Three-dimensional constellation probabilistic shaping scheme based on fine-grained data reconstruction
Hyun Woo Ji, Jianxin Ren, Bo Liu, Yaya Mao, Jianye Zhao, Shuaidong Chen, xiumin song, Rahat Ullah, Lilong Zhao, Yilan Ma · Optics Express · 2025
This paper proposes a three-dimensional (3D) constellation probabilistic shaping (PS) scheme based on fine-grained data reconstruction (FGDR). The goal is to integrate 3D orthogonal frequency division multiplexing (3D-OFDM) technology with PS techniques to improve system spectral efficiency and capacity. The proposed FGDR-PS technique utilizes the fine-grained characteristics of the bitstream at the transmitter side to adjust the proportion of each bit, thereby achieving constellation shaping. To verify the scheme's effectiveness, experiments were conducted over a 25 km single-mode optical fiber, testing the performance of different schemes at the same net rate. Experimental results indicate that at a bit error rate (BER) of 3.8 × 10 −3 , the proposed 3D-PS-32QAM-Scheme2 achieves a gain of 1.9 dB, 1.2 dB, 0.8 dB, and 0.5 dB compared to 2D-PS-32QAM, 3D-32QAM-Scheme1, 3D-32QAM-Scheme2, and 3D-PS-32QAM-Scheme1, respectively.