Communication Using the Nonlinear Fourier Transform

ongoing

High-speed optical fiber networks form the backbone of the information and communication technologies, including the Internet. More than 99% of the Internet data traffic is carried by a network of global optical fibers. Despite their great importance, today's optical fiber networks face a looming capacity crunch: The achievable rates of all current technologies characteristically vanish at high input powers due to distortions that arise from fiber nonlinearity. The solution of this long-standing complex problem has become the holy grail of the field of the optical communication. The aim of this project is to develop a novel foundation for optical fiber communication based on the nonlinear Fourier transform (NFT). The NFT decorrelates signal degrees-of-freedom in optical fiber, in much the same way that the conventional Fourier transform does for linear systems. My collaborators and I have recently proposed nonlinear frequency-division multiplexing (NFDM) based on the NFT, in which the information is encoded in the generalized frequencies and their spectral amplitudes (similar to orthogonal frequency-division multiplexing). Since distortions such as inter-symbol and inter-channel interference are absent in NFDM, it achieves data rates higher than conventional methods. The objective of this proposal is to advance NFDM to the extent that it can be built in practical large-scale systems, thereby overcoming the limitation that fiber nonlinearity sets on the transmission rate of the communication networks. The proposed research relies on novel methodology and spans all aspects of the NFDM system design, including determining the fundamental information-theoretic limits, design of the NFDM transmitter and receiver, algorithms and implementations. The feasibility of the project is manifest in preliminary proof-of-concepts in small examples and toy models, PI's leadership and track-record in the field, as well as the ideal research environment.