A fast and accurate method for evaluating joint second-order PMD statistics

A fast and efficient way for evaluating the joint probability density function of the polarization-mode dispersion (PMD) vector derivative components parallel and orthogonal to the PMD vector itself is presented. This allows for a fully analytical evaluation of outage probabilities, overcoming thus the limitation of the Monte Carlo method, which do not allow for reliable evaluation of probabilities smaller than about 10/sup -4/.