Dynamic Channel Modeling for Mode-Division Multiplexing

Abstract

Environmental perturbations, such as wind, mechanical stress, temperature, and lightning, impose microsecond-time-scale changes to the transfer matrix of a multimode fiber (MMF), necessitating adaptive multiple-input multiple-output (MIMO) equalization to track the time-varying channel. It is of significant interest to accurately model channel dynamics so that adaptive MIMO equalization algorithms can be optimized and their impact on digital signal processing complexity and performance can be assessed. We propose a dynamic channel model using coupled-mode theory to describe time-varying polarization- and spatial-mode coupling in MMF caused by fast environmental perturbations. Our method assumes that the MMF has built-in refractive index and geometric defects that are responsible for random birefringence and mode coupling. Various time-varying perturbations modify the coupling coefficients in the coupled-mode equations to drive channel dynamics. We use our dynamic channel model to simulate an example aerial fiber link subject to a sudden gust of wind and show that increasing the strength of the perturbation, the number of modes, and the length of exposed fiber all lead to higher error at the output of an adaptive MIMO equalizer. We employ the model to study the convergence and tracking performance of the least-mean-squares adaptive equalization algorithm.