Multi-Task Learning Convolutional Neural Network and Optical Spectrums Enabled Optical Performance Monitoring

Optical signal-to-noise ratio (OSNR) monitoring is essential to both the operation of reliable and reconfigurable network and the supply of high quality-of-service. Recently, deep learning technique has been proposed for the implementation of OSNR monitoring. However, the potential of deep learning technique for OSNR monitoring has not been fully exploited in terms of the monitoring accuracy and robust operation. Here, we propose an OSNR monitoring scheme with high accuracy and short response time using the long short-term memory neural network (LSTM-NN). The use of LSTM-NN is helpful to identify the relationship between the time-varied data and corresponding OSNR without manual feature extraction. We investigate the optimal number of time steps for the LSTM-NN-based OSNR monitoring and the monitoring accuracy with respect to various modulation formats with variable baud-rates. Finally, we implement an experimental verification for 34.94-GBd polarization division multiplexing-16 quadrature amplitude modulation signal, achieving a mean absolute error of 0.05 dB over the OSNR range from 15 to 25 dB. The advantages of LSTM-NN-based OSNR monitoring include high accuracy and smart response, which is ideal for future agile optical network.

This paper presents a novel joint optical performance monitoring (OPM) scheme for multiple spatial modes in mode division multiplexing elastic optical network (MDM-EON) systems, which integrates the optimal Radon transform (ORT) image with the improved multi-task multi-input multi-output residual neural network (IMT-MIMO-ResNet) model. The ORT images effectively enhance the differentiation among various parameters across different spatial modes. Additionally, based on the output characteristics and intrinsic correlations among spatial modes of the MDM-EON system, we design the IMT-MIMO-ResNet model. The proposed scheme overcomes the limitations of existing approaches, where a single OPM device can monitor only one spatial mode. It enables simultaneous monitoring of bit rates (BRs), modulation formats (MFs), mode coupling (MC) intensities, and optical signal-to-noise ratio (OSNR) values for each MF across five spatial modes. Moreover, the scheme demonstrates favorable robustness under conditions with large chromatic dispersion, large differential group delay, and strong fiber nonlinear effects. Both simulation and experimental results show that the accuracy of baud rate identification (BRI) and modulation format identification (MFI) for each spatial mode reaches 100%, and the average mean absolute errors (MAEs) for OSNR values and MC intensities corresponding to each MF are below 0.4 dB and 0.010, respectively. The computational complexity of the proposed scheme, evaluated using floating-point operations (FLOPs) and parameters (Params), is approximately 1/7 of that required by a typical OPM scheme.

The reference optical spectrum (ROS) based in-band optical signal-to-noise-ratio (OSNR) monitoring methods are accurate and inherently robust to fiber chromatic dispersion and polarization mode dispersion. They are also very simple and suitable to be deployed ubiquitously in the transmission system. In this paper, a ROS-based in-band OSNR monitoring method is proposed for erbium doped fiber amplifier (EDFA) amplified multispan dense wavelength division multiplexing systems with cascaded filtering effect (CFE) caused by cascaded add-drop filters or wavelength selective switches. In such systems, the optical signal as well as the optical noise may experience significant CFE making the ROS-based methods proposed before ineffective. The new method solves this problem and performs very well for OSNR in the range of 10–30 dB when the fiber nonlinear effect and nonideal filtering effect are present. The OSNR monitor based on the new method is also more convenient to use as only one-time calibration is required when used at different monitoring locations along the optical link.

Cost-effective OSNR monitoring with large chromatic dispersion tolerance using random forest for intermediate nodes

OSNR monitoring technique based on multi-order statistical moment method and correlation function for PM-16QAM in presence of fiber nonlinearities

We propose an optical performance monitoring technique for simultaneous monitoring of optical signal-to-noise ratio (OSNR), chromatic dispersion (CD), and polarization-mode dispersion (PMD) using an artificial neural network trained with asynchronous amplitude histograms (AAHs). Simulations are conducted to demonstrate the technique for both 40-Gb/s return-to-zero differential quadrature phase-shift keying (RZ-DQPSK) and 40-Gb/s noneturn-to-zero 16 quadrature amplitude modulation (16-QAM) systems. The OSNR, CD, and PMD monitoring range and root-mean-square (rms) errors are 10-30 and 0.43 dB, 0-400 and 9.82 ps/nm, and 0-10 and 0.92 ps, respectively, for RZ-DQPSK systems. For 16-QAM system, the monitoring range and rms errors are 10-30 and 0.2 dB, 0-400 and 9.66 ps/nm, and 0-30 and 0.65 ps for OSNR, CD, and PMD, respectively. As the generation of AAH does not require any clock or timing recovery, the proposed technique can serve as a low-cost option to realize in-service multiparameter monitoring for the next-generation transparent optical networks.

Recently, degree of polarization (DOP) of light has been utilized as a parameter to monitor optical signal-to-noise-ratio (OSNR). However, in the presence of polarization-mode dispersion (PMD), when using DOP to determine OSNR, OSNR is always underestimated due to the depolarization effect induced by PMD. In this paper, we propose and experimentally demonstrate a simple PMD-insensitive OSNR monitoring technique based on DOP measurement. By the assistance of the transmitter-side polarization scrambling, the in-service OSNR parameter can be accurately derived from the measured maximum DOP value within the polarization scrambling period, which is immune to PMD effect. The monitoring performance is experimentally evaluated by statistical method at OSNR of 25 dB. Experimental results show that OSNR monitoring with about 1-dB standard deviation can be achieved in a 10-Gb/s NRZ-OOK system with DGD varying from 0 to 80 ps.

We experimentally demonstrate meta-learning-enabled accurate optical signal-to-noise ratio (OSNR) monitoring of directly detected 16QAM signals with extremely few training data. When one-shot training, where one amplitude histogram (AH) for each OSNR value includes only 2000 data samples, is implemented for a 16QAM signal within a variable OSNR range of 15-24 dB, the experimental root mean squared error (RMSE) of the retraining technique is 1.53 dB. For transfer learning from the 16QAM simulation to the experimentally generated AH, the RMSE can be reduced to 1.11 dB. In comparison with both the retraining and transfer learning techniques, the RMSE of meta-learning-enabled OSNR monitoring can be further reduced by 42.8% and 22.3%, respectively. In order to reach the optimal accuracy with an RMSE of 0.66 dB, the meta-learning technique requires only 15 AHs for each OSNR value to be monitored, while the retraining and the transfer learning techniques need 20 and 25 AHs, respectively.

As critical communication infrastructure, optical networks have a vital role in safe and dependable transmission of massive amounts of data, supporting essential societal services. However, these networks are inherently vulnerable to a multitude of deliberate, man-made attacks targeting service disruption at the physical layer. Physical-layer attack techniques can range in their scope and effects, level of sophistication, locality, detectability, etc. An example of a relatively unsophisticated attack method is a deliberate fiber cut, typically targeting critical network elements (e.g., links with the highest betweenness) and resulting in straightforward transmission interruption [1]. More refined attack techniques rely on the insertion of harmful signal (e.g. in- and out-of-band jamming) [2], or on external tampering with the fiber to degrade the transmission quality (e.g., polarization scrambling via fiber squeezing) [3]. Diverse attack techniques cause different effects, which complicates their detectability. For example, some attacks add unfilterable noise, some reduce the power of the affected optical channels, while some inflict changes in the state of polarization too quick for the coherent receiver to compensate [3]. Therefore, monitoring only the spectrum [4], or individual signal parameters such as the power, optical signal-to-noise ratio (OSNR), or presence of errors may result in inaccurate diagnostics and root cause attribution. This obstacle in quick recovery of affected services is further pronounced for newly emerging attack techniques whose effects may deviate from the attack signatures previously known to the network management system [5].The complexity of the evolving physical-layer security landscape and the intricate interplay of different optical performance monitoring (OPM) parameters in the presence of diverse attack methods can greatly benefit from the application of machine learning techniques capable of deep data analysis. In this talk, we present how different data analytics and machine learning approaches can be applied to interpret the OPM data reported from the commercially available coherent receivers to identify anomalous operation and trigger security threat warnings. The analytical techniques are applied to experimental data obtained from an operator’s metropolitan testbed subjected to in- and out-of-band jamming, and external polarization scrambling attacks. We begin with an analysis of the optical signal degradation caused by the different attack methods. We then investigate the application of several supervised learning approaches that, once trained on the experimental data, can detect the presence of an attack and identify its type and intensity. The accuracy of several classifiers is scrutinized, along with the relevance of OPM parameters reported by the coherent receivers and the impact of missing features. To gain insight into the potential of the network to detect emerging, previously unseen attack techniques, we further analyse the performance of unsupervised learning techniques that detect the anomalies in signal parameters introduced by attacks. The presented findings help achieve timely and accurate detection of physical-layer attacks and serve as a prerequisite for fast and effective attack response and network recovery.

Optical signal to noise ratio (OSNR) is a critical performance parameter of an optical fiber communication system. A method for OSNR monitoring based on coherent optical sampling technology is presented. The estimated error of a 10-GBd QPSK signal is <1 dB with the actual OSNR in range of 12 to 30 dB. The influence of uneven spectrum of a local oscillator on the measured accuracy is investigated, and a compensated algorithm for such unevenness is provided. Our method has the advantage of lower complexity and can be extended to higher symbol rate communication system requiring no high bandwidth ADC hardware.

In this work, we investigate the effects of estimation uncertainties in the optimization of the optical signal-to-noise ratio (OSNR) at the wavelength division multiplexing (WDM) optical networks. The OSNR optimization increases the network throughput and energy efficiency and it enables to minimize the number of retransmissions by higher layers. The OSNR optimization algorithms are related to the optical channel estimation performed by the combinations of optical performance monitors (OPM) measurements, analytical models and numerical simulations. Furthermore, the estimation uncertainties are introduced by several factors such as the availability of monitoring information, monitoring accuracy, imperfection of physical layer and dynamic of channel allocation. Our numerical results have demonstrated the relation between the uncertainties in the channel error estimation, normalized mean squared error (NMSE) and power penalty. The developed methodology can be utilized to determinate the maxim level of acceptable uncertainties in the channel error estimation according the power penalty permitted in the OSNR optimization.

In this paper, asynchronous complex histogram (ACH)-based multi-task artificial neural networks (MT-ANNs), are proposed to realize modulation format identification (MFI), optical signal-to-noise ratio (OSNR) estimation and fiber nonlinear (NL) noise power estimation simultaneously for coherent optical communication. Optical performance monitoring (OPM) is demonstrated with polarization mode multiplexing (PDM), 16 quadrature amplitude modulation (QAM), PDM-32QAM, as well as PDM-star 16QAM (S-16QAM) for the first time. The range of launched power is −3 to −2 dBm with a fiber link of 160–1600 km. Then, the accuracy of MFI reaches 100%. The average root mean square error (RMSE) of OSNR estimation can reach 0.37 dB. The average RMSE of NL noise power estimation can reach 0.25 dB. The results show that the monitoring scheme is robust to the increase of fiber length, and the solution can monitor more optical network parameters with better performance and fewer training data, simultaneously. The proposed ACH MT-ANN has certain reference significance for the future long-haul coherent OPM system.

Optical performance monitoring (OPM) aims to estimate the amount of distortion in optical networks. Its importance relies on building robust and efficient networks with self and dynamic diagnosis. In this work, channel impairment monitoring is investigated for optical wireless communication. The monitoring is achieved using a support vector machine (SVM) regressor. A cost-effective and straightforward acquisition system is used to build the training features, which are asynchronous amplitude histograms. Three different parameters related to three common channel impairments are monitored using these features:optical signal-to-noise ratio (OSNR), visibility range, and $\xi$ parameter related to pointing error. First, each parameter is monitored when there is only one isolated channel impairment. Then, each parameter is monitored when two and three jointly channel impairments occur. The results showed that using this low complex machine learning (ML) technique, the achieved prediction accuracy was very high ( $>$ 0.98) for most channel conditions except for the monitoring of the OSNR parameter, where the prediction accuracy dropped to 0.86 under harsh channel conditions. Moreover, the superiority of ML-based techniques is compared with non-ML-based techniques for the OSNR parameter monitoring. The results indicated that the ML-based technique achieved high prediction accuracy than the non-ML-based technique, especially for harsh channel conditions.

Based on the peak to valley ratio (PTVR) of the average magnitude difference function (AMDF), we present a novel optical signal to noise ratio (OSNR) and symbol rate (SR) estimation method for commonly used auxiliary amplitude modulations (AAMs). Moreover, it is demonstrated that the influence of chromatic dispersion (CD) on the method can be mitigated by maximizing the PTVR of the AMDF with additional tunable dispersion compensators. The results of simulations show that the OSNR estimation error can be kept within 0.8 dB in the wide OSNR range of (12, 32) dB, while the SR estimation error is below 0.079% for four widely used 10 Gsymbol/s AAM signals.

Optical performance monitoring (OPM) is the estimation and acquisition of different physical parameters of transmitted signals and various components of an optical network. OPM functionalities are indispensable in ensuring robust network operation and plays a key role in enabling flexibility and improve overall network efficiency. We review the development of various OPM techniques for direct-detection systems and digital coherent systems and discuss future OPM challenges in flexible and elastic optical networks.