Nonlinear phase noise compensation in single-span digital coherent optical systems employing artificial neural networks

Abstract

As digital coherent technology gets mature and its cost reduces, it is becoming a competitive solution for future implementations of high-capacity long-reach passive optical networks (LR-PON). In the case of single-channel LR-PONs, the system performance is ultimately limited by the combined effect of the receiver additive noise and the nonlinear phase noise, which in turn is a consequence of the interplay between dispersion and Kerr-induced self-phase modulation. In this paper, we show that by employing a three-layer artificial neural network (ANN) to mitigate the effect of nonlinear phase noise, the bit error rate is reduced from 6.8e-4 to 4.9e-4