Experimental Investigation of Gain Offset Behavior of Feedforward-Controlled WDM AGC EDFA Under Various Dynamic Wavelength Allocations

Hybrid Machine Learning EDFA Model

“…The EDFA gain spectrum can be modelled using a center of mass function based on single channel and fully channel loaded ripple functions, shown in Eq. 1 [5].…”

“…Wavelength dependent gain in EDFAs is a significant source of variation and uncertainty in channel performance estimation. Recently there has been much interest in developing machine learning based models to account for the detailed EDFA behavior [1][2][3], replacing lookup tables or analytical models [4][5][6][7][8][9]. However, these machine learning models are built solely from a-posteriori knowledge which is trained from the experimental data, ignoring the existing a-priori knowledge.…”

Machine Learning Assisted EDFA Gain Ripple Modelling for Accurate QoT Estimation

“…The dataset used for training those models were either i) analytical without considering fiber effects/NLI noise contribution, or ii) experimental, requiring the collection of data for each individual span EDFA, which practically limits the usability of these approaches. In [12] the gain offset of a AGC-EDFA was experimentally investigated and a method was derived aiming at reducing the gain offset up to 0.5 to 0.7dB when encompassing up to five cascaded EDFA spans. The scenario targeted in [12] was dynamic with add/drop of new/existing connections.…”

“…In [12] the gain offset of a AGC-EDFA was experimentally investigated and a method was derived aiming at reducing the gain offset up to 0.5 to 0.7dB when encompassing up to five cascaded EDFA spans. The scenario targeted in [12] was dynamic with add/drop of new/existing connections. However, for fixed load EDFAs, with static network conditions, the gain ripples effect was not explored.…”

Dual-wavelength source based optical circuit switching and wavelength reconfiguration in multi-hop ROADM systems

“…Analytical models have also been developed to predict and minimize power excursions. The wavelength dependent gain was estimated by measuring the fully loaded WDM channel gain and single channel gain, and generated an estimated gain spectrum to perform wavelength assignment in order to minimize power excursions [14,15]. 5%-15% power excursion reductions were reported in a single-hop end-to-end network with five cascading EDFAs.…”