Modeling the Bit-Error-Rate Performance of Nonlinear Fiber-Optic Systems

We present a detailed statistical model of nonlinear interference noise (NLIN) in optical communication systems. We demonstrate an efficient method of calculating second-order statistics of the NLIN coefficients, particularly their temporal autocorrelation and cross correlation. The model is highly accurate in predicting system performance metrics such as bit-error-rate and signal-to-noise ratio, and is shown to provide better accuracy with respect to models that use the NLIN variance alone, particularly when accounting for the adaptive filtering of realistic receivers.