Nouvelles techniques de compensation des effets nonlinéaires pour les transmissions optiques à longue distance

par Wasyhun asefa Gemechu

Projet de thèse en Réseaux, information et communications

Sous la direction de Yves Jaouen et de Mansoor Isvand yousefi.

Thèses en préparation à Paris Saclay en cotutelle avec l'University of brescia , dans le cadre de École doctorale Sciences et technologies de l'information et de la communication (Orsay, Essonne) , en partenariat avec Laboratoire de Traitement et Communication de l'Information (laboratoire) , Comelec/GTO : Télécommunications Optiques (equipe de recherche) et de Télécom ParisTech (établissement de préparation de la thèse) depuis le 01-04-2016 .

  • Titre traduit

    Advanced nonlinearity Compensation techniques for long-haul optical Transmission systems


  • Résumé

    Context: In recent years, the exponentially increasing data traffic and consequent need to increase the capacity of optical communication networks has added more pressure on network infrastructure and design engineers to come up with better signal processing and transmission techniques. A new transmission technology was introduced in 2005, namely the orthogonal frequency division multiplexing (OFDM) method for long-haul high-speed optical transmission systems in the intensity modulation-direct detection (IM/DD) environment. The revival of coherent detection transmissions led to CO-OFDM as a highly competitive technology for next generation high capacity optical systems. CO-OFDM has shown tremendous potential for Tb/s, polarization diversity multiplexed transmissions, based on the compensation at the receiver of linear signal impairments due to dispersion and PMD. However compensating the ASE-noise accumulated in the fiber network remains challenging, unless a substantial increase of the OSNR was permitted, which in turn is prevented by the presence of nonlinear effects in the fiber channel. To date, it is clear that the Kerr fiber nonlinearity poses the major threat to optical communication systems. Hence novel and possibly breakthrough solutions for compensating the Kerr effect of the optical fiber channel are absolutely necessary for solving the upcoming capacity crunch. Thesis Work Description: This thesis work will develop new techniques for the mitigation first, and possibly full compensation next, of nonlinear fiber impairments in wavelength and polarization multiplexed CO-OFDM transmissions. Building from already existing fiber models and compensation methods like electronic compensators and DCF modules that increase circuitry complexity and fiber length, improved versions of the Volterra Series Transfer function (VSTF) method will be developed and numerically demonstrated as nonlinear-equalizer modules. Unlike other compensation techniques, the VSTF method has the capability to simultaneously compensate for linear attenuation and dispersion, weak nonlinear effects and accumulated ASE-noise from optical amplifiers, both in the single channel and in the multi-channel (WDM) environment. The spot light will be on the truncated third-order VSTF, to model and compensate for the most relevant nonlinear effects as described by the nonlinear Schrödinger equation (NLSE), namely: Self-Phase Modulation (SPM), Cross-Phase Modulation (XPM) and Four-Wave Mixing (FWM), both individually and combined. Higher-order terms will also be included in the VSTF method, and simulated for a better approximation of (supposedly weak) nonlinear effects. The performance of such equalizers will be evaluated using BER-metric and computational latency for both 100 GB and 400 GB-Nyquist channel single-channel and WDM transmissions. However, the VSTF only represent a first step towards the ultimate goal of reaching the highest spectral efficiency, which is fundamentally limited by the nonlinear Shannon channel capacity theorem. Present design of optical communication systems is based on the hypothesis that nonlinearity is a small perturbation (nonlinear noise) to a quasi-linear system. The development of a new kind of nonlinear optical communication system, such that nonlinearity can be fully compensated for, will enable a breakthrough increase of the OSNR, and a large leap in spectral efficiency. Therefore the second part of the thesis will focus on the development of a practical nonlinear communication method, based on the theory of the inverse spectral transform. This method, originally proposed by Hasegawa in 1993 and called eigenvalue (or multi-soliton) communication, is based on the fundamental observation that the discrete nonlinear spectrum of an optical signal is invariant (except for a trivial linear phase shift) upon propagation in the fiber channel, as described by the scalar NLSE. The method was never implemented, since it is was not suitable for the IM/DD systems of the time. However the recent development of practical coherent detectors based on fast digital signal processing enables the real-time measurement of both quadrature components of the received field. This means that the direct spectral transform (also known as nonlinear Fourier transform) of the received signal can be computed, and the eigenvalue spectrum fully recovered. Our first goal will thus be to find the optical input signal constellations, which lead to a signal alphabet such that their members enjoy the largest distance (and possibly so that they are even orthogonal) in the complex plane of their eigenvalues. The second goal will be to numerically evaluate and optimize the transmission performance and spectral efficiency of practical transmission systems based on the eigenvalue communication technique, both in the single and in the multi-channel environment. The third and final goal will be to extend the eigenvalue communication technique to include the polarization dimension, so as to double the channel transmission capacity by polarization modulation or multiplexing. To that end, the nonlinear Fourier transform of the vector NLSE, or Manakov system, will be numerically developed and implemented, and optimal polarization modulation formats will be developed.