Nonlinear frequency-division multiplexing: theoretical aspects, numerical algorithms, and experimental demonstration

The exponential growth of global data traffic is challenging the optical network, currently limited by Kerr nonlinearity. This thesis investigates a novel transmission paradigm for long haul optical fiber communication, expected to outperform conventional systems. This technique, commonly known as nonlinear frequency-division multiplexing (NFDM), uses the nonlinear Fourier transform (NFT) to encode information on the nonlinear spectrum, whose evolution along the optical fiber is, under some circumstances, trivial. For this reason, NFDM masters nonlinearity, and is expected to increase the capacity of the current optical networks. The first part of this thesis is dedicated to the NFT as a tool for optical communication: the mathematical framework is introduced and numerical algorithms for the NFT are developed. Secondly, the implementation of an NFDM scheme is described and the performance, obtained through simulations, is shown. The results are discussed and some important limitations are highlighted. Next, in the light of the above, three novel detection strategies are introduced that provide significant performance improvements (up to 6.2 dB). Then, the main concepts are extended to dual polarization systems, and a reduced complexity paradigm is introduced. Finally, the first experimental demonstration of a dual-polarization NFDM scheme modulating both the continuous and the discrete spectra is described and the results are shown.