Modeling nonlinear phase noise in differentially phase-modulated optical communication systems

Abstract

Using an alternative approach for evaluating the Bit-Error Rate (BER), we present a numerical and experimental investigation of the performance of phase-modulated optical communication systems in the presence of nonlinear phase noise and dispersion. The numerical method is based on the well known Karhunen-Lo;eve expansion combined with a linearization technique of the Nonlinear Schr odinger Equation (NLSE) to account for the nonlinear interaction between signal and noise. Our numerical results show a good agreement with experiments.

Impact of PMD and Nonlinear Phase Noise on the Global Optimization of DPSK and DQPSK Systems

“…1, the BER can be exactly evaluated using the Karhunen-Loéve (KL) method [1] only if the received optical noise is white and Gaussian. If fiber nonlinearities are of concern, linearization techniques of the NLSE can be applied together with KL method for BER evaluation in the presence of nonlinear phase noise (NPN) [2]. This method will be called here Extended KL method.…”

“…However, if the BER has to be computed several times, the overall computational effort is so large that it becomes prohibitive for system optimization. In this case, the BER can be calculated very fast and accurate by using linearization techniques of the Nonlinear Schrödinger Equation (NLSE) together with Karhunen-Loéve series expansion [2].…”

“…In this paper, we use a global optimization algorithm [7] together with an extended Karhunen-Loéve method [2] and the nonlinear phase-shift criterion [4] to find the maximum reach of single-channel DPSK and DQPSK systems including PMD and nonlinear phase noise, induced by the interaction between the signal and ASE noise. We extend the results presented in [7] and also show the impact of both effects on the maximum reach for data rates ranging from 5 Gbit/s to 230 Gbit/s.…”

BER estimation for multilevel modulation formats

“…Another limitation of the deterministic analysis is that it is necessary to keep the noise information separated from the signal samples during simulation, and thus, neglecting potential nonlinear interactions between signals and noise, like nonlinear phase noise (NLPN). Note that modeling NLPN as a parametric gain process affecting the ASE noise (colored instead of white) rather than the signal, as proposed in [12], [13] allows to estimate the performance of a system being affected by NLPN with deterministic methods.…”

Moment-generating function method used to accurately evaluate the impact of the linearized optical noise amplified by EDFAs

“…5, we also simulate the experiments of the multi-span DPSK systems discussed in Ref. [7] and show that, to fit the experimental data, one needs to take into account the nonlinearity induced phase difference between noise and noise-free signal, which will affect the signal-noise beating significantly.…”