Low complexity exponential pruning learned digital back-propagation method for fiber nonlinearity mitigation

• DOI: 10.1364/ol.555981
• Publication Date: 2025-02-24T13:00:41Z
• Authors: Lyu Li; Zekun Niu; Hang Yang; Junzhe Xiao; Guozhi Xu; Weisheng Hu; Lilin Yi
• Keywords: Pruning

Learned digital back-propagation (LDBP) is emerging as a promising solution to mitigate the nonlinear fiber distortions that limit capacity of optical communications. However, computational complexity LDBP, which affected by dispersion compensation, will increase significantly communications move toward higher baud rates and longer transmission distances. Herein, we propose what believe be novel method called exponential pruning LDBP (EP-LDBP), achieved redundant taps with adjustable parameters. Experimental results show EP-LDBP achieves 61.5% reduction in compared frequency domain without sacrificing compensation performance 21-channel wavelength division multiplexing (WDM) over 1600 km using 60 Gbaud 16-ary quadrature amplitude modulation (16QAM). Furthermore, our analysis under varying (30-60 G) distances (400-1600 km) demonstrates its superior potential reducing complexity, thus aligning more effectively evolving landscape