels. In particular, our focus is on the recently proposed sliding window bidirectional RNN (SBRNN) optical fiber autoencoder. We show that adjusting the processing window in the sequence estimation algorithm at the receiver improves the reach of simple systems trained on a channel model and applied "as is" to the transmission link. Moreover, the collected experimental data was used to optimize the receiver neural network parameters, allowing to transmit 42 Gb/s with bit-error rate (BER) below the 6.7% hard-decision forward error correction threshold at distances up to 70km as well as 84 Gb/s at 20 km. The investigation of digital signal processing (DSP) optimized on experimental data is extended to pulse amplitude modulation with receivers performing sliding window sequence estimation using a feed-forward or a recurrent neural network as well as classical nonlinear Volterra equalization. Our results show that, for fixed algorithm memory, the DSP based on deep learning achieves an improved BER performance, allowing to increase the reach of the system.
Optical nonlinear Fourier transform-based communication systems require an accurate estimation of a signal's nonlinear spectrum, computed usually by piecewise approximation methods on the signal samples. We propose an algorithm, named successive eigenvalue removal, to improve the spectrum estimation of a multi-soliton pulse. It exploits a property of the Darboux transform that allows removing eigenvalues from the nonlinear spectrum. This results in a smaller pulse duration and smaller bandwidth. The spectral coefficients are estimated successively after removing the eigenvalues of a signal. As a beneficial application, we show that the algorithm decreases the computational complexity by iteratively reducing the pulse duration.
We perform an experimental end-to-end transceiver optimization via deep learning using a generative adversarial network to approximate the test-bed channel. Previously, optimization was only possible through a prior assumption of an explicit simplified channel model.
In this paper, we apply deep learning for communication over dispersive channels with power detection, as encountered in low-cost optical intensity modulation/direct detection (IM/DD) links. We consider an autoencoder based on the recently proposed sliding window bidirectional recurrent neural network (SBRNN) design to realize the transceiver for optical IM/DD communication. We show that its performance can be improved by introducing a weighted sequence estimation scheme at the receiver. Moreover, we perform bit-to-symbol mapping optimization to reduce the bit-error rate (BER) of the system. Furthermore, we carry out a detailed comparison with classical schemes based on pulse-amplitude modulation and maximum likelihood sequence detection (MLSD). Our investigation shows that for a reference 42\,Gb/s transmission, the SBRNN autoencoder achieves a BER performance comparable to MLSD, when both systems account for the same amount of memory. In contrast to MLSD, the SBRNN performance is achieved without incurring a computational complexity exponentially growing with the processed memory.