Multiuser Coding and Signal-Processing for Fiber-Optic WDM Communication Systems

University Of Virginia Main Campus, Charlottesville VA

Investigators

Abstract

Multiuser Coding and Signal-Processing for Fiber-Optic WDM Communication Systems by Maite Brandt-Pearce University of Virginia Abstract: The need for high speed data transmission has motivated the desire to utilize the full capacity of optical fibers. The fiber optic channel holds the potential for information transfer orders of magnitude faster (in bits per second) than is currently exploited. Signal degradation through the fiber channel is a key problem in accomplishing high capacity, especially in long-haul links (transcontinental or transoceanic) and large networks. This research explores pre- and post-fiber coding and signal processing possible to reduce the effect of channel degradation. Mastering these physical limitations of the fiber medium can translate into a tremendous increase in the capacity of existing infrastructure. This research addresses the design and optimization of multi-channel wavelength division multiplexed (WDM) fiber-optic systems. Multiuser systems are typically broadband and high power, sometimes leading to severe linear and nonlinear degradation in the signal due to properties inherent to the fiber. The approach taken is a communications theoretic approach, considering signal constellation, coding, channel equalization, and joint signal detection. The research is to be performed in three phases. The first is to identify best and worst time/wavelength sequences for WDM. Effective jointly optimal line-codes for WDM systems will then be designed. The research will conclude with a study of nonlinear equalization methods for crosstalk cancellation. The tools used for optimization and system performance analysis are input/output descriptions of the fiber nonlinearity and system simulation.