Dario Pilori

Accurate estimation of Non-Linear Interference (NLI) is critical for optimizing ultra-wideband optical networks. This paper investigates the performance of Longitudinal Power Monitoring (LPM)-based NLI estimation, focusing on heterogeneous links. While LPM offers robust, link-agnostic estimation of Self-Channel Interference (SCI), it requires correction for Cross-Channel Interference (XCI) to fully assess NLI. Two correction approaches are compared: a simplified Gaussian Noise (GN)-model-based approximation and the more accurate Polynomial Closed-Form GN/EGN model (PCFM). Experimental validation is conducted over both homogeneous and heterogeneous links, including scenarios with unequal channel power, on two different modulation formats. Results show that while estimation accuracy decreases in complex scenarios, LPM-based methods still provide sufficiently reliable performance for practical optimization tasks, such as determining optimal launch power. These findings support the use of LPM-based NLI estimation as a promising tool for real-time network monitoring and digital twin modeling.

Accurate estimation of Non-Linear Interference (NLI) is critical for optimizing ultra-wideband optical networks. This paper investigates the performance of Longitudinal Power Monitoring (LPM)-based NLI estimation, focusing on heterogeneous links. While LPM offers robust, link-agnostic estimation of Self-Channel Interference (SCI), it requires correction for Cross-Channel Interference (XCI) to fully assess NLI. Two correction approaches are compared: a simplified GN-modelbased approximation and the more accurate PCFM model. Experimental validation is conducted over both homogeneous and heterogeneous links, including scenarios with unequal channel power, on two different modulation formats. Results show that while estimation accuracy decreases in complex scenarios, LPMbased methods still provide sufficiently reliable performance for practical optimization tasks, such as determining optimal launch power. These findings support the use of LPM-based NLI estimation as a promising tool for real-time network monitoring and digital twin modeling.

We experimentally demonstrate the extraction of per-span nonlinear interference (NLI) contributions using Linear Least Squares Longitudinal Power Monitoring (LLS-LPM). This receiver-side DSP technique enables accurate estimation of the nonlinear signal-to-noise ratio (SNRNL) for each span in an optical link. By combining these estimates with optical signalto-noise ratio (OSNR) measurements, we propose a simple and effective method to determine the optimal launch power on a per-span, per-channel basis. This approach improves power optimization in wavelength-division multiplexed (WDM) systems, particularly in scenarios with span heterogeneity or channel power imbalance. Experimental validation over a 5×85 km G.652 SMF link confirms the effectiveness of the method and highlights its advantages over cumulative metrics traditionally used at the receiver.

We experimentally assess the accuracy of NLI estimation using Longitudinal Power Monitoring over a 1483-km heterogeneous link (comprising SMFs and PSCFs) with unequalized channel powers, comparing two different SCI/XCI correction factors.

We experimentally assess the accuracy of NLI estimation using Longitudinal Power Monitoring over a 1483-km heterogeneous link (comprising SMFs and PSCFs) with unequalized channel powers, comparing two different SCI/XCI correction factors.

This paper presents an experimental demonstration of the Linear-Least Squares Longitudinal Power Monitoring (LLS-LPM) technique over a mixed fiber link. The link is made by five spans, including four spans of G.652 Single-Mode Fiber (SMF) and one span of G.655 Non-Zero Dispersion-Shifted Fiber (NZDSF) with negative dispersion in the C-band. The study highlights the challenges posed by the negative dispersion of G.655 fiber and its impact on the convergence of the LLS-LPM algorithm. The experimental results demonstrate that the LLS-LPM algorithm can accurately estimate the power profile in the portion of the link where the cumulated chromatic dispersion is unambiguous. This work provides valuable insights into the application of LLS-LPM in real-world fiber networks with mixed fiber types.

This letter presents an experimental demonstration of the Linear-Least Squares Longitudinal Power Monitoring (LLS-LPM) technique over a mixed fiber link. The link is made by five spans, including four spans of G.652 Single-Mode Fiber (SMF) and one span of G.655 Non-Zero Dispersion-Shifted Fiber (NZDSF) with negative dispersion in the C-band. The study highlights the challenges posed by the negative dispersion of G.655 fiber and its impact on the convergence of the LLS-LPM algorithm. The experimental results demonstrate that the LLS-LPM algorithm can accurately estimate the power profile, only in the link portion where there is a one-to-one relation between the cumulated chromatic dispersion and the distance. This work provides valuable insights into the application of LLS-LPM in real-world fiber networks with mixed fiber types.

Given the ubiquity of optical fiber networks in both terrestrial and submarine environments, leveraging these facilities for sensing anomalous conditions alongside telecommunications can provide significant added value. In this context, distributed acoustic sensing (DAS) systems have been widely employed and discussed due to their sensitivity and ability to locate events. However, integrating them within existing networks is complex and expensive. On the other hand, the received state of polarization (SOP) is also sensitive to external factors, and it can be used for sensing: in this case, no extra hardware would be required since the SOP is already estimated in coherent receivers for data demodulation. The sensing information is provided “for free” by the already installed hardware, potentially requiring only a software upgrade. In this work, we analyze the feasibility of using polarization-based sensing to detect anomalous conditions in metropolitan environments. A polarimeter was used to evaluate SOP noise induced by urban factors, while a commercial coherent transceiver was employed to assess SOP estimation noise. We propose two algorithms for processing polarization data: a time-based method called SOP angular speed (SOPAS) and an adaptive, frequency-based approach named SOP-power spectral density gap (SOP-PSDG). These algorithms were compared by processing Stokes vector samples from the polarimeter when different sinusoidal vibrations are applied to the fiber through a mechanical shaker. Results demonstrate that a sampling rate of just a few tens of Hz is sufficient to effectively identify various hazardous conditions, with SOP-PSDG consistently outperforming SOPAS. Additionally, preliminary findings on the performances of these algorithms using SOP samples from a commercial coherent receiver are discussed.

A novel method for the estimation of nonlinear interference (NLI) penalties in optical networks is presented. Its theoretical derivation is grounded in linear least squares (LLS) longitudinal power monitoring (LPM), an algorithm capable of estimating the absolute power evolution of a signal along the link. We demonstrate that closed-form NLI estimation is inherently integrated into LPM. However, this method does not account for the cross-channel interference (XCI) contribution to NLI, since the receiver typically lacks knowledge of other WDM channels. To address this limitation, simple analytical expressions to compensate for the bias originating from neglecting XCI are proposed for the scenario of homogenous WDM combs. These formulas are derived from closed-form NLI models (e.g., the GN-model) and require only limited information about the signal configuration. The effectiveness of the proposed approach is validated through a wide range of realistic numerical simulations and experiments, including C-band transmissions over 300-km and 1100-km EDFA-amplified optical links. The obtained results confirm the reliability of the method, yielding nonlinear SNR estimates with a maximum absolute error consistently below 0.97 dB when compared to well-established analytical NLI models or measured values. The mean absolute errors in simulation and experiments are 0.3 dB and 0.5 dB, respectively. Additionally, a practical demonstration of transmit power optimization is provided, leveraging the values obtained through the proposed method.

We review methods for separating ASE and nonlinear interference noise in coherent receivers, emphasizing techniques and potential pitfalls. Accuracy and practical challenges are evaluated, with a focus on the Longitudinal Power Monitoring-based approach.