RBF neural network nonlinear equalizer for 100G 16-QAM coherent optical system

We evaluate the performance of radial basis function neural network (RBFNN) based nonlinear equalizer (NLE) for single carrier (SC) 16-QAM system. We provide the comparison of nonlinearity compensation results of RBFNN-NLE and K - means clustering algorithm in SC 16-QAM system over dispersion managed (DML) and unmanaged links (DUL). The simulation results show that nonlinear penalty reduces by nearly 2 dB and 1 dB for DUL and DML cases respectively.