Exploiting optical signal analysis for autonomous communications
(English) Optical communications have been extensively investigated and enhanced in the last decades. Nowadays, they are responsible to transport all the data traffic generated around the world, from access to the core network segments. As the data traffic is increasing and changing in both type and...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Estado: | Versión publicada |
| Fecha de publicación: | 2022 |
| País: | España |
| Institución: | CBUC, CESCA |
| Repositorio: | TDR. Tesis Doctorales en Red |
| OAI Identifier: | oai:www.tdx.cat:10803/687343 |
| Acceso en línea: | http://hdl.handle.net/10803/687343 https://dx.doi.org/10.5821/dissertation-2117-379456 |
| Access Level: | acceso abierto |
| Palabra clave: | Àrees temàtiques de la UPC::Informàtica 004 621.3 |
| Sumario: | (English) Optical communications have been extensively investigated and enhanced in the last decades. Nowadays, they are responsible to transport all the data traffic generated around the world, from access to the core network segments. As the data traffic is increasing and changing in both type and patterns, the optical communications networks and systems need to readapt and continuous advances to face the future data traffic demands in an efficient and cost-effective way. This PhD thesis focuses on investigate and analyze the optical signals in order to extract useful knowledge from them to support the autonomous lightpath operation, as well as to lightpath characterization. The first objective of this PhD thesis is to investigate the optical transmission feasibility of optical signals based on high-order modulation formats (MF) and high symbol rates (SR) in hybrid filterless, filtered and flexible optical networks. It is expected a higher physical layer impairments impact on these kinds of optical signals that can lead to degradation of the quality of transmission. In particular, the impact of the optical filter narrowing arising from the node cascade is evaluated. The obtained simulation results for the required optical-signal-to-noise ratio in a cascade up to 10 optical nodes foresee the applicability of these kinds of optical signals in such scenarios. By using high-order MF and high SR, the number of the optical transponders cab be reduced, as well as the spectral efficiency is enhanced. The second objective focuses on MF and SR identification at the optical receiver side to support the autonomous lightpath operation. Nowadays, optical transmitters can generate several optical signal configurations in terms of MF and SR. To increase the autonomous operation of the optical receiver, it is desired it can autonomously recognize the MF and SR of the incoming optical signals. In this PhD thesis, we propose an accurate and low complex MF and SR identification algorithm based on optical signal analysis and minimum Euclidean distance to the expected points when the received signals are decoded with several available MF and SR. The extensive simulation results show remarkable accuracy under several realistic lightpath scenarios, based on different fiber types, including linear and nonlinear noise interference, as well as in single and multicarrier optical systems. The final objective of this PhD thesis is the deployment of a machine learning-based digital twin for optical constellations analysis and modeling. An optical signal along its lightpath in the optical network is impaired by several effects. These effects can be linear, e.g., the noise coming from the optical amplification, or nonlinear ones, e.g., the Kerr effects from the fiber propagation. The optical constellations are a good source of information regarding these effects, both linear and nonlinear. Thus, by an accurate and deep analysis of the received optical signals, visualized in optical constellations, we can extract useful information from them to better understand the several impacts along the crossed lightpath. Furthermore, by learning the different impacts from different optical network elements on the optical signal, we can accurately model it in order to create a partial digital twin of the optical physical layer. The proposed digital twin shows accurate results in modeled lightpaths including both linear and nonlinear interference noise, in several lightpaths configuration, i.e., based on different kind of optical links, optical powers and optical fiber parameters. In addition, the proposed digital twin can be useful to predict quality of transmission metrics, such as bit error rate, in typical lightpath scenarios, as well as to detect possible misconfigurations in optical network elements by cooperation with the software-defined networking controller and monitoring and data analytics agents. |
|---|