SOA pattern effect mitigation by neural network based pre-equalizer for 50G PON

Semiconductor optical amplifier (SOA) is widely used for power amplification in O-band, particularly passive networks (PONs) which can greatly benefit its advantages of simple structure, low consumption and integrability with photonics circuits. However, the annoying nonlinear pattern effect degrades system performance when SOA needed as a pre-amplifier PONs. Conventional solutions mitigation are either based on filtering or gain clamping. They not sufficiently flexible practical deployment. Neural network (NN) has been demonstrated impairment compensation communications thanks to powerful fitting ability. In this paper, first time, NN-based equalizer proposed mitigate 50G PON intensity modulation direct detection. The experimental results confirm that effectively significantly improve dynamic range receiver, achieving 29-dB budget FEC limit at 1e -2 . Moreover, well-trained NN model receiver side be directly placed transmitter line terminal pre-equalize signal transmission so simplify digital processing unit.