Modified QPSK partition algorithm based on MAP estimation for probabilistically-shaped 16-QAM

Abstract

Probabilistic shaping (PS) is investigated as a potential technique to approach the Shannon limit. However, it has been proved that conventional carrier phase recovery (CPR) algorithm designed for uniform distribution may have extra penalty in PS systems. In this paper, we find that the performance of QPSK partition algorithm is degenerated when PS is implemented. To solve this issue, a modified QPSK partition algorithm that jointly optimizes the amplitude decision threshold and filter weight is proposed, where the optimization of decision threshold is based on maximum a posterior probability (MAP) estimation. Different from the conventional decision methods which commonly use Euclidean distance metric, the MAP-based decision introduces the statistical characteristics of the received signals to obtain an accurate amplitude partition. In addition, the filter weight is optimized for different decision thresholds to enhance the tolerance of ASE-induced phase noise. We verify the feasibility of the proposed algorithm in a 56 GBaud PS 16-ary quadrature amplitude modulation (16-QAM) system. The proposed algorithm reduces the error of phase noise estimation by nearly half. Compared with conventional QPSK partition, the proposed algorithm could narrow the gap with theoretical mutual information (MI) by more than 0.1 bit/symbol. The channel capacity is increased by 4.2%, 4.3% and 3.6% with signal-to-noise ratio (SNR) from 8 dB to 10 dB respectively. These observations show that the proposed algorithm is a promising method to relieve the penalty of QPSK partition algorithm in PS systems.