Nonlinear-tolerant two-dimensional distribution matcher scheme for probabilistic shaping

Abstract

This Letter proposes a nonlinear-tolerant two-dimensional distribution matcher (2D-DM) scheme. It removes the corner points of probabilistically shaped quadrature amplitude modulation (QAM) to obtain better nonlinear tolerance. Because the remaining number of points is not a power of 2, we propose to divide constellation points into different layers and symbols. Then, the proposed 2D-DM performs matching using one-dimensional shapers, which generates the in-phase and quadrature components of QAM together. In fact, it realizes two-dimensional shapers from one-dimensional shapers. Simulation results show that two-dimensional shapers generated by the proposed 2D-DM have higher tolerance to power amplifier nonlinearity and fiber nonlinearity compared to one-dimensional shapers.