Multi-Impairments Equalization Enabled by a Pruning Neural Network in Single Sideband-Direct Detection 200Gbps PON Downlink

The SSB-DD scheme is expected to be a potential solution for the next generation PON over 50Gbps, because it can overcome the power fading. However, due to various impairments such as chromatic dispersion, SSBI, fiber nonlinearity and bandwidth limitation existed in the system at the same time, it is difficult for the traditional equalization scheme to effectively mitigate them. This paper proposed a multi-impairment equalization scheme based on a pruning neural network. It can simultaneously equalize multiple impairments and has excellent equalization performance. The 200Gbps simulation results under 20km fiber transmission show that the system can achieve a power budget of 42.5dB with EDFA power-amplifier and SOA pre-amplifier.