Fuzzy clustered time-domain dispersion compensation for long-haul coherent optical links

TY - JOUR

T1 - Fuzzy clustered time-domain dispersion compensation for long-haul coherent optical links

AU - Liu, Zhiheng

AU - Yang, Aiying

AU - Guo, Peng

AU - Wan, Wenkai

AU - Zhao, Zhe

AU - Xu, Tianjia

AU - Dong, Yi

N1 - Publisher Copyright: © 2026 Elsevier Inc.

PY - 2026/7

Y1 - 2026/7

N2 - In long-haul coherent optical fiber communication systems, chromatic dispersion compensation (CDC) is essential but computationally intensive. Conventional time-domain CDC (TD-CDC) requires lengthy filters, while frequency-domain CDC (FD-CDC) introduces latency and integration issues. Although hard-decision tap-clustered scheme reduces the TD-CDC complexity, its benefit diminishes over very long links where tap coefficients disperse widely. To address this, we propose a novel two-stage fuzzy-clustered TD-CDC scheme. It employs a soft-decision strategy, allowing filter taps to partially belong to two nearest neighboring clusters. Simulation results for 32 GBaud under 3200 km fiber length demonstrate 45% and 15% complexity reductions compared to clustered TD-CDC and FD-CDC, respectively, at the 6.7% HD-FEC threshold. Furthermore, across various baud rates and fiber lengths, the fuzzy-clustered TD-CDC scheme reduces complexity compared to clustered TD-CDC without compromising performance. Experimental validation on a 20 GBaud, single-polarization 16QAM standard single-mode fiber communication system over 1800 km demonstrates 53.8% and 40% complexity reductions compared to clustered TD-CDC and FD-CDC, respectively, at the 20% HD-FEC threshold. Moreover, the proposed scheme matches the optimal Q-factor of FD-CDC with 27% lower complexity.

AB - In long-haul coherent optical fiber communication systems, chromatic dispersion compensation (CDC) is essential but computationally intensive. Conventional time-domain CDC (TD-CDC) requires lengthy filters, while frequency-domain CDC (FD-CDC) introduces latency and integration issues. Although hard-decision tap-clustered scheme reduces the TD-CDC complexity, its benefit diminishes over very long links where tap coefficients disperse widely. To address this, we propose a novel two-stage fuzzy-clustered TD-CDC scheme. It employs a soft-decision strategy, allowing filter taps to partially belong to two nearest neighboring clusters. Simulation results for 32 GBaud under 3200 km fiber length demonstrate 45% and 15% complexity reductions compared to clustered TD-CDC and FD-CDC, respectively, at the 6.7% HD-FEC threshold. Furthermore, across various baud rates and fiber lengths, the fuzzy-clustered TD-CDC scheme reduces complexity compared to clustered TD-CDC without compromising performance. Experimental validation on a 20 GBaud, single-polarization 16QAM standard single-mode fiber communication system over 1800 km demonstrates 53.8% and 40% complexity reductions compared to clustered TD-CDC and FD-CDC, respectively, at the 20% HD-FEC threshold. Moreover, the proposed scheme matches the optimal Q-factor of FD-CDC with 27% lower complexity.

KW - Chromatic dispersion compensation

KW - Coherent optical communication systems

KW - Computational complexity

KW - Digital signal processing

KW - Optical fiber transmission

UR - https://www.scopus.com/pages/publications/105029054166

U2 - 10.1016/j.yofte.2026.104581

DO - 10.1016/j.yofte.2026.104581

M3 - Article

AN - SCOPUS:105029054166

SN - 1068-5200

VL - 99

JO - Optical Fiber Technology

JF - Optical Fiber Technology

M1 - 104581

ER -