Low complexity equalization in coherent optical communication based on Fermat number transform

In order to reduce the power consumption of the receivers in coherent optical communication system during digital signal processing (DSP), a low complexity equalization scheme in DSP flow of a 400Gb/s-DP-16QAM system has been proposed. This scheme is based on Fermat number transform (FNT), which sequentially performs static equalization (SE) and dynamic equalization (DE) in the transform domain. For different distances, the proposed scheme finds the optimal solution under the condition that transform length and data bitwidth are mutually restricted under different transmission distances, while achieving low complexity and optimal performance. The experimental results show that the adopted TrDE scheme has better computational complexity than traditional frequency domain equalization (FDE) and time domain equalization (TDE) nearly without any performance loss. In the 80km, 160km, and 240km scenarios, the number of multiplier is reduced by more than 72%, and the advantage becomes more obvious as the transmission capacity increases.