Digital Nonlinear Compensation Based on the Modified Logarithmic Step Size

Enhancing the Signal Power Symmetry for Optical Phase Conjugation Using Erbium-Doped-Fibre-Assisted Raman Amplification

“…The system FIGURE 7. Experimentally measured and simulated Q 2 factors versus signal launch power in inline transmission systems with/without a midlink OPC using different DRA schemes was simulated by solving the coupled nonlinear Schrödinger equations (Manakov equations) using the well-known splitstep Fourier method [28] with a step size of 0.1 km and the simulated signal power profiles shown in Fig. 3(a).…”

Time-Domain Digital Back Propagation: Algorithm and Finite-Precision Implementation Aspects

“…Using LS-CO, it is possible to optimize the signal in-band response while controlling the out-of-band response, which is very effective when using spectrally limited pulse-shaping such as rootraised cosine (RRC) pulses. In the ensuing evaluation, we focus on a two steps-per-span (StPS) DBP implementation, using modified logarithmic [6] distribution of the steps, with the adjustment factor k=0.4 as in [6].…”

Finite-Precision Optimization of Time-Domain Digital Back Propagation by Inter-Symbol Interference Minimization

“…In this work, we use a two-step-per-span TD-DBP algorithm ( Fig. 1) with modified logarithmic distribution 4 of the steps with the adjustment factor k=0.4. EachD uses a 17-tap least-squares constrained-optimization (LS-CO) FIR filter 3 to compensate for accumulated chromatic dispersion (CD) corresponding to the step length, with the in-band response optimized with respect to the pulse-shaped spectrum.…”

“…The Kerr non-linearity is one of the major reachlimiting factors in fiber-optic communication systems. While there have been significant research efforts on digital signal processing (DSP) nonlinear compensation techniques, such as digital back propagation (DBP), these simulations and experiments generally use floating-point off-line DSP 2,4 . Practical implementation aspects such as finite-precision effects, which are crucial due to the low-power and high-throughput requirements in real-time systems, remain largely unexplored.…”