Phase Noise Estimation and Compensation in 100G 4-QAM CO-OFDM system using Radial Basis Function Network

We propose an efficient and less complex phase estimation and compensation method using radial basis function network (RBFN). The RBFN is used to exploit the channel correlation in the frequency domain. In this paper, we provide the compensation of laser phase noise (LPN) and nonlinear phase noise in 4-QAM coherent optical orthogonal frequency division multiplexing (CO-OFDM) system. Simulation results show that the proposed RBFN algorithm efficiently estimates the LPN and trajectory matches with the actual LPN. The RBF network also reduces the computational complexity and the RBF subnetworks to a great extent compared to $K$-means based RBFNN-NLE technique. With RBFN-PE algorithm, the optimum launch power (LP) increases to 1 dBm and at this LP, Q-factor improves by 2.6 dB. Furthermore, simulation results show that at an FEC limit (Q = 8 dB), the transmission reach increases by 4 spans with the RBFN-PE algorithm. Thus, we show that the system performance improves by increase in optimum LP and transmission reach with the proposed RBFN-PE.