We introduce wavelength conversion technology for wideband transmission over deployed optical fibers by extending the wavelength range within the optical fiber transparent window which is not yet supported by practical transceivers.

We have demonstrated real-time S+C+L triple-band WDM transmission of coherent 200-Gb/s (34-Gbaud DP- 16QAM) signals by employing all-optical wavelength converters. 100-nm wavelength range (1504.98-1606.82 nm) transmission over 40 km is realized by using only transceivers operating at C-band.

We focus on the introduction of a multi-wavelength-band transmission system based on all optical wavelength conversion, which increases the transmission capacity in the deployed fiber infrastructure without the need for new wavelength band transceivers.

Accurate optical monitors are critical for automating operations of fiber-optic networks. Deep neural network (DNN) based optical monitors have been investigated as accurate optical monitors to leverage a large amount of data obtained from fiber-optic networks. Although DNN-based optical monitors have been trained and tested to ensure the given accuracy criteria, this does not ensure sufficient accuracy under unexpected conditions, that is, out of test conditions, e.g., a newly developed modulation format that is not included in the test dataset. Thus, it is necessary to prepare a monitor to assess the current accuracy of a DNN-based optical monitor's output for robust automation of networks. We present a DNN-based optical monitor that simultaneously outputs an optical signal-to-noise ratio and its uncertainty information using a dropout method at the inference phase. This monitor was evaluated in cases in which the DNNs were trained with either a limited number of records or partially missing records in a training dataset. The proposed monitor successfully informed that own output has large uncertainties due to a limited amount of training data or a missing part in training dataset. Additionally, to improve an accuracy of estimated uncertainty, the number of partial neural networks by dropout at the inference phase was optimized. This is a valuable step toward designing robust "self-driving" optical networks.

Wavelength conversion using all-optical phase modulation in a fiber driven by two pump waves is investigated. The operation features are analyzed using an all-optical phase modulation model with two parallel-/cross-polarized pump waves to generate a phase-preserving copy of the optical signal at an exact frequency up-/down-shifted by the two-pump detuning. The conversion efficiency is experimentally verified using a 300-m highly-nonlinear fiber. The results agree well with a theoretical prediction. The conversion bandwidth over 4 THz is achieved and error-free wavelength conversion for a 32-GBd polarization-division multiplexed 16QAM signal is demonstrated. The technique's applicability to a large-capacity wavelength-division multiplexed signals is also discussed.

To address the open and diverse situation of future optical networks, it is necessary to find a methodology to accurately estimate the value of a target quantity in an optical performance monitor (OPM) depending on the high-level monitoring objectives declared by the network operator. Using machine learning techniques partially enables a trainable OPM; however, it still requires the feature selection before the learning process. Here, we show the OPM that uses a convolutional neural network (CNN) with a digital coherent receiver to deal with the abundance of training data required for convergence and pre-processing of input data by human engineers needed for feature (representation) extraction. To proof a concept of the OPM based on CNN, we experimentally demonstrate that a CNN can learn an accurate optical signal-to-noise-ratio (OSNR) estimation functionality from asynchronously sampled data right after intradyne coherent detection. We evaluate bias errors and standard deviations of a CNN-based OSNR estimator for six combinations of modulation formats and symbol rates and confirm that the proposed OSNR estimator can provide accurate estimation results (<0.4 dB bias errors and standard deviations). Additionally, we investigate filters in the trained CNN to reveal what the CNN learned in the training phase. This is a valuable step toward designing autonomous "self-driving" optical networks.

Real-time S+C+L triple-band WDM transmission of coherent 200-Gb/s (34-Gbaud DP-16QAM) signals is demonstrated for the first time. With wide-band wavelength converters, 100-nm wavelength range (1504.98–1606.82 nm) transmission over 40 km is realized without a transceiver operating at S- or L-band.

We investigate maximum transmission reach of wavelength converted systems for variable conversion efficiencies. We propose an SBS suppression scheme to improve the conversion efficiency without generating signal crosstalk.

We use a distributed superchannel aggregation scheme to demonstrate ultra-wideband single-photodiode reception based on inherently polarization-aligned Kramers-Kronig carrier generation at the receiver. The impact of polarization rotations is discussed. 400-Gb/s capacity is demonstrated for optimized conditions. (C) 2018 The Author(s)

We experimentally demonstrate that a convolutional neural network (CNN) can acquire an accurate OSNR estimation functionality from asynchronously sampled data right after intradyne coherent detection, while the feature selection before the CNN can be omitted.

We present continuously tuneable signal frequency shifting up to 4 THz using a dispersion and dispersion-slope reduced highly nonlinear fibre. A 1.6-Tb/s DP-16QAM WDM signal is arbitrarily translated to different wavelengths across the C-band during a 160-km error-free transmission.

We present a deep-learning-based optical monitor that simultaneously outputs an optical signal-to-noise ratio and its confidence level using a dropout technique at inference time. Its behaviour is analysed with partially missing and limited number of records in training dataset.

We propose a nonlinear fiber-based optical frequency shifter using two CW pumps orthogonally polarized to signals and demonstrate THz-range error-free frequency shifting of 1.6-Tb/s DP-16QAM WDM signal without guard-band midway through 160 km transmission.

It is necessary to guarantee the operational range of machine learning (ML)-based optical physicallayer monitors (OPMs). To declare high-level monitoring objectives and obtain their values from OPMs, finding a methodology to accurately estimate the value of a target quantity and ensure their operational range is necessary. We introduce a deep neural network (DNN) with a digital coherent receiver to ML-based OPMs to deal with the abundance of training data needed for convergence and the preprocessing of input data by human engineers needed for feature (representation) extraction. However, guaranteeing the operational range of trained models on DNN-based OPMs was left for another investigation. To address this issue with DNN-based OPMs, we propose an "operational range expander," a simple treatment of the link between pre-processing training datasets and their specified operational range. We assess the operational range expander by performing simulation and experimentation using a DNNbased optical signal-to-noise ratio (OSNR) estimator. We select a laser frequency offset between a signal and a local oscillator in digital coherent receivers as an example quantity for a practical operational range expander in this study. This is because the OPMs need to work before digitally compensating frequency offset despite the difficulty in fully controlling frequency offset in practical situations. We evaluate bias errors and standard deviations of OSNR estimation from different frequency offsets ranging from -3.5 to 3.5 GHz and confirm that the provided operational range expander specified the operational range of DNNbased OSNR estimators through their training phase.

In its most simple implementation, the Kramers-Kronig (KK) scheme allows the reception of complex modulation formats with a single photodiode. Thereby, in analogy to a heterodyne receiver, an additional continuous wave signal is used that allows under certain conditions the distortion-free reconstruction of the data signal. We present experimental investigations on a novel KK scheme based on inherently polarization-aligned KK carrier generation at the receiver. We use a distributed super channel aggregation scheme to demonstrate the ultra-wideband single-photodiode reception of a 3 x 33-CBd single-polarization 32 quadratic-amplitude modulation superchannel with this scheme. The impact of polarization rotations is discussed, and we show that variations in data or KK carrier polarization will only affect the carrier-to-signal power ratio (CSPR), leading to an increase or decrease from the optimum CSPR by about 6 dB. A net capacity of 400 Gb/s is demonstrated for optimized conditions. In addition, we demonstrate transmission measurements over a 60-km standard single-mode fiber link with a 344-Gb/s net rate superchannel (3 x 28 CBd sub-carriers) with the novel KK scheme.

DOI:10.14923/transcomj.2018PHI0002 イバ中の非線形補償 [3], [4],プログラマブル光送受信 器 [5], [6],エラスティック光ネットワーキング [7], [8] など,近年ネットワーク物理層に投入された多数の技 術に関連している. このような複雑性を隠蔽し,許容可能な運用管理コ ストで多数の自由度の調整を可能にするため,光ネッ トワークへの機械学習技術 [9], [10]の適用が検討され ている.例えば,文献 [11]では,伝送距離やファイバ 入力光パワー,光ファイバ種などの情報を,伝送品質 (QoT)に抽象化する技術が,文献 [12]では各ドメイ ンで機械学習技術によって推定された QoTを更に階 層的に解析し,マルチドメイン光ネットワークを制御 する検討がなされている. 本論文では特に,これらの抽象化の中で最も物理層 に近いものの一つである,測定された光波形データか らそのデータの要約,すなわち変調方式,波長分散, 偏波モード分散,光信号対雑音比(Optical Signal-toNoise Ratio, OSNR),偏波依存損失などの光ファイ バチャネルの物理層パラメータを機械学習技術により 推定することを検討する.推定されたパラメータは, 将来における物理層パラメータ変化の予測,上位抽象 化層への入力,ネットワーク制御の判断などに利用さ れる.これにより,より効率的なリソース配分及び光

We experimentally realize a wavelength-shift-free optical phase conjugator exploiting inherent suppression of the original signal in a polarization-diversity loop. We achieve 0.5-dB Q(2)-factor improvement in 200 Gb/s PDM-16QAM transmission over 800 km.

Enhancing the capacity of optical communication networks is essential to achieving a hyperconnected world in which people, information, and things are connected and to enabling continued development of information and communications technologies (ICT) such as the Internet of Things (IoT), big data, artificial intelligence, and 5G mobile communications. In particular, increasing optical fiber transmission capacity to more than 100 Tbps by 2020 or shortly thereafter is needed to handle the ever-increasing volume of digital data traffic. Since conventional technologies are getting close to the transmission limit, technological breakthroughs enabling higher capacity must be made. Given this requirement, we are researching and developing key technologies for optical transceivers and optical nodes that will enable transmission capacity to be increased. In this paper, we introduce our recent advances in optical modulation and demodulation technologies for sending and receiving large-capacity signals and in optical node technologies for achieving energy-saving broadband optical signal switching.

We propose an optical frequency shifter for dual-polarization signals using frequency conversion in fiber. The proposed optical frequency shifter is achieved by transmitting input signal combined with two orthogonally polarized continuous wave pump lights through a nonlinear fiber and a polarizer. In the scheme, a polarization switched and frequency shifted copy of an original signal is generated in the nonlinear fiber and the subsequent polarizer removes the original signal in a wavelength independent fashion. It allows a fractional bandwidth shift, in which the shifted signal is overlapped with the original signal. We achieve THz range frequency shift with original signal suppression without an optical filter by using a polarization diversity loop. A polarization beam splitter is used as a splitter and a combiner of the loop as well as for original signal rejection. A contention resolution of two dual-polarization 400-Gb/s superchannels is demonstrated by applying the optical frequency shifter with acceptable in-band crosstalk. For any-to-any optical frequency shift, wavelength arrangement for pumps and required zero dispersion and dispersion slope of nonlinear fibers are numerically simulated.

We propose and experimentally demonstrate a proof-of-concept of a programmable optical transceiver that enables simultaneous optimization of multiple programmable parameters (modulation format, symbol rate, power allocation, and FEC) for satisfying throughput, signal quality, and latency requirements. The proposed optical transceiver also accommodates multiple sub-channels that can transport different optical signals with different requirements. Multi-degree-of-freedom of the parameters often leads to difficulty in finding the optimum combination among the parameters due to an explosion of the number of combinations. The proposed optical transceiver reduces the number of combinations and finds feasible sets of programmable parameters by using constraints of the parameters combined with a precise analytical model. For precise BER prediction with the specified set of parameters, we model the sub-channel BER as a function of OSNR, modulation formats, symbol rates, and power difference between sub-channels. Next, we formulate simple constraints of the parameters and combine the constraints with the analytical model to seek feasible sets of programmable parameters. Finally, we experimentally demonstrate the end-to-end operation of the proposed optical transceiver with offline manner including low-density parity-check (LDPC) FEC encoding and decoding under a specific use case with latency-sensitive application and 40-km transmission.