Nonlinear compensation technologies for future optical communication systems

Digital nonlinear compensation techniques have been thought to be keys to realize further spectrally efficient optical fiber communication systems. The most critical issue of the digital nonlinear compensation algorithms has been their computational complexity, or gate count of digital signal processing circuit. Among several approaches, digital nonlinear compensation algorithms based on perturbation analysis are attractive in terms of the hardware efficiency because the algorithms can compensate the accumulated nonlinear noise over all transmission spans with only one stage. In this paper, we discuss three approaches to sophisticate the perturbation nonlinear compensation. First, we illustrate a perturbation-based post-equalization method to improve the robustness to transceiver device imperfections. We next propose and numerically evaluate a symbol degeneration method to extend the perturbation nonlinear compensation methods to higher-order QAM without increasing the computational complexity. Finally, we discuss a sub-band processing of perturbation nonlinear compensation for further computational complexity reduction. By combining the perturbation method with Nyquist frequency division multiplexing, the computational complexity of perturbation calculation is reduced by a factor of more than 10 for 3000-km single-channel transmission of 128 Gbit/s dualpolarization QPSK with only 0.1 dB performance degradation.